Dynamic Syntax

Contents
Preface ix
1 Towards a Syntactic Model of Interpretation 1
1.1 Natural Language as a Formal Language? . . . . . . . . . . . . . 1
1.2 Underspecication in Language Processing . . . . . . . . . . . . . 3
1.3 The Representational Theory of Mind . . . . . . . . . . . . . . . 7
1.4 Pronominal Anaphora: Semantic Problems . . . . . . . . . . . . 9
1.4.1 The Problem of Multiple Ambiguity . . . . . . . . . . . . 9
1.4.2 The Problem of Uniqueness . . . . . . . . . . . . . . . . . 12
1.4.3 The Problem of Indirect Reference . . . . . . . . . . . . . 13
1.4.4 Quantication . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4.5 Syntactic Processes of Anaphora . . . . . . . . . . . . . . 15
1.5 The Anaphora Solution { Towards a Representational Account . 16
2 The General Framework 18
2.1 A Preliminary Sketch . . . . . . . . . . . . .. . .. .. .. . .. 19
2.2 The Data Structures of the Parsing Model .. . .. .. .. . .. 27
2.2.1 Atomic Formulae . . . . . . . . . . . .. . .. .. .. . .. 31
2.2.2 Tree Modalities . . . . . . . . . . . . .. . .. .. .. . .. 36
2.2.3 Basic Tree Structures . . . . . . . . .. . .. .. .. . .. 39
2.2.4 Partial Tree Structures . . . . . . . . .. . .. .. .. . .. 41
2.2.5 Requirements . . . . . . . . . . . . . .. . .. .. .. . .. 42
2.2.6 Descriptions of Tree Structures . . . .. . .. .. .. . .. 49
3 The Dynamics of Tree Building 55
3.1 The Parsing Process { A Sketch . . . . . . . . . . . . . . . . . . . 56
3.1.1 A Basic Example . . . . . . . . . . . . . . . . . . . . . . . 56
3.1.2 A Left-Dislocation Example . . . . . . . . . . . . . . . . . 64
3.1.3 Verb-nal Languages and the Grammar-parser Problem . 67
3.2 The Parsing Process Dened . . . . . . . . . . . . . . . . . . . . 76
3.2.1 Computational Rules . . . . . . . . . . . . . . . . . . . . . 80
3.2.2 Lexical Transitions . . . . . . . . . . . . . . . . . . . . . . 89
3.2.3 Pragmatic Actions and Lexical Constraints . . . . . . . . 95
3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
iv
Contents v
4 Linked Tree Structures 103
4.1 Relative Clauses { Preliminaries . . . . . . . . . . .. .. . .. . 103
4.1.1 The LINK Relation . . . . . . . . . . . . . .. .. . .. . 104
4.1.2 The Data Reviewed . . . . . . . . . . . . . .. .. . .. . 105
4.2 The Analysis { A Sketch for English . . . . . . . . .. .. . .. . 109
4.2.1 Dening Linked Tree Structures . . . . . . .. .. . .. . 110
4.2.2 Relativizers Annotating Unxed Nodes . . .. .. . .. . 111
4.3 Relatives: Towards a Dynamic Typology . . . . . . .. .. . .. . 121
4.3.1 Relativizers Projecting a Requirement . . . .. .. . .. . 121
4.3.2 Variation in Locality . . . . . . . . . . . . . .. .. . .. . 128
4.3.3 Topic Structures and Relatives . . . . . . . .. .. . .. . 130
4.3.4 Variation in Order { Head-nal Relatives . .. .. . .. . 133
4.3.5 Head-internal Relatives . . . . . . . . . . . .. .. . .. . 139
4.3.6 The Potential for Lexical Variation . . . . . .. .. . .. . 142
4.4 Genitive Constructions as LINK Structures . . . . .. .. . .. . 144
4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . .. .. . .. . 148
5 Wh Questions: A General Perspective 150
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 150
5.2 The Semantic Diversity of wh Questions . . . . . . . . . . . .. . 151
5.2.1 Scopal Properties of wh Expressions . . . . . . . . . .. . 154
5.3 Wh -initial vs wh -in-situ Structures . . . . . . . . . . . . . . .. . 156
5.3.1 Wh -in-situ Structures . . . . . . . . . . . . . . . . . .. . 158
5.3.2 Wh -in-situ from a Dynamic Perspective . . . . . . . .. . 161
5.4 Expletive wh Structures . . . . . . . . . . . . . . . . . . . . .. . 163
5.4.1 Partial Movement . . . . . . . . . . . . . . . . . . . .. . 163
5.4.2 Partial Movement as a Re ex of a Requirement . . . .. . 169
5.5 Wh Expressions and Scope Eects . . . . . . . . . . . . . . .. . 186
6 Crossover Phenomena 190
6.1 Crossover { The Problem . . . . . . . . . . . . .. .. .. . .. . 190
6.2 Crossover { The Dynamic Account . . . . . . . .. .. .. . .. . 196
6.2.1 Crossover in Relatives . . . . . . . . . . .. .. .. . .. . 196
6.2.2 Crossover Phenomena in Questions . . . .. .. .. . .. . 213
6.3 Summary . . . . . . . . . . . . . . . . . . . . . .. .. .. . .. . 221
7 Quantication Preliminaries 223
7.1 Introduction . . . . . . . . . . . . . . . . . .. . .. .. .. . .. . 223
7.2 Scope Eects and Indenites . . . . . . . .. . .. .. .. . .. . 224
7.3 Quantication . . . . . . . . . . . . . . . . .. . .. .. .. . .. . 231
7.3.1 Quantied NPs . . . . . . . . . . . .. . .. .. .. . .. . 234
7.3.2 Scope . . . . . . . . . . . . . . . . .. . .. .. .. . .. . 239
7.3.3 Term Reconstructions . . . . . . . .. . .. .. .. . .. . 245
7.3.4 Applications { E-type Anaphora . .. . .. .. .. . .. . 249
vi Contents
8 Re ections on Language Design 253

8.1 The Overall Perspective . . . . . . . . . . . . . . . . . . .. . .. 253
8.2 Underspecication and the Formal Language Metaphor .. . .. 259
8.2.1 English is not a Formal Language . . . . . . . . .. . .. 260
8.3 Well-formedness and Availability of Interpretations . . . .. . .. 262
8.4 Universals and Language Variation . . . . . . . . . . . . .. . .. 264
8.5 On Knowledge of Language . . . . . . . . . . . . . . . . .. . .. 266
9 The Formal Framework 268
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 268
9.2 Declarative Structure . . . . . . . . . . . . . . . . . . . . . . . .. 273
9.2.1 Feature-decorated Tree Construction . . . . . . . . . . .. 273
9.2.2 Goal-directedness . . . . . . . . . . . . . . . . . . . . . .. 291
9.2.3 The Structure of Goal-directed Partial Tree Models . .. 297
9.2.4 Tree Descriptions . . . . . . . . . . . . . . . . . . . . . .. 302
9.3 Procedural Structure . . . . . . . . . . . . . . . . . . . . . . . .. 307
9.3.1 Actions over Goal-directed Partial Tree Models . . . . .. 308
9.3.2 Natural Languages . . . . . . . . . . . . . . . . . . . . .. 312
9.4 Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 317
9.4.1 Finite Binary Trees . . . . . . . . . . . . . . . . . . . . .. 317
9.4.2 Partial Trees . . . . . . . . . . . . . . . . . . . . . . . .. 319
9.4.3 Requirements . . . . . . . . . . . . . . . . . . . . . . . .. 320
9.4.4 Actions . . . . . . . . . . . . . . . . . . . . . . . . . . .. 321
9.4.5 Partial Order . . . . . . . . . . . . . . . . . . . . . . . .. 322
9.4.6 Logical Forms . . . . . . . . . . . . . . . . . . . . . . . .. 322
9.4.7 Computational Rules . . . . . . . . . . . . . . . . . . . .. 323
9.4.8 Update Actions . . . . . . . . . . . . . . . . . . . . . . .. 325
9.4.9 Pragmatic Actions . . . . . . . . . . . . . . . . . . . . .. 325
Bibliography 326
General Index 338
Symbol Index 347
Preface
Knowing a language consists in having a capacity to understand what a speaker
of the language is saying. The present book seeks to develop a formal basis
for this common-sense view, characterizing structural properties of language in
terms of the incremental process of building up interpretation from the sequence
of words. The emphasis is on the process of establishing some structure as inter-
pretation, rather than just specifying the result, which is the structure itself.
This is a departure from orthodox assumptions. In these more established ap-
proaches, the perspective is static { a characterization of strings and some xed
structure and interpretation. Upon this new view, the perspective is dynamic {
about how structures grow from some initial skeletal form of interpretation to
some richly structured object from which a full interpretation can be derived.
The goal that we have set ourselves is to dene a dynamic architecture within
which structures are established as interpretation for an uttered language string
on a left{right basis, and in which explanations of properties of sentences are
couched in terms of how interpretation of a string can be progressively estab-
lished from some starting point to a logical form as outcome representing an
interpretation of the string in context.
The motivation for this shift arose from early work in which two of the three
authors set out to substantiate informal ideas from pragmatics by dening a
formal model of the step-by-step process of understanding a linear sequence
of words in context, re ecting insights of work by Sperber and Wilson and
colleagues (Sperber and Wilson 1986). Arguably the main challenge in the task
we set ourselves was to re ect the way in which hearers can take relatively little
information encoded in the words and enrich it into some selected interpretation.
The idea of goal-directed labelled deduction as a central tool arose, because we
were looking for ways to model not merely what interpretation a string has in a
context but how it gets to be so understood, on the assumption that utterance
interpretation is a left to right inferential processing task.
The labelled deductive system (LDS) methodology of Gabbay (1996) seemed
particularly well suited for this modelling task, using as it does a labelled natural
deduction mode of presentation, modelling the inference task from the inside as
it were, rather than from a more global external perspective. In this early work,
interpretation was seen as built up largely by deduction with words acting as
constraints on that process. The surprise of this work was that time and again,
though our intentions were to address pragmatic problems, the step-wise process
ix
x Preface
of building up interpretation proved fruitful in addressing phenomena which

linguists have classied as syntactic puzzles. However, many aspects of the
structures we were relying on remained informal, and the challenge of providing
detailed specication of the steps involved in interpretation was not taken up.
In particular we were not able to address what it meant to have incomplete and
partial structural information. All we had were itemized building blocks which
then combined together in speciable ways. What we needed was to be able to
say what it means to incrementally establish the building blocks themselves.
The present book reports an experiment in modelling interpretation as a pro-
cess of tree growth, a departure made possible by incorporating Blackburn and
Meyer-Viol's work on modal tree logics (Blackburn and Meyer-Viol 1994, Black-
burn et al. 1996). The endpoint of the interpretation process is a tree structure
corresponding to a logical form, but all preceding stages involve trees which are
incomplete in some way, lacking annotations to nodes, lacking nodes, or lacking
xed tree relations between nodes. The result has been very striking. By taking
concepts of underspecication and the dynamics of their resolution beyond the
semantic/pragmatic areas (where they are familiar) into the syntactic formal-
ism, we have a natural basis for exploring a whole range of puzzles that currently
defy an integrated form of explanation. One after another, phenomena which
have had in other models to be characterized as heterogeneous { merely discrete
phenomena warranting quite dierent forms of analysis { we have been able to
model as minor variants in the options which the tree growth process makes
available. The result is this book. Here we try and convince our readers that
the time has come to shift from the static perspective of formalisms based on
the metaphor of the familiar classical logics to more dynamic formalisms where
the emphasis is on the process and the incremental development of structure.
The consequence of incorporating dynamics into syntax is, we will argue, that
the principles of syntax denitive of natural language systems should no longer,
as is standard, be seen as providing some fully articulated structure as the
input to semantic evaluation of contextually relative specications of content,
as is assumed in all other formalisms. In its place, we suggest, syntax has to be
seen as an articulation of the dynamics involved in progressively building up tree
structures for interpretation. In putting forward this case, we will dene a formal
framework and apply it to a number of outstanding puzzles. We deliberately
chose to consider a range of phenomena across a broad typological range, at the
cost of not dening a complete formal fragment, in order to give a sense of its
general applicability. So we develop typologies for relative clause constructions
encompassing English, Japanese, Arabic, Hebrew, Italian and Romanian, and
for wh questions encompassing English, German, Hungarian, and Arabic. Along
the route, we explore the similarities and dierences between our paradigm and
other more established paradigms, placing the formalism we propose within a
landscape of alternative analyses. And we give a preliminary indication of how to
meet the challenge of expressing quantier construal within the adopted parsing-
oriented perspective. We then tentatively start out on the task of exploring the
consequences for linguistic theory and philosophy of language in general. The
book closes with a chapter that spells out in more detail the assumptions of the
Preface xi
framework set out in chapters 2 and 3.
There is a real sense in which this book could not have been written at
an earlier period. First, it has been developed following a period in which
pragmaticists and semanticists have paid increasing attention to the partial
nature of information available in the interpretation process, and this book is
an extension of that type of work. Secondly, the tools with which to develop
formal accounts of the dynamics of information growth are only now becoming
better understood. While a considerable amount of recent work has been done in
the development of semantic models of information growth, the development of
syntactic systems for modelling such concepts are only now beginning to become
available, the rapid expansion of the development of programs for computers
providing the trigger for a developing interest in the formal specication of
procedures. It is against the computer science background that Gabbay dened
Labelled Deductive Systems, and it is against the joint background of work on
underspecication and the fruitfulness of a modular LDS methodology that the
present framework has been developed.
As we hope this narrative of its birth process has indicated, this book is at
one and the same time an exercise in linguistics, applied logic, and perhaps also
philosophy of language. We hope it is readable by anyone working in any one of
these elds, though the styles of the various readers may vary quite dramatically.
We anticipate that linguists will be best suited by the step-by-step unfolding of
logic formalism developed in tandem with empirical argumentation. A logician,
to the contrary, might want to begin at the end with the denitions, and then
leap about amongst the early chapters. And a philosopher might prefer to
read the early and later chapters, while only dipping more sporadically into the
empirical cross-linguistic argumentation developed in the central parts of the
book. A computer scientist, we would hope, might be inspired to address the
challenge of dening an implementation of this highly modular system. Our
minimal hope is that our book will at least entice people into exploring the
very rich potential of looking at syntactic properties of language in terms of the
dynamics of structural growth.
As an essentially interdisciplinary exercise, this book has been some time in
gestation. In its early form, it comprised joint work by Ruth Kempson and Dov
Gabbay; and the rst steps in articulating a formal framework were taken as
part of an ESRC-funded project (R000232069). Marcelo Finger subsequently
did preliminary work on dening a prototype implementation of this early model
(for which a number of visits to London were made possible by the British
Academy). The decision to look at the potential provided by modal tree logics,
taken upon the arrival of Wilfried Meyer-Viol as part of the team, is what has
led to the ability to express structural concepts of underspecication and subse-
quent update (work which was carried out as part of EPSRC research projects
GR/K67397, GR/K68776). This has led to the detailed articulation of a formal
system making use of tools of Labelled Deduction, the Logic of Finite Trees, and
the Epsilon Calculus. As part of these EPSRC-funded projects, Roger Kibble
revised and very much expanded the earlier Finger implementation to yield a
rst proto-type implementation of the present model. And the nal steps of the
xii Preface
completion of the book were supported by the Leverhulme Foundation's award

to the rst author of a ve-year Leverhulme Research Professorship.
During this period of its growth we have been very grateful to logic and lin-
guistics colleagues for their support, criticism and challenges presented. Amongst
those we have to thank for these various forms of stimulation are: Shalom Lap-
pin, Abbas Benmamoun, Andrew Simpson, Wynn Chao, Ronnie Cann, Johan
van Benthem, Jan van Eijck, Patrick Blackburn, Gisbert Fanselow, Tim Fer-
nando, Marcelo Finger, Rodger Kibble, Uwe Reyle, Hans Kamp, David Swin-
burne, Lutz Marten, Denise Perrett, Masayuki Otsuka, Akiko Kurosawa, Anili
George, Ruy de Quieroz, Antony Hunter, Nigel Vincent, David Adger, Rita
Manzini, Annabel Cormack, Neil Smith, Deirdre Wilson, Robyn Carston, An-
drew Davidson, Stavroula Tsiplakou, Larry Moss, Jonathan Ginzburg, Natasa
Rakic, Ian Ravenscroft, Elisabet Engdahl, Nicolas Nicolov, David Weir, Asl
Goksel, Yan Jiang, Akiko Yoshimura, Ivan Sag, Jacqueline Lecarme, Malcolm
Edwards, Hans-Jurgen Ohlbach, Hans-Martin Gartner, Michael Moortgat, Dick
Oehrle, Glyn Morrill, Samson Abramsky, and the many students who have
been exposed to these ideas over the last three years. Ronnie Cann, Jonathan
Ginzburg and Marcelo Finger have to be thanked in particular for the very
considerable help they gave in commenting upon a pre-nal version.
The preparation of the manuscript has been a community eort to which
so many people have contributed that this often painful process has seemed
almost like fun. In particular we have to thank Jane Spurr and Anna Maros for
their exemplary responsiveness to the challenge of getting camera-ready copy
prepared in a short space of time; and the Blackwell's team of Philip Carpenter,
Steve Smith, Tami Kaplan, Lisa Eaton, Simon Eckley and Jean van Altena for
the pleasure of their distinctive combination of unswerving professionalism and
supportive friendliness at every stage right through from conception to the nal
product. And thanks also to the band of Mike Gabbay, Anili George and Laura
McMillan for carrying out the thankless task of proof-reading with such good
cheer.
Last in the list but without doubt deserving the largest vote of thanks, we
thank our families, parents, and closest friends, for putting up with the intensity
of our commitment to this project.
We are painfully aware that an interdisciplinary endeavour such as this has a
harder time pleasing its deliberately broad audience than books with a more
specialized focus, but we hope you nd it fun.
Ruth Kempson
Wilfried Meyer-Viol
Dov Gabbay
King's College London, July 2000
1
Towards a Syntactic Model

of Interpretation
1.1 Natural Language as a Formal Language?
Knowing a language means being able to segment sounds in that language into
units, recover information from those units, and use this information to work
out what someone communicating in that language has intended to convey. This
statement about what it means to know a language might seem such common
sense as to be banal. In fact, however, it is almost universally rejected by
linguists. Standardly for linguists, knowing a language is not to be explained
directly in terms of having an ability to use it. Rather, the order of explanation
is reversed. First, a characterization of a language is given, then what knowing
a language is, and, nally, how language can be used, given such knowledge. As
will become apparent in due course, this book is to be an implicit plea to return
to the common-sense view. But we start from what is familiar.
According to a view which has dominated linguistic methodology right across
dierent theoretical paradigms since the 1950s, natural languages are taken to
be denable as a type of formal language. A formal language is devised to
describe some object of study, providing the means of giving it a complete
characterization. Such a language is usually presented by itemizing some set of
basic expressions, from which all and only the well-formed expressions of the
language, the sentences, can be induced through recursively applying a set of
syntactic rules. Relative to these rules, semantic rules (if dened) characterize
the interpretation of the sentence-sized objects in terms of the interpretation of
the component parts and their mode of combination.1
1 By no means all formal systems have accompanying formal denitions of semantics. But
a system that lacks formal specication of its semantics may nevertheless be formal in the
sense that the statements that can be expressed within the system are articulated to a level of
detail such that further statements (indeed the complete set of such further statements) can
be provably derived. Lacking such rules of proof, a formal semantics is essential.
1
2 Towards a Syntactic Model of Interpretation
By analogy with formal languages, a natural language may be characterized

as a set of sentence strings generated by a grammar { a nite body of rules
which assigns all and only such strings a structure (and a phonological char-
acterization), relative to which a semantic interpretation is dened. There are
a number of variations on this theme, depending on the form of the syntactic
rules, the detailed form of correspondence between syntactic and semantic rules
for a given system, and so on; but it is none-the-less the standard conception of
natural languages that they are a specialized type of formal language. Accord-
ing to this completely orthodox view (shared at least by categorial grammar
and HPSG paradigms; see Morrill 1994, Pollard and Sag 1994), a language is
characterized independently of any agent knowing or understanding that lan-
guage, as some kind of abstract object: knowing a language must then involve
standing in a certain propositional attitude { that of `knowing' { to some stored
set of rules which comprise the language.
According to a second view which is taken to be in sharp contrast with this
view of language,2 a language does not have any status as an entity independent
of the individual that uses that language { it is exclusively a psychological con-
struct, a body of principles, I-language, which a speaker/hearer has internalized
(and in this sense `knows'). Though this second characterization of language is
self-consciously cognitive, the concept of a linguistic system is essentially sim-
ilar to that of the formal language characterization. This I-Language that the
speaker knows is, as in the formal language concept of grammar, a body of
principles which, in some sense, a speaker has knowledge of { albeit subcon-
scious and unrecoverable. On both concepts of language, the syntax determines
what strings of the language are well-formed, assigning them structures rela-
tive to which semantic interpretation is xed.3 On both of these views, the link
between knowledge of a language and its application to understanding or speak-
ing is obscure: the only direct evidence for the knowledge said to constitute a
grammar of a language is the ability of individual speakers to make judgements
of grammaticality, such judgements being supposedly a direct re ection of this
subconscious knowledge, accessible independently of the additional constraints
and vagaries of performance (see Higginbotham 1987 amongst many others). It
is this body of knowledge { said to be describable independent of any appli-
cation to particular language tasks { which is put to use in performance and
understanding, these being said to be not pertinent to the characterization of
the knowledge of language itself. The articulation of properties associated with
language in use are thus essentially secondary, dened only relative to some
prior articulation of a grammar.
2 As the Chomskian philosophy of language (Chomsky 1986), this cognitive rewrite of the
view that natural languages are formal languages has dominated the eld for nearly half a
century.
3 Note that this is true even in models which have movement, say, to a level of LF as the
interface level between syntax and semantics, for even with such movement processes, the
structures generated are inhabited by expressions of the language, and not by some distinct
representation of interpretation assigned to those expressions. As in other elds of theory
construction, not all grammars incorporate a formal assignment of interpretation to such
syntactic structures: see n.1.
Underspecication in Language Processing 3
In this book we are going to be introducing a model of language which re-
ects the common-sense view of language. We shall be setting out a model of
natural language understanding in which the development of an interpretation
for a string is dened as an incremental left-to-right process of constructing a
logical form representing one possible content attributable to the string { in
short, a form of parser. The process will involve the top-down development of
tree structure representing the logical form. Each node as it is introduced is
inhabited by formulae constructed from the words as these become available on
a left{right basis, where each such formula is a representation of content which
may have been determined in part through a choice from some set of contextu-
ally provided values. The only concept of structure is the sequence of partial
logical forms, in terms of which this process is dened. As a representational
model of interpretation, this framework is one among several (e.g. Jackendo
1992). What is novel about this formal articulation of the parsing process is
that it purports also to provide an explanation of syntactic properties of nat-
ural language. Having set out the framework, what we shall argue is that the
dynamics of how representations of content are incrementally built up provides
a basis for explaining core syntactic phenomena { specically long-distance de-
pendency and a cluster of related phenomena. The result will be a new twist on
the view that the level of representation explanatory of syntactic distribution is
a level of semantic representation.4 The only level of representation advocated
will indeed be that re ecting some assigned interpretation; but that is only the
lesser part of the story, for syntactic explanations will make reference not merely
to semantically interpretable representations but also to the process of building
up such representations on a left{right basis from a sequence of words.
Our demonstration of the fruitfulness of this approach will involve taking
a number of current syntactic puzzles and showing how a solution to them
emerges naturally from the dynamic perspective we have adopted. Finally, we
shall follow up on the consequences of these results, and return to the point
we have started from, closing with a reconsideration of the question of what it
means to know a language.
1.2 Underspecication in Language Processing
Central to our account will be a concern not merely with what interpretation is
projected by a natural language expression, where that interpretation is taken
as some composite entity; but with how the components are successively set
into a conguration following a left{right sequence and then combined. With
this concern uppermost, one particular property of natural languages is imme-
diately apparent. Language processing as a task of establishing interpretation
involves manipulating incomplete objects at every stage except at the very end,
as shown by the set of partial trees schematically displaying a parsing sequence
for (1.1):
4 This was the view of the generative semanticists arguing in the late 1960s and 1970s that
deep structure was the level of semantic representation (see McCawley 1968).
(1.1) John upset Mary.
(a) ? (b) P AST :?
John ? John ?
Upset ?
(c) P AST :? (d) P AST : Upset(Mary )(John)
John ? John Upset(Mary )
Upset Mary Upset Mary

Figure 1.1: The processing of John upset Mary
In every tree representation up to the last, we have a partially specied structure
for a logical form, in which something is represented as being missing. At step
(a), one node is annotated by some logical constant denoting an individual
named John, and two nodes are not annotated at all, one of which will be
subject to further expansion. At (b), a further pair of nodes is introduced,
a representation of the relation Upset is introduced, but there is as yet no
indication of how its second argument is to be saturated, leaving three nodes
as yet to be annotated. At (c), the process of unfolding the tree on a top-down
basis is complete, the terminal nodes have sub-terms of some logical form, but
no interpretation has as yet been assigned to the intermediate nodes. And,
nally, at (d), all nodes have been assigned a representation, the interpretation
of the root node having duly been compiled by combining the various component
parts according to ways directed by the structural conguration. Thus, while
the completed tree represents the nally constructed logical form of the string
and the mode of combination of its parts, it is the sequence of partial trees
which re ects what information is available at each of the intermediate steps.
Furthermore, the tree thus constructed represents the logical form for a string,
representing one possible interpretation: it is not a syntactic tree (or a parse
tree) in the conventional sense of having terminal nodes which are annotated
by the words of a string.
This incompleteness of interpretation at all non-nal stages of an interpre-
tation process is endemic to the task of parsing. The partial projection of in-
terpretation for a string { simply assuming for the moment that interpretation
can legitimately be represented as some sort of annotated tree structure { may
involve, as here, partially annotated nodes: some nodes are in some appropriate
sense complete, others as yet incomplete, or even yet to be constructed.
Underspecication in Language Processing 5
The trees displayed in this sequence, however, by no means exhaust the types
of incomplete specication which natural language expressions contribute to the
build-up of interpretation. On the one hand, this process may also involve taking
as input intrinsically incomplete representations of content which get replaced by
some antecedent representation already established in the context in which the
utterance is being interpreted. This is the phenomenon of anaphora. If we model
this phenomenon by dening anaphoric expressions as projecting the trigger that
leads to some such process of substitution, then the content lexically projected
by such expressions will have to be systematically weaker than any selected
interpretation, being rather an indication of the type of substitution which will
lead to some appropriate interpretation being established. On the other hand,
there may be trees in which individual parts are introduced without specication
of how they are related to each other. This phenomenon, we shall argue, is a
major part of what has come to be known as the left-periphery problem (see
Rizzi 1990), whereby structure associated with left-peripheral constituents in
a sentential string may stand in a range of relatively underspecied relations
to the structure projected from expressions immediately subsequent to them
in the string, at least as initially describable in the parsing process. And this
is by no means the end of the list of possible forms of underspecication. A
major consideration, later in the book, will be the articulation of such forms
of underspecication and how the processes that update these various forms
interact.
The need for partial descriptions of tree structure is well recognized in the
parsing literature since Marcus (1980) and Marcus et al. (1983), but the trees so
described dier from those of the present enterprise in a number of ways. First,
standardly, a tree is annotated by words in a string. Secondly, the tree descrip-
tions manipulated in parsing formalisms are dened relative to an independent
grammar formalism which denes well-formedness solely in terms of assignment
of hierarchical structure to the string. Trees, that is, may be developed into
new trees by rules of the grammar, but any such process is invariably dened
over (sub)trees which are complete in themselves: there is no concept of a tree
in which not all tree relations are specied { this is indeed a distinguishing
property of parser formalisms. What will be new in what we introduce is the
extension of incomplete specications from semantics and pragmatics, where
they have become familiar through work on anaphora in particular,5 to the do-
main of syntax. Unlike other models, we shall be taking the dynamics of tree
growth on a left-to-right basis as the core of an account of intrinsic properties
of natural language structure.
On the view we shall put forward, the build-up of interpretation for a string
will be projected as a sequence of successively richer descriptions of a logical form
represented as a tree structure. The starting point will be a description of a tree
as consisting of only a root node, this node decorated merely with a statement
5 See the postulation of discourse referents in Discourse Representation Theory (DRT) and
its variant UDRS (Kamp and Reyle 1993, Reyle 1993, 1996) set up to articulate the process
of establishing a DRS, and the postulation of infons, situations, constraints and parameters
in Situation Semantics (Barwise and Perry 1983, Cooper 1996).
of the goal { to derive a propositional formula as decoration for that node on

the basis of the tree yet to be introduced. The endpoint will be a completely
annotated tree structure representing a particular interpretation assigned to that
string. The type of example round which much of the subsequent discussion will
revolve is that of left-dislocated expressions as in (1.2):
(1.2) John, Mary is sure that Sue admires.
From a parsing perspective, examples such as this present a particular form of
underspecication, in that a language processor, in representing the contribu-
tion to the interpretation of the whole made by the left-dislocated expression,
is not initially able to identify the role which that expression plays in some
propositional structure to be established. Re ecting this informal observation,
we will take left-dislocated expressions to annotate a node in the emerging tree
which is not initially assigned a xed tree position, an underspecication which
has subsequently to be resolved. This incorporation into the description of nat-
ural language structure of underspecied tree relations and their subsequent
resolution will turn out to provide a natural basis for explaining the cluster
of phenomena associated with long-distance dependency eects across the full
array of language types. Such long-distance dependency phenomena are agreed
by all to be central to any explanation of the syntactic properties of natural
language.6 Yet, despite extensive study of long-distance dependency over many
years, its characterization has remained surprisingly heterogeneous in all the
major frameworks, requiring, as we shall see in due course, a number of dier-
ent statements with discrete forms of explanation. These include strong versus
weak crossover, resumptive pronoun versus gaps, partial movement, wh-in-situ,
and expletive wh, etc. In this new, explicitly dynamic perspective, these phe-
nomena will be seen as falling together in a principled way, while nevertheless
allowing systematic cross-language variation.
The sequence of chapters is as follows. In chapters 2 and 3 we introduce the
basic framework, with rules for building tree representations providing a logical
form for simple clausal sequences, and for long-distance dependency structures.
In chapter 4 we give a characterization of the process of building up inter-
pretation for relative clause sequences, with a preliminary indication of how
a cross-language typology for these structures might be developed in terms of
how information may be copied from one tree structure to another. Setting
out this typology will involve surveying restrictive and non-restrictive relatives,
head-initial, head-nal and even head-internal relatives. In chapter 5 we look at
the range of wh questions across languages in the light of the dynamic perspec-
tive we have been setting out. In chapter 6 we look at the interaction between
the processes involved in establishing the interpretation of left-dislocated wh
expressions and anaphoric expressions { that is, the phenomenon of crossover.
What we shall see in these analyses of relative clauses and wh questions is that
6 Recall, back in the 1960s, that it was long-distance dependency phenomena, with their
associated island constraints, which it was granted could not be explained in purely denota-
tional terms, with the consequent admission that there were at least some syntactic phenomena
whose explanation was irreducibly syntactic (see Partee 1976).
The Representational Theory of Mind 7
natural typologies emerge as individual language systems are analysed as mak-
ing available dierent underspecied descriptions, with consequent variation in
the dynamics of their resolution. In chapter 7 we outline ways in which quanti-
cation might be re ected in this system. In chapter 8 we step back to consider
the general issues raised by this dynamic view. We explore the consequence of
this model { a variant of the common-sense view of language { that a grammar
of a language provides the architecture within which parse sequences leading to
an interpretation can be dened. Indeed, on this view, a string is well-formed
if, using all its words in a left-to-right order, it creates at least one logical form.
The syntax of natural languages will thus be characterized in terms of incre-
mental building up of structures representing interpretation. The nal chapter
then provides a more explicit statement of the rules and assumptions set out in
chapters 2 and 3.
1.3 The Representational Theory of Mind
The point of departure for this account is the assumption that the process
of understanding a natural language string should be dened as a process of
incrementally constructing representations. Relatively uncontentious though
this assumption may seem, it nevertheless is founded upon some variant of a
Representational Theory of Mind, which is taken by many to be controversial
(see J.A. Fodor 1981, 1983, 1998, Silvers 1989, Dennett 1987). According to
this type of view, assigning interpretation to a signal by a cognitive system nec-
essarily involves pairing the signal with some object (or more abstractly some
denotatum of the signal) via the construction of a representation of that ob-
ject/denotatum. Very approximately, we see a pigeon ahead of us ying across
the car, if the given stimulus to the retina causes some cognition-internal rep-
resentation to be set up which we take to denote that pigeon. The information
that can then be derived from such stimuli lies not in the objects themselves,
nor in the stimuli, but in the drawing of inferences over such internally con-
structed representations. As Fodor (1998) expressed it, concepts are essentially
mind-dependent. Despite the many controversies surrounding Fodor's particu-
lar variant of cognitive psychology, in particular over the semantic properties
such a system must have to substantiate what it means to be a representation
(see Dennett 1993, Fodor and Lepore 1992), the Representational Theory of
Mind is quite generally assumed as a working methodology for cognitive sci-
ence. As both Fodor and Dennett have put it from opposing stances within
the family of representational theories of mind, there does not seem to be any
serious alternative, if we are to make any progress in cognitive science. Natural
language systems, according to Fodor's own representationalist theory, dene a
mapping from some relatively low-level representation of the incoming stimulus
onto some internal representation within a formal system, over which inference
can be syntactically dened, this system of representations being discrete from
the natural language itself. It is in this sense that a natural language is what
Fodor (1983) calls an input system.
A property of all natural language systems that Fodor fails to stress is that
there is a systematic gap between information integral to the natural language

expression and the interpretation that is attributed to that expression in a par-
ticular context. This is the well-known phenomenon of context dependence,
displayed by anaphora, tense, nominal construal and ellipsis, in which the par-
ticular interpretation of the expression or structure depends on its processing
within the given context:
(1.3) He ignored her.
(1.4) Would you like a red apple? Or would you prefer a green one?
(1.5) I'm going to go home soon. Are you going to, too?
Understanding each of these sentences involves an asymmetry between the in-
formation projected by the expressions themselves, sui generis, and their in-
terpretation in the context in which they occur. The enrichment of the ini-
tial specication to yield some context-specic content has been argued to be
constrained by quite general pragmatic considerations such as presumption of
relevance (Sperber and Wilson 1986, 1995), and so not a task achieved solely
through computations internal to the language input system.7 As we shall see in
due course, there are various aspects of interpretation which may be incomplete
at non-nal stages of the parsing process; and the challenge of modelling the task
of building up interpretation for natural language strings is to provide a general
characterization of the process involved in updating incomplete specications of
interpretation. Preserving the general spirit of the Fodor/Sperber/Wilson Rep-
resentational Theory of Mind, we shall model all updating of representations
as replacement of one description of a partial structure by some description of
a richer structure, so that the nal representation of content is the result of
progressive updates of structure from some starting point. Seen in this light,
anaphora resolution and ellipsis construal will be seen as structural processes,
which induce a copy procedure from some antecedent representation of content.8
In adopting this stance, and moreover anticipating that it might provide
a basis for structural properties of natural language systems, the account to
be provided will be a departure from current orthodoxies about the nature of
natural language interpretation. In particular it involves a departure from the
assumption that natural language systems are formal languages whose syntactic
properties are dened over strings of words (= sentences) with semantic inter-
pretation dened relative to their syntactic structure. Furthermore, we shall be
arguing that the representations resulting from anaphora construal will be the
result of interaction between the system-internal operations and quite general
pragmatic operations, so that the grammar will systematically underdetermine
the full logical forms.
7 Sperber and Wilson argue, in the face of Fodor's scepticism about the possibility of there
ever being a theory of central processing, that interpretation of signals is constrained by a
balancing of cognitive eort and inferential eect { a constraint they advocate as the content
of the concept of relevance.
8 See Kamp 1996, which argues for the same conclusion on rather dierent grounds.
Pronominal Anaphora: Semantic Problems 9
1.4 Pronominal Anaphora: Semantic Problems
It might be argued that any shift to a representational account of the phe-
nomenon of context dependence is a retreat to a purely formal solution that
does not provide a semantic explanation. In the past couple of decades, work
has been done in particular in Dynamic Semantics, Situation Semantics and
Discourse Representation Theory (DRT) on characterizing the relation between
an anaphoric expression and its denotational content as part of a general charac-
terization of the context relativity of natural language interpretation (see Kamp
1984, Kamp and Reyle 1993, Groenendijk and Stokhof 1991, Barwise and Perry
1983, and many others). These paradigms vary as to the importance of the par-
tiality integral to the intrinsic content of an anaphoric expression.9 However,
all these various approaches to the problem of anaphora retain a commitment
to characterizing the construal of anaphoric expressions as a property of seman-
tic interpretation, an account which is articulated relative to an independent
syntax, within which structural properties of anaphoric expressions are charac-
terized. It should not go unnoticed that this commitment preserves the natural
language-as-formal-language methodology, with its assumption that syntax {
dened independently of semantics { determines the well-formedness of strings
and all their structural properties, leaving the interpretation of the strings to
be dened in the light of some such syntactic characterization. Even in DRT, a
discourse representation structure itself is a representation of a partial model,
and not a level over which syntactic generalizations are dened.
In what follows, we shall be articulating in some detail a representational-
ist stance on interpretation and, as part of that, a substitutional account of
anaphora construal. In the remainder of this chapter, therefore, we provide
preliminary evidence for this position, by setting out the problems which face
accounts of anaphora in terms of some direct attribution of denotational content
to the natural language expression.10
1.4.1 The Problem of Multiple Ambiguity
The rst problem that arises is that pronouns, being interpreted relative to
some contextually determined value, will have to be assigned discrete denota-
tional contents according as the context varies, and they will accordingly be
9 The primary focus in Dynamic Semantics, for example, is solely a characterization of the
semantic value of some anaphoric expression relative to what is taken to be a syntactic indexing
of the pronoun { without any semantic characterization of the content of a pronoun qua
pronoun in the absence of such indexing. Situation Semantics, to the contrary, articulates a
whole novel ontology within which a range of partial objects are introduced { situations, infons,
parametrized objects, etc. Discourse Representation Theory, preferring a more conservative
semantic ontology, articulates an interim construct, the concept of a discourse representation
structure, which has a hybrid status both as a form of representation and as a partial model
interpretable by embeddability conditions within the total model. In its articulation of an
intermediate structural conguration, and only relative to this there being any model-theoretic
evaluation, Discourse Representation Theory is much closer to the enterprise we are embarked
on here than either Dynamic Semantics or Situation Semantics (see Dekker 1996, Kamp 1996).
10 This section is largely taken from Kempson et al. 1998.
characterized as multiply ambiguous. The challenge, then, for a semantic ac-

count of anaphora is whether such an account can provide a unitary base from
which to explain the range of interpretations available to a single lexical form.
We claim that it cannot. Though pronominal anaphora has been the focus of
attention over a considerable period, initiated by Evans (1980), Heim (1982) and
Kamp (1984), it is universally assumed that it cannot be reduced to a unitary
phenomenon. It is seen as dividing into at least two, if not three, discrete seman-
tic phenomena according as the pronoun is construed as a bound variable, as
an indexical referring to some xed individual, or as a so-called E-type pronoun
denoting a witness set of some quantied assertion provided by the antecedent
sentence (solutions vary as to how many natural classes are required { see Heim
1990, Groenendijk and Stokhof 1991, Chierchia 1991, Lappin and Francez 1994,
Kadmon 1990, van der Does 1996, Neale 1990).11 With these three types (pos-
sibly reducible to two), the various forms of intra- and cross-sentential forms
of co-reference as in (1.6){(1.10) can be characterized. However, in addition,
anaphoric expressions may have an interpretation established through inference
without any direct matching of content, a phenomenon which does not corre-
spond straightforwardly to any of the three categories. The range of types is
illustrated in (1.6){(1.12). (1.6) is an example of a bound variable construal,
(1.7) an indexical construal, (1.8) an E-type construal; (1.9) and (1.10) involve
intra-sentential co-reference and cross-sentential co-reference; (1.11) is an exam-
ple of cross-sentential anaphora involving intermediate steps of inference, and
the construal of (1.12) involves what has been called bridging cross-reference
(Clark and Haviland 1977, Clark 1977, Garrod and Sanford 1981):
(1.6) Every girl worries that she might get pregnant.
(1.7) She is pregnant.
(1.8) Most girls passed with distinction. They had worked very hard.
(1.9) Sue is worried that she might get pregnant.
(1.10) Sue came in. She was pregnant.
(1.11) She has either rented a car or she's taken a van. She has to return it
on Monday.
(1.12) The Smiths are nice. He is a doctor.
What is less often pointed out is that these phenomena are replicated right across
dierent forms of anaphoric construal. (1.13){(1.18) are examples displaying the
11 The force of Dynamic Predicate Logic (DPL) has been to dene a concept of variable
binding which allows some cases of E-type anaphora to fall together with more regular cases
(see Groenendijk and Stokhof 1991 and other references). However, this extended concept of
variable binding (involving pairs of variables) is not applicable to all cases, leaving a remain-
der unaccounted for (see van der Does 1996, Kibble 1997). Even DRT, the pioneer in the
systematic representation of context dependence and the ability thereby to unify apparently
dierent types of anaphora, presumes a quite dierent characterization for indexical pronouns
(see Kamp and Reyle 1993).
same array of anaphoric eects for denite NPs (in the same relative sequence),
and (1.19){(1.23) and (1.24){(1.28) exemplify the same phenomena as displayed
by the two forms of demonstrative NP (the same putative ambiguity problem
also extends to tense construal { see van Eijck and Kamp 1997, Kempson et al.
1998):
(1.13) Every house I have put on the market I have checked, in order to make
sure that the house won't be hard to sell.
(1.14) The idiot was sick.
(1.15) Most students were there. The entire group got drunk.
(1.16) John came in. The idiot had drunk too much.
(1.17) You can have a car or you can have a van, but the vehicle must be
returned on Monday.
(1.18) The Smiths are nice. The husband is a doctor.
(1.19) Every day I drink a Coke, I know that later that day I shall have a
migraine.
(1.20) That idiot is sick.
(1.21) Most people who came early left well before a few people got drunk.
That group were no problem.
(1.22) John came in with his partner. That idiot had drunk too much.
(1.23) The Smiths are not as nice as the Jones. That marriage has been a
disaster.
(1.24) Every time I don't take my pills, I think that this time I am better
and will not need them.
(1.25) This idiot is sick.
(1.26) Most people who came early left well before a few people got drunk.
These drunkards unfortunately spoiled the occasion.
(1.27) John came in with his partner. This new guy is a disaster.
(1.28) The Smiths are not as nice as the Jones were. This new couple are
never around.
Unless these multiple forms can be explained in terms of some unitary semantic
basis, discrete specications will have to be advocated for every pronoun, every
demonstrative determiner, and indeed every tense form. Yet, with an anaphoric
expression characterized as a set of discrete lexical items, one would antici-
pate parametric variation across languages according as one pronoun projects a
bound variable interpretation, a discrete pronoun an indexical interpretation, a
third an E-type interpretation. And similarly demonstrative pronouns and tense

forms. Yet no such systematic variation exists.12 The generality of this phe-
nomenon suggests that an optimal solution will involve a unitary base for such
lexical anaphora, explaining the semantically divergent forms by some theory of
how weakly specied lexical items can give rise to this array of interpretations.
For a framework in which lexical content is specied as some xed denotation
(whether as a variable, a witness set or an individual uniquely referred to), this
array of interpretations remains problematic.
1.4.2 The Problem of Uniqueness
Turning now to the phenomenon of discourse co-reference, and setting on one
side the problem of multiple ambiguity, context dependence is seen quite gen-
erally as requiring a semantic characterization in the form of an update mecha-
nism. Sentences are not evaluated for truth in isolation, but rather with respect
to a context. Evaluation of a sentence within such a context then results in
the creation of a new context.13 With the shift of focus from a truth-based
conception of content to an evaluation in terms of context change, the claim is
that the denotational content of an expression can be dened in a unitary way
as its context change potential, and that it is this shift that provides the needed
exibility in characterizing the intrinsic content of an expression. In several
variants of this form of analysis, discourse-based anaphora is analysed in terms
of a Russellian denite description with entailment of uniqueness (see Neale
1990, Heim 1990, Cooper 1996, van der Does 1996, Milward 1995, F. Breheny
1999), the anaphoric expression being dened as a context-update mechanism
mapping some context containing a unique individual (hence the presupposed
implication of existence) onto a new context, with the predication said to hold
of that individual.
Such accounts however face the well-recognized problem that many examples
fail to full the condition of uniqueness while nevertheless constituting successful
instances of pronominal construal. A pronoun (or denite NP) can be success-
fully used despite there being more than one individual in the context which
could provide a possible value for the anaphoric expression, as in (1.29) said to
refer to a coat which the speaker is just taking o in a changing room:
(1.29) I like it, but it's too similar to this other one.
The response by Kadmon (1990) that the situation of coat liking is not the
same as that in which the two coats are similar is unattractive, for the concept
12 The only putative exception we know of occurs in Malagasy, which is argued by Zribi-Hertz
and Mbolatianavalona (1999) to contain a pronoun restricted to bound variable interpreta-
tions.
13 In Dynamic Predicate Logic and Stalnaker variants of interpretation, in which the context
is a set of world-assignment pairs, the sentence is an update function which acts as a pro-
gressive ltering on such sets (see Stalnaker 1974, 1984, Heim 1982, Groenendijk and Stokhof
1991). However in Situation Semantics and DRT, with inference being dened over situa-
tions/discourse representations, the situation/discourse representation becomes part of the
context, though DRT can also be interpreted in Stalnaker terms, as pointed out by Fernando
(1994), van Eijck and Kamp (1997), and Kamp (1996).
of situation supporting the sentence appears to have been narrowed down to
just the entities of which the relation is said to hold, and not to any broader
partial model relative to which the assertion is evaluated.14
Also problematic are examples such as (1.30) said in a discourse situation
in which there are15 two boys and two dogs, with just one boy holding one of the
dogs too tightly:
(1.30) He's holding it too tightly.
How can a pronoun in such circumstances succeed in picking out the dog in ques-
tion, given the single scenario? Given that there are two male individuals and
two further non-human individuals, the assertion made is incorrectly predicted
to be false. A move that is made in Situation Semantics, for example by Mil-
ward (1995), is to suggest that a situation must allow objects to carry indices,
thus enabling the uniqueness relation between the linguistic expression and the
semantic object which constitutes its referent to be retained. But this move
involves treating individuals in situations as syntactic objects, discretely named
by the index assigned to them, with a notable shift to a more representational
perspective.
1.4.3 The Problem of Indirect Reference
Suppose we also set aside this problem, there remains yet the problem of indirect
reference, but here the context may have to be broader than is provided by the
previous sentence. (1.31) displays the problem:
(1.31) John had a heart attack right outside the hospital, and they refused to
treat him without an insurance card.
The problem is that, if the pronominal is taken to dene an update from a
context containing a xed individual to a new context which incorporates the
assertion made about that individual, there should be a clearly denable sense
in which the rst context contains that individual. But in this respect (1.31) is
problematic. The hospital in the situation corresponding to the content of the
rst sentence of (1.31) is a building. Yet it is apparently suÆcient to license the
use of the pronoun they. Milward's (1995) proposal within situation-theoretic
assumptions is to assume that situations relative to which sentences are eval-
uated contain not only the entities depicted, possibly carrying distinguishing
14 Perhaps the most discussed case is that of the two bishops:
(i) If a bishop meets another bishop, he blesses him.
where to be true, both bishops must bless each other (see Kadmon 1990, Chierchia 1991, Heim
1990, Milward 1995).
15 The example is discussed by Milward (1995) as a problem for the uniqueness implication
of the denite NP; but the problem carries over directly to a pronominal variant, unless such
a use is said to involve rigid reference to some individual demonstratively picked out { a move
which side-steps the uniqueness problem only to re-confront the ambiguity problem. Such
cases have been used to motivate a discrete concept of resource situation (see Cooper 1996).
But, as with Milward's suggestion of indexed objects, the level of abstraction is such that the
constructs posited are arguably nothing more than a representation of constraints to be met
by some semantic object, rather than themselves constituting a semantic object.
indices, but also relations dened on those entities, the situation providing the
interpretation for the sentence John had a heart attack right outside the hospital
accordingly, having not only the entities John, a heart attack and a (unique)
witness satisfying the expression the hospital, but also the relations between
hospitals, doctors and medical conditions needing treatment, on the one hand,
and between heart attacks and medical conditions needing treatment, on the
other. But if such relations are dened as part of the situation satisfying John
had a heart attack right outside the hospital, then the situation becomes quite
unlike the very narrow concept of situation required to yield the right result for
the earlier coat-choosing case.
Evidence that no extension of the notion of a situation to incorporate entities
is suÆcient for a complete characterization of anaphora is presented by cases in
which the situation described may not include an individual, and yet reference
to it by a pronoun can be successful. Consider (1.32) uttered to a friend who
has just split up from his partner and is paranoid about meeting her at every
turn:
(1.32) I'm having a party. Don't worry: she won't be there.
The update made available by the rst sentence (ignoring problems of modality)
is some party. The only set of individuals that can in any sense be said to be
specied as part of the context is some set as witness to the expression a party,
or in some circumstances its complement set.16 But even allowing that the
complement of a witness set might exceptionally form part of a specication
of context, this would not alleviate the problem that this is not suÆcient to
guarantee the unique individuation of the person successfully referred to by
the use of the pronoun she as part of the context relative to which it denes
an update. What is required here is a concept of inference dened over the
information provided by the utterance of the antecedent sentence { in this case
along the lines Ìf A is having a party, C will be there. A is having a party.
Therefore C will be there.' But the need to construct premises from which
the implication of the presence of C at the party is derived is driven by the
presence of the pronoun itself. The specic entities depicted in the situation are
in no sense available in the context within which the sentence is uttered. So the
concept of context is quite unlike the content-independent context presumed in
a semantics in terms of context-updates.17
16 Note the use of the complement set as the basis for establishing anaphor-antecedent rela-
tions in negative sentences such as (i), an observation noted by Corblin (1996), and developed
by Kibble (1997):
(i) They didn't invite a professor. She would have spoiled the party.
17 This is the problem of presupposition accommodation (see Beaver 1997 for an overview).
But notice that in this case accommodation must not have the eect of updating some party
context to include the designated individual. Formal constructive theories of context have
been proposed by van der Does (1996) and Rooth (1992, 1996).
1.4.4 Quantication
The nal problem for the semantic account of anaphora is the use of quantied
expressions which enable an antecedent to be made available for construal of
a following pronoun that is not part of the interpretation of the quantied
expression itself. In (1.33) two situations are described disjunctively, and the
pronominal her then refers to some individual indicated in whichever of the two
situations holds:
(1.33) Either we can get a girl to read a poem, or we can get an older woman
to rattle o a passage from Yeats. If it takes her 10 minutes,...
In such cases, one might argue, a pair of situations is introduced as a context
of evaluation (as in Milward 1995), with no means of choosing between them.18
What is needed is a step of inference derivable from the two situations severally
described to a third situation containing someone who reads for whoever ùs'
is. But no such individual is available in the situation corresponding to the
interpretation of the sentence. To side-step such problems, Milward suggests
a concept of meta-level disjunction, without any ability to distinguish between
the disjuncts. He posits an arbitrary name mechanism, the name in some sense
picking out either one of the two individuals, both being ìn attention'. But this
is to directly incorporate proof-theoretic, and hence syntactic, constructs into
the semantic characterization. It is simply the rule of Disjunction Elimination
of natural deduction systems in a dierent guise (see Prawitz 1965, Lemmon
1965).
1.4.5 Syntactic Processes of Anaphora
Notwithstanding all these possible extensions to the semantic vocabulary, some
cases of pronominal anaphora remain intransigent. Anaphora may be integrated
into the syntactic system through the use of resumptive pronouns in both topic
and focus constructions and in relative clauses, and all such phenomena are
standardly taken to be unrelated to the central cases of anaphora. In certain
circumstances, a pronoun may function as a gap, with the pronoun appearing to
be subject to the very same subjacency constraints as syntactic gaps (Swedish
{ Maling and Zaenen 1981):
(1.34) *Lisa talar jag med pojken som kysst (henne) [Swedish]
Lisa talked I with the boy who kissed (her)
Às for Lisa, I talked with the boy who kissed her.'
18 These cases are highly problematic for DRT also, which imposes the requirement of intro-
ducing a new discourse referent for each indenite, thereby guaranteeing that there is not a
uniquely determined antecedent for the pronominal. Moreover, the embeddability conditions
for the conditional which follows require that all partial models in which the antecedent holds
must be extendable to that of the consequent. But the individual to be constructed in that
subsequent DRS cannot be identied as some girl or some woman as previously set up, be-
cause no such entity is introduced at the top level of the DRS constructed from the previous
sentence. The analogous problem applies to Dynamic Predicate Logic, for which they are
equally problematic.
In others, they serve as a means to enable gap devices to be available in those

circumstances in which a syntactic gap is precluded (Shlonsky 1993, Aoun and
Choueiri (1997), Aoun and Benmamoun 1998):
(1.35) qara?tu ll-maqaalat llaÆii saafara S-
p
Saabu
llaÆii
read-I the article that travelled the young man that
kataba-ha [Classical Arabic]
wrote-it
Ì read the article that the young man who wrote it travelled.'
This type of phenomenon is not generally taken to be a problem given the
syntax{semantics divide, but it should be noted that no semantically based
explanation of anaphora purporting to provide a uniform explanation will pro-
vide an account for this type of anaphora, for the licensing of this form of
pronominal construal appears to be determined by uncontentiously syntactic
properties, and hence cannot be characterized in purely denotational terms.
Yet this phenomenon runs right across dierent language families, suggesting
a systematic interaction between certain types of syntactic processes and the
general anaphora phenomenon, a puzzle which singularly lacks explanation if
the two sets of phenomena are taken to be unrelated.
1.5 The Anaphora Solution { Towards a
Representational Account
Though no one of these arguments taken on its own is a knock-down argu-
ment against characterizing an anaphoric expression in terms of its denotational
properties { after all, ambiguity of linguistic expressions can always be invoked
{ the cumulative evidence leaves wide open the challenge of providing an ac-
count of anaphora which captures the phenomenon in an integrative way. In
what follows, we shall be articulating an account of interpretation which is set
within a representationalist methodology. We shall assume that linguistic input
projects instructions for building up logical forms, and that these instructions, in
combination with information independently established in context, yield some
completed propositional structure. Any one such logical form is then combined
with some larger set of premises, the context, to derive context-specic inferences
(Sperber and Wilson 1995). This overall set of premises represents the full prag-
matic content of that utterance (including so-called contextual implications), a
set which then forms (part of) the context for the following utterance.19 On
this view, the computational/lexical system of a natural language systematically
underdetermines the set of possible interpretations for any sentence, and it is
only interaction between pragmatic and system-internal actions which yields a
context-particular representation of content as articulated in some logical form.
19 Our focus in this book will be exclusively on the process of logical form construction, not
on the deduction of contextual implications.
The Anaphora Solution { Towards a Representational Account 17
Against this general background, the diversity in pronominal interpretation
can be modelled by articulating pronouns as projecting representations with
only a skeletal indication of an interpretation { merely that of a place-holder to
be substituted by some proper representation. The only structural restriction
on this substitution process is one of locality; and antecedent representations
respecting the specied locality restriction may be selected from a wide vari-
ety of sources { other terms in the same conguration (cf. (1.6), (1.7), (1.9)
above), terms constructed during the processing of some previous utterance (cf.
(1.8), (1.10)), terms representing objects in the utterance situation (1.30), or
even terms derived by some context-specic inference (cf. (1.31), (1.32)). The
manipulation of steps of inference to derive a term to take as antecedent (1.33)
is not in principle problematic, given a deductive perspective on interpretation.
Re ecting the supposedly syntactic nature of the interpretation task, there is no
attempt, on this view, to provide a characterization of anaphora resolution in
denotational terms, for semantic evaluation is dened not on the linguistic string
itself, but on the logical form which results from the interpretation process of
which anaphora resolution is a part.
All such statements about the interpretation task are of course nothing more
than a promissory note without a formal characterization of both the system
of representations in terms of which such interpretations are expressed and the
process of building up such representations. Indeed, one of the purposes of this
book is to articulate a formal model of the process of interpretation to give sub-
stance to some of these claims. It remains to be seen, when such a framework
has been articulated, whether construals of anaphoric expressions which involve
syntactic processes can be described in the same terms as what we might broadly
refer to as discourse anaphora, and this question is taken up in chapters 4 and
6. However, whatever the answer to that question, the evidence of underspeci-
cation intrinsic to natural language expressions and the apparently structural
nature of the enrichment process that resolves this underspecication suggest
that the enterprise of setting out a model of natural language understanding in
representationalist terms is well worth exploring in some detail. We now turn
to the task of articulating such a model.
2
The General Framework

Our aim is to dene a formal model of the left-to-right process of natural
language understanding. The primary purpose of this model is to character-
ize how the information associated with an initial specication given by the
rst expression in a string grows to an interpretation of the whole. So we shall
dene at each step what incomplete specication of interpretation has been
built up at that point and what additional requirements remain to be fullled
in order to complete the interpretation process. Central to this task is to re ect
the way the intrinsic content of expressions in natural language characteristi-
cally underspecies their construal in context. We now adopt as background
the following set of assumptions, endorsing the representationalist perspective
sketched in chapter 1.
(i) A model of natural language understanding involves the articulation of
a structured representation, hence a mapping from linguistic expressions
onto representations of content. (It is not a mapping dened from linguis-
tic form direct onto model-theoretic denotations).
(ii) These representations of content are expressions in a system discrete from
the natural language itself, for which a syntactic denition of inference is
provided.
(iii) The processing task is goal-driven and incremental. For each string, the
goal is to establish some propositional formula representing its meaning,
using each item of information as projected by the words in sequence.
(iv) Lexical specications of content are to be in a form that provides suitable
input to this building process.
(v) This rule-driven process interacts with choice mechanisms (re ecting the
fact that the properties intrinsic to individual expressions only partly de-
termine the interpretation assigned to them in any particular context of
use). These choice processes will be dened as the completion of a partial
18
A Preliminary Sketch 19
specication of some structure or formula by an independently accessible
representation.
As these ve assumptions demonstrate, the commitment is resolutely repre-
sentationalist, and the model will duly be dened in terms of how structured
representations are progressively established.
Having set out this statement of what we are going to be modelling, we must
swiftly add a codicil. What we shall not be modelling is how a parser succeeds
in making choices between dierent possible interpretations of a string or in
pragmatically establishing indirect inferences of various sorts. This, we assume,
depends on a broad range of factors, by no means exclusively structural, con-
strained by overarching considerations of relevance. What we are assuming in
putting such relevance-based considerations aside is that the modelling of nat-
ural language understanding presents two separate problems. The rst is one
of dening on a left-to-right basis how a structural representation of interpre-
tation for a string can grow { setting out, in other words, a formal architecture
for describing the parsing process. The second is to take such a framework and
articulate within it criteria which enable a parser to make choices that lead to
appropriate interpretations of a given string with the further pragmatic eects
that such choices give rise to. Our concern in this book is only with the rst of
these: we will articulate a formal model of interpretation growth for a sequence
of expressions, dening possible interpretations for strings of a language on a
left-to-right basis.
To give substance to the assumptions we have set out, a number of early de-
cisions have to be made. We have to choose some logical language for expressing
the representations which re ect the content assigned to any given sentence. We
have to decide how to induce the structure, relative to which the rules of com-
bination for such formulae can be dened. And we have to be able to dene
the ways in which structures can be incomplete, with an appropriate process of
completion.
2.1 A Preliminary Sketch
The deductive system we need will be a composite one { dened in terms of
sub-modules and their interaction. We shall introduce it step by step. First
we shall sketch out the various sub-modules and the dynamics of the system
relatively informally, to give the more casual reader an overview. Then we shall
return to give more detailed denitions and exemplication.
The overall intuition we are modelling is that interpretation is built up from
left to right as each individual word is processed, following the combinatorial
properties of the words as specied by their logical type. Words are assumed
to project expressions in some logical language, and it is these that combine
together to result in a logical form corresponding to the interpretation of the
sentence. This much is familiar, like many categorial grammar systems in par-
ticular. However, we do not assume (as in Steedman 1996, Milward 1995) that
20 The General Framework
(a)
John
(b)
John
xy [Admire(x)(y )]
(c)
John
xy [Admire(x)(y )] Mary
Figure 2.1: Development of an unreduced lambda term
this left-to-right process directly projects model-theoretic content. Rather, we

assume a two-step process.
First, a tree structure is induced from the linear sequence of words, with
words introducing tree structure and lambda terms as annotations for nodes.
Figure 2.1 shows the series of trees that are constructed for a simple example
such as (2.1):
(2.1) John admires Mary.
The words of the sentence can be seen as transitions between partial decorated
trees:
( ) John 7 ! (b) Mary
7 ! (a) admires 7 ! (c):
Note the notational device: all terms decorating the tree are indicated by an
initial capital, e.g. Man, whereas the words of the language are indicated by
the lack of a capital, e.g. man.1 The nal tree structure, (c), corresponds to
the unreduced lambda term
((xy[Admire(x)(y)]; Mary); John)
1 The one exception is proper names, which are always written with an initial capital letter
as both words and as logical constants annotating nodes in the projected structure.
Now, secondly, with the nal tree re ecting the interpretation of the sen-
tence, the non-terminal nodes are consecutively annotated { with `reductions' {
through a bottom-up process, (d){(e) in gure 2.2, that leads to a propositional
formula as annotation at the root node.
(d)
John y [Admire(Mary )(y )]

(e) Admire(Mary )(John)
Figure 2.2: Evaluation of the term of gure 2.1

The heart of this process is the concept of goal-directed tree growth. At the
outset of the construction of a logical representation of a natural language string,
all that is available is a specication of the required outcome: namely, the
construction of some propositional structure. In more familiar terms, what is
required is the construction of a representation of type t. In the course of the
parsing process, these requirements become progressively more specic. For
instance, once the word John has been processed and assigned subject position,
what is required is a predicate phrase, in other words, a phrase of type e ! t.
This process is exemplied in Figure 2.3, where requirements for types at a node
are represented by that type preceded by a question mark.
With each additional word the induced tree becomes progressively richer. At
all but the last stage, the tree is incomplete, or partial, in one or more aspects:
A tree node may be specied at a xed position but lack an annotation
and be decorated merely by a requirement. In this case, a location is
created for a sub-term of a certain type (i.e. with certain combinatorial
properties), where the term itself has as yet to be supplied. This is the
well-known phenomenon of sub-categorization, but we shall generalize it
as the standard way in which nodes in a tree are introduced.
A xed node may have an associated annotation which is itself incomplete.
This is the case, for example, when a pronoun is processed. Re ecting the
structural nature of anaphora construal (see chapter 1), we shall take pro-
nouns and other anaphoric expressions to project a place-holding variable,
?t
John ?(e ! t)
?t
John ?(e ! t)
xy [Admire(x)(y )] ?e
?t
John ?(e ! t)
Figure 2.3: Development of an unreduced lambda term with requirements
to be replaced by an actual term. In all such cases, the interpretation given

by the lexical item itself is incomplete, a mere starting point in the as-
signment of interpretation, that has to be updated by some formula of
appropriate type.
A node may be specied simply as occurring in the tree, its position not yet
fully xed: indeed, this is the account we shall shortly propose in detail
for left-dislocation structures. Such structural underspecication comes
down to the information that a term is a sub-term of the representation
under construction without a specication of its exact relation to that
encompassing term.
All such incomplete characterizations of a node are updated during the construc-
tion process, and the resulting outcome is a tree whose nodes are annotated with
lambda terms, in such a way that the whole decorated tree structure represents
an unreduced lambda term. Evaluating that tree (i.e. normalizing the term)
results in a tree the root node of which is decorated with a propositional formula
representing the reduced term. Evaluation of the tree in this way takes place
once all words of the string have been processed.
To probe a bit more what structural underspecication entails, let us see
what it might mean to characterize a node as annotated but unxed. We take
up the problem of parsing the following sentence from left to right.
(2.2) Mary, John admires.
What information is established after parsing the word Mary if the next word
John is to be construed as subject? All that is known is that some term de-
noted by the proper name Mary is to be a part of the proposition expressed by
the resulting logical form. Supposing that, as before, we represent logical forms
through the annotated tree structure, what has been established in parsing this
rst expression is that a representation of some individual named Mary is pro-
jected as an annotation for a node within that tree { and in consequence a part
of the resulting logical form. But it is not known what position in the tree that
node is to adopt; its structural relation to the nodes in the eventual tree is not
xed. Supposing we model this at present with dotted lines:
?t
Mary
Then, with the words that follow, we get rst a pair of nodes, the rst anno-
tated by a formula projected by John, the second decorated by the requirement
?(e ! t):
?t
Mary John ?(e ! t)

Next we get a second pair of nodes of which the rst is annotated by a formula
projected by admires:
?t
Mary John ?(e ! t)
xy [Admire(x)(y )] ?e
At this point, the string ends; so, according to our goal-directed enterprise of
establishing some propositional formula, the string has supposedly presented
enough information to establish a proposition of type t. With one node that
requires an annotation and an annotated node requiring a location in the tree,
the latter can be taken to supply the annotation for the rst:
?t
John ?(e ! t)
and the tree is duly updated to yield the propositional formula Admire(Mary)
(John) { in other words, the same conguration that results from (2.1):
Admire(Mary )(John)
This string is of course the simplest possible example of a long-distance

dependency; but we shall take the dynamics in this account of long-distance
dependency as symptomatic of syntactic description in general. In this simple
sketch, there is no projection of tree structure dened over the surface sequence
of words: the only structure is that corresponding to the resulting logical for-
mula Admire(Mary)(John). The initial constituent of (2.2) as described from
its `fronted' position, at that early juncture in the left-to-right projection of
an annotated tree lacks a xed position, a lack which is subsequently remedied
to yield a tree identical to that projected from the sequence of words in (2.1).
This underspecied assignment of structure to left-dislocated constituents will
be used in due course to explain an array of syntactic puzzles associated with
long-distance dependency phenomena (chapters 4{6). Generalizing this, we shall
argue that structural properties of natural language expressions are not to be
given in terms of a single xed structure dened over a string. To the contrary,
syntactic phenomena are explained in terms of a sequence of transitions from
some very weak initial specication (matching what is taken to be the parser's
starting point in the interpretation process) to a tree structure establishing a
propositional interpretation, using the individual update specications provided
by the words in sequence. The syntactic analysis of a string thus involves the
concept of growth of increasingly rich tree structures.
In order to dene the type of partial trees which make up such sequences
and the accompanying concept of tree growth, we have to set up a tree descrip-
tion language. The necessity of the shift from an analysis which makes direct
reference to the objects to an analysis which makes essential reference to the
description of those objects can be seen, by analogy, from the description of
partiality in semantic characterizations, where it is more familiar. In semantic
description, the growing recognition of the need for a partial specication of de-
notational content in the case of anaphoric expressions (and in the case of scope
assignment to quantied phrases) has led to the mushrooming of languages for
partial specications (see the use of meta-variables in Alshawi and Crouch 1992,
Alshawi 1996, and Pinkal 1995, the UDRS language for describing lack of scope
assignment within DRT (see Reyle 1993, 1996) and the use of Minimal Recur-
sion Semantics in HPSG (see Copestake et al. 1999)). In extending the concept
of denotational underspecication into the domain of structural explanations,
we have, for similar reasons, to set up a tree description language. In what fol-
lows, then, one of the early steps will be to introduce such a language, so that
we can subsequently talk about how trees `grow'. And central to the interest
in exploring this perspective on natural language structure will be the cluster
of long-distance dependency phenomena, with their initial characterization as
lacking a xed tree relation within the developing tree.
In one sense, there is nothing special about the proposal that the process-
ing of left-dislocation (and more generally long-distance dependency) structures
involves an initial uncertainty characterizable merely as an underspecied tree
relation. Very similar accounts have already been set out within the parsing
literature in several guises. Chief amongst these is that of Marcus (1980) and
subsequent work on D-Tree Grammars (see Marcus et al. 1983, Marcus 1987,
Rogers and Vijay-Shanker 1996). In all these accounts, however, the partiality
of information available in parsing an initial dislocated noun phrase is taken
to be indicative of a larger problem: in parsing an initial noun phrase, there
is always insuÆcient information at the outset to know its relative position in
some resultant tree. On the one hand, the initial noun phrase might simply be
the subject, as in (2.1) above. It might be dislocated as in (2.2). However, it
might also turn out to be part of some indenitely larger NP, as in:
(2.3) John, Sue and Mary left.
(2.4) John's brother's dog's collar has dropped o.
If the parser is to wait until a denitive decision can be made between these
various alternatives, then no characterization of its structural role can be made
at that point. In these parsing accounts, the distinct forms of uncertainty are
both expressed through structural underspecication.
The question that arises is how a theory purporting to be a model of the in-
terpretation process can justify characterizing long-distance dependency eects
in terms of such underspecication, at the expense, apparently, of the ability
to explain the general phenomenon of parsing uncertainty. Should we model
the progressive build-up of a sentence interpretation, setting aside the selection
task of disambiguation? Or must the model directly re ect the way in which
uncertainty in the interpretation is progressively eliminated? Almost all recent
approaches to parsing have adopted the latter option. It has been generally
assumed that the point of departure in any account of parsing is the radical
underspecication of discrete alternative interpretations of a string, and that
the prime focus is to articulate resolution strategies that identify a denitive
structural description of some syntactic structure (see e.g. Marcus 1980 and
the following tradition of syntactic parsing and the characterization of semantic
underspecication explored in e.g. Pinkal 1995). However, we have chosen the
former option, and have set aside the problem of disambiguation as a secondary
task.
One of the principal reasons for setting the phenomenon of disambiguation
aside is the commitment to dene content of a linguistic expression as the contri-
bution it makes to building up an interpretation (see the discussion of anaphora
in chapter 1). An important distinction which we believe needs to be preserved
in such an account is that some expressions may make two or more distinct
contributions to the interpretation and so are ambiguous, while others have a
only a single contribution which may, however, not be fully determined. In
both cases the supercial phenomenon is one of availability of choice and hence
parsing uncertainty, but this masks the fact that the underlying phenomena in
the two kinds of case are dierent. This distinction is familiar in the case of
lexical content, where it is well established that a distinction needs to be made
between the English word bank (with its two quite discrete interpretations) and
the vacillation in interpretation of a pronoun relative to context, despite the fact
that, from a parsing perspective, it may be equally impossible to assign an in-
terpretation to either, in default of an appropriate context. Any account which
re ects parsing uncertainty directly without dierentiating these cases has no
basis for distinguishing idiosyncratic lexical ambiguity from systematic cross-
linguistic underspecication of interpretation.2 We believe the same distinction
applies at the structural level. While the initial constituent in a left-dislocation
structure may underspecify its interpretation within some unfolding structure,
this is dierent in kind from the uncertainty confronting a parser at the on-
set of a string whose rst constituent is a noun phrase, where the uncertainty
is a consequence of the availability of choice between more than one discrete
structure.
This view is buttressed by work in pragmatics, where there is evidence that
disambiguating a structure and enriching weak specications of content are pro-
cesses subject to the same psychological constraints, but that these none-the-less
constitute quite dierent phenomena (see Sperber and Wilson 1995). Moreover,
the problem of establishing a particular interpretation in any given context is an
essentially cognitive one { according to Relevance Theory concerning interaction
between minimizing cognitive eort for establishing a suÆciency of inferential
eects. Whatever the precise form of these constraints, the problem of dening
them is secondary to the task of dening the process of tree growth involved in
the interpretation process and the form of input that expressions contribute to
the growth of such structures.
2 See the discussion in chapter 1. It should be noted that some accounts of semantic
underspecicity adopt the stance that there is no distinction in principle between these two
cases: see Poesio 1996, van Deemter 1996, Pinkal 1995.
The Data Structures of the Parsing Model 27
Accordingly, we have elected to dene parsing of a sentence as the unfold-
ing of tree descriptions for all possible interpretations. In so doing, we choose
to distinguish not merely between lexical underspecication and lexical ambi-
guity, but also between structural underspecication and structural ambiguity.
And in setting out the landscape of possible construction steps, the framework
articulates all possible transitions for a given sequence of words that yield a
well-formed logical structure, presuming on a pragmatic theory such as Rele-
vance Theory (see Sperber and Wilson 1995) to provide the criteria for choosing
between them. In the process we develop an account of long-distance depen-
dency in terms of initial structural underspecication and subsequent resolution
of uncertainty. We return to the question of whether this system is too liberal
at several points during the book, in particular in chapter 6.
The system we describe is modular, and uses the terminology of Labelled
Deductive Systems (LDS), of Gabbay (1996). Labelled deductive systems are
systems in which deduction is dened for Label{Formula pairs. The components
represented by the Label and the F ormula are independently dened systems,
and the Labelled Deductive System in which they are embedded provides the
logic of their combination. In our system, the formulae represent sub-terms
of logical forms, and the labels represent specications that drive their com-
bination. The operations that combine formulae are set within a goal-directed
enterprise of constructing an overall logical form. In this construction process, a
tree structure is progressively introduced and decorated with labelled formulae.
The process is driven by instructions projected by the sequence of words of a
natural language string.
We introduce the system in three stages, represented diagrammatically in
gures 2.4{2.6. Initially we present the basic structures and a vocabulary for de-
scribing them (in this chapter). This characterization will necessarily be rather
abstract, since every phenomenon explained in subsequent chapters involves the
characterization of how partial structures grow; and to see this, we have rst to
set in place the vocabulary for describing individual partial trees. We then de-
ne the dynamics involved in building up individual tree structures (chapter 3) {
covering simple clauses, complement clauses and long-distance dependency. We
then extend the system to account for relative clauses (chapter 4), introducing
a concept of linked tree structures.
2.2 The Data Structures of the Parsing Model
In our parsing system, so-called Formulae are used to express the denotational
content of individual terms, phrases and sentences. Arguably, the construction
of these Formulae is the central goal of the parsing process, and in our system
they will take the form of terms of a familiar typed lambda calculus. The mode
of combination of these formulae is determined by their logical type, and these
In this chapter the two types of Data Structures are introduced that are used by our
(theoretical) parser: rstly Basic Tree Structures, the representations of Logical
forms (denition 2)and secondly Partial Tree Structures (denition 5), the
objects the parser constructs on its way to full Basic Tree Structures.
Basic Tree Tree Structures consist of binary trees the nodes of which are decorated
by Declarative Units (section 2.2.1). The following structure represents the meaning
of John admires Mary:
h: : : ti : Admire(Mary)(John)
h: : : ei : John h: : : e ! ti : y[Admire(Mary)(y)]
h: : : e ! (e ! t)i : xy[Admire(x)(y)] Mary

Pulling the tree and the decorations apart we get the following
Basic Tree Declarative Units
h: : : ei : John
h: : : ! ti : y[Admire(Mary)(y)]
h: : : e ! (e ! t)i : xy[Admire(x)(y)
The tree structure is addressed by tree The tree decorations are addressed by La-
modalities (section 2.2.2), tree node bel and Formula predicates, for instance,
identiers (section 2.2.6.2), and by the the Label predicates T y (section 2.2.1)
(sub)language ADR (denition 7). and T n (section 2.2.6.2) and F o, the For-
mula predicate (section 2.2.1).
Figure 2.4: Basic tree structures

Partial Tree Structures consist of structures that can develop to Basic Tree Struc-
tures. The nodes of Partial Tree Structures are decorated by Partial Declarative
Units and Requirements (section 2.2.5)
[ T y(t); h#1 i(T y(e ! t))]
[h:::ei : John ] [ T y(e ! t)]

Partial Tree Partial Declarative Units
[ T y(t); h#1 i(T y(e ! t))]

[h:::ei : John ]
[ T y(e ! t)]
The Partial Tree Structures are addressed The Tree Decorations consist of pairs of
like Basic Tree Structure, but there is a (partial) declarative unit, to the left of
an additional distinction between Exter- the bullet and a set of requirements to
nal and Internal Tree Modalities (section the right of the bullet (denition 5).
2.2.5.1).
Figure 2.5: Partial tree structures
types we will display as Labels to the lambda term which is the corresponding
Formula.3
The simplest mode of combination might be represented as the following
application of what is in eect a step of Modus Ponens on the side of the Type
label and function application on the Formula side:
e : John e ! t : Smile
t : Smile(John)
Lexical items of the natural language project, among other things, the requisite
labelled formulae to act as premisses. These are not the words themselves
3 The types, taken from type theory as in the Montague grammar tradition, are expressed
as in type-logical grammars as propositional formulae in a conditional logic: the semantic
type he; ti corresponds to the formula e ! t, he; he; tii as (e ! (e ! t)), etc. See Morrill 1994,
Moortgat 1997, Oehrle 1995.
Partial Tree Structures are connected by the relation of development or exten-

sion. This relation is denoted by (denition 5):
[ T y(t); h#1 i(T y(e ! t))] h: : : ti : Admire(Mary)(John)

h: : : ei : John
[h:::ei : John ] [ T y(e ! t)]
h: : : e ! ti : y[Admire(Mary)(y)]
Both Basic and Partial Tree Structures are represented by Tree Descriptions as sets
of sets of formulae in the language DU (section 2.2.2.1).
fT n(a); ?T y(t)g
fh" iT n(a); F o(John); T y(e); [#]?g fh"1 iT n(a); ?T y(e ! t)g

A Node Description (section 2.2.6.1) consists of a set of formulae containing a Tree
Node Identier (denition 7). A Tree Description consists of a set of Node Descrip-
tions satisfying some consistency and coherency constraints (section 2.2.6.2).
Figure 2.6: Tree developments and tree descriptions
but concepts which the word is taken to express. Though the types provide
the major label driving the combinatorial process, other labels may provide
additional control, including gender specications, scope restrictions, labels for
discrete sentence types, etc. The unifying characteristic of these labels is that
they serve as controls on the process of setting out and combining the formulae
projected by the lexical expressions. Initially we shall assume that the only mode
of combination consists of the twinned operations of functional application and
type deduction, as indicated in the above step of labelled Modus Ponens, but,
in chapter 4 we will dene a mode of combination which resembles an update
action to represent the incorporation of information of a relative clause into the
head.
We will introduce in tandem our basic data structures, decorated partial
trees and the language DU to talk about them (DU for Declarative Unit). In
this section we will not be wholly formal but will introduce our structures by
highlighting three typical features of the account which together bring out the
form of the descriptions we shall give and the dynamics of how such descriptions
grow:
the atomic propositions of the language DU
the tree modalities the language DU uses to describe relations within a
tree structure
the representation of requirements in DU to dictate directions in which a
tree is to develop
2.2.1 Atomic Formulae
In the following paragraphs, we rst dene the Type and Formula predicates
and introduce the propositions which serve as decorations of the tree structure.
We shall dene Labels and Formulae initially by means of monadic predicates
in a predicate-argument notation, but in the nal chapter we shall provide a
more general denition in terms of feature structures. These propositions we
then embed in a tree description language which will be adequate for describing
logical forms in decorated tree representation.
As we have so far sketched, the aim of our parser is to construct decorated
tree structures. The decorations on the nodes of such trees consist of declarative
units. Declarative units are pairs consisting of a sequence of labels followed by
a (content) formula:
h|l1 ; :{z: : ln}i : |{z}
.
Labels F ormula
The general methodology behind such labelled formulae is that the formula is the
primary term, with labels providing additional information about that formula
{ its potential for combination with other formulae, its semantic evaluation, its
position in a structure, and so on (see Gabbay 1996). We have seen examples of
formulae in the denotations John, Mary and xy[Admire(x)(y)]. The types
e, t and e ! t from these examples are instances of labels. For instance, the
decoration of the root node of gure 2.2 is the declarative unit
h: : : ; ti : Admire(Mary)(John):
It is the language DU which expresses facts about these declarative units. The
atomic propositions of the language have the form P (a), where P is a monadic
predicate letter and a is some appropriate term. For each Label category of the
declarative units we will have a predicate Lai, and for the Formula category we
have the predicate F o. The above declarative unit, for instance, supports the
facts F o(Admire(Mary)(John)) and T y(t), which we will display as
h: : : ti : Admire(Mary)(John) j= T y(t) ^ F o(Admire(Mary)(John)).
Here du j= Lai(li ) means that the formula Lai(li ) holds at, or is satised by,
declarative unit du.
The parser we are constructing takes a natural language string to produce
a logical term in some formal language. This term we represent in the form
of a decorated binary tree. So our data structures will be decorated trees. To
describe these structures, however, we need a (second) formal language, so
we introduce a language to talk about data structures which themselves are
formulae in some language.
When a proposition, say P (a), is true of an object hl1 : : : lni : we will use
the notation
hl1 : : : ln i : j= P (a):
That is, we separate the object on the left from the proposition on the right by
the symbol `j='. If P (a) is true on object hl1 : : : lni : , we will also say that
hl1 : : : ln i : satises P (a) or that P (a) holds on hl1 : : : ln i : .
So, as displayed above, the declarative unit h: : : ti : Admire(Mary)(John)
satises both the proposition T y(t) expressing that the unit has the Type label t
and the proposition F o(Admire(Mary)(John)) stating that the content formula
of the unit is Admire(Mary)(John). This is an instance of a general pattern.
A declarative unit hl1 ; : : : lni : can be completely represented by a description,
the nite set of atomic propositions satised by that unit:
fLa1(l1 ); : : : ; Lan (ln ); F o( )g:
So in this description, both labels and formulae are represented by monadic
predicates. A partial declarative unit, an object naturally arising in the course
of a parse, is now merely a subset of a description of a declarative unit. The
descriptions of declarative units determine the atomic vocabulary of the lan-
guage DU . So, this language has monadic predicates La1; : : : Lan; F o, standing
for n Label dimensions and a Formula dimension and individual constants from
DLa1 ; : : : DLa ; DF o respectively, denoting values on these dimensions. The sub-
sequent paragraphs will discuss the Type label T y with values in the domain
n
DT y = fe; t; cn; e ! t; : : : g and the Formula predicate, which takes values in

DF o .
The Type Predicate T y

With a process of node annotation which is a re ex of the isomorphism of type
deduction and function application, the Type predicate T y is the backbone that
underpins the general process of combination of information.
The predicate T y takes as argument values logical types from a nite set
DT y , expressed as terms in a conditional logic. We assume that the basic types
e, t, cn are in Dty . We have no general recursive step determining typehood,
but ll DT y with some nite number of further types determined by the lexicon
of the language. There are no general rules of type inference: there are, that
is, no operations of type-lifting, or of composition of functions.4 We will use
the predicate types e ! t, e ! (e ! t), e ! (e ! (e ! t)), t ! (e ! t);
the determiner type (cn ! e), the noun type e ! cn, the sentential modier
type t ! t, and adverbial modier type (e ! t) ! (e ! t). As we shall
see, expressions in other frameworks analysed as NP or DP are in this volume
assigned the type e. This is made to work by using variable binding term
operators rather than quantiers as denotations of determiners, a matter to
which we return in a while.
Bearing in mind the focus on partial specications, a general feature of the
language will be the incorporation of meta-variables in the language itself as
place-holders for the proper terms: that is, terms belonging to the domains
DLa [ DF o . In the case of the Type predicate too, we allow their use. In that
case they expect to be replaced by elements of DT y . So we allow propositions
i
of the form
T y(U ! (e ! t))
as annotations of tree nodes. Notice that U ! (e ! t) 62 DT y : this schematic
type can end up as types e ! (e ! t); t ! (e ! t), (e ! t) ! (e ! t), etc. by
updating, but is not itself a type (see Marten 1999).
The Formula Predicate F o
The Formula predicate F o has as values elements in the set DF o consisting of
terms of a typed lambda calculus. These terms express the content assigned
to a natural language string, and, arguably, the construction of a value of the
F o predicate, given a natural language string, is the whole point of the parsing
process. The Formula predicate and the Type predicate work in tandem in the
basic tree conguration
fF o( ()); T y(Y )g
fF o(); T y(X )g fF o( ); T y(X ! Y )g

As we have seen already, the content formulae we will be using are terms of
a typed lambda calculus. For instance, the transitive verb admire projects a
lambda expression of the form xy[Admire(y)(x)] (with type (e ! (e ! t))
which requires two arguments. As is standard, an argument like Mary (of type
e) is absorbed into the predicate by -reduction: xy[Admire(y)(x)](Mary) =
y[Admire(y)(x)](Mary=x) = y[Admire(y)(Mary)].
4 In Categorial Grammar, type-lifting and composition of functions is universally assumed
as the basis for characterizing non-constituent coordination (see Steedman 1996, Dowty 1987,
Morrill 1994, Milward 1995). Though we do not address the issue here, we might assume that
co-ordination is handled through a form of linked structure (see chapter 4).
There are several idiosyncratic features of this representation language. The

rst is the use of meta-variables as values of the predicates. As established in
chapter 1, a central task for this framework is to capture the underspecication
of content projected from the lexicon, for example by pronominals and other
anaphoric expressions; and this content underspecication is expressed through
meta-variables which are (to be) replaced during the interpretation process.
These are not object-level variables, but substitution sites for other DF o ex-
pressions. For instance, the word he does not encode a given logical formula,
but a meta-variable, with constraints on the possible substituends.
A proposition containing a meta-variable will be explicitly represented as
incomplete in virtue of an outstanding requirement. For example, a proposition
of the form F o(U) will be accompanied by a requirement for a value from DF o,
here represented by a question mark (see section 2.2.5):
fF o(U); ?9xF o(x)g:
In order to end the parsing process without outstanding requirements, an actual
term has to replace the meta-variable U, for the existential quantier will only
range over elements of DF o.
A further domain in which meta-variables prove useful is the characterization
of tense. In what follows, we will by and large ignore tense, but we wish to
tentatively adopt the possibility of capturing temporal relations through the
addition of a label internal to the formula associated with the type t, Si : A.
The values of the F o predicate upon this alternative view would consist of tuples
fF o(Si : A); ?9yF o(y : A)g;
where Si is some `time-variable'. Whenever = S : A, we set for F o():
T i() = Si and for() = A. This re ects the semantics of time by positing
an additional sortally restricted argument, on which can be predicated an array
of dierent properties and relations as indicated by temporal adverbials of one
sort or another. Informally, Si : P is true if there is a modal/temporal unit Si
relative to which the formula P is true.5 To express this, we precede such a F o
predicate with a statement about the temporal relation in which this variable
is involved, for instance
fT e(Si < Se ); F o(Si : A); ?9yF o(y : A)g:
This choice of notation directly follows van Eijck and Kamp (1997) in having
an additional argument position for an event variable, but without abstracting
over it. It also re ects work by Finger and Gabbay (1992) and Russo (1996) on
5 The motivation for some such temporally restricted argument is now well established: see
Partee 1973 on the anaphoric properties of tense, Kamp and Reyle 1993 for arguments against
analysing natural language tense construal in terms of propositional operators; Davidson 1967
and work following in that tradition on the need to posit event variables; and for the claim
within Situation Semantics that sentences are evaluated relative to situations Barwise and
Perry 1983, Cooper 1996. See Perrett 2000 for a detailed application of this approach to tense
in the characterization of the so-called verb-chaining phenomenon.
combining predicate logic and temporal logic in a single system, with the label
representing the index of evaluation for a given formula, and relations between
such indices expressed in a sortally restricted predicate logic system externally
to the formula. As with anaphoric expressions, such time variables, at least
as projected from the lexicon, underdetermine the information pertaining to
the temporal ow within which the associated proposition is construed, and so
must be represented by meta-variables as place-holders, standing in for whatever
indexical or quantied expression represents the assigned content. The set of
temporal statements that are argument to the predicate T e are of the form
S < S0 (S precedes S0 ); S O S0 (S overlaps S0 ), etc. In all that follows, we skate
over issues involving temporal interpretation, noting merely that there are a
number of ways in which tense could be addressed in the framework.
A further distinguished set of meta-variables are those of the form WH, which
we shall take to be the formula projected by wh expressions in wh questions.
These meta-variables are exceptional amongst meta-variables in being retained
in the formula that results. The assignment of interpretation to a sentence such
as Who did John like? thus gives rise to a tree whose root node is annotated
with the formula Si : Like(WH)(John).
A second idiosyncratic feature of the language we use for content formulae
are the variable binding term operators. These operators are projected by the
determiners and are of type cn ! e. Examples of such operators are the ,
and operators. These bind variables within an individual expression to create
terms with bound variables.6 For example, a phrase like some man will project
a formula (; (x; Man(x))). This is a term consisting of three components: the
operator indicating the mode of quantication (e.g. indenite, existential), the
variable x that is bound by the operator, and the context in which this variable
is bound, the restrictor Man(x). The determiner some itself will project the
formula X (; X ) (of type cn ! e) which expects the projection of a common
noun (e.g. (x; Man(x))) as argument. So all noun phrase expressions are taken
to map onto a formula of type e in the computation of the tree structure, and
thus combine as arguments with some n-place predicate expression to form at
some root node of type t a propositional formula. The eect of this is that we do
not type-raise expressions corresponding to noun phrases to match the type of
generalized quantiers (of type (e ! t) ! t)), but rather assign all noun phrases
a type re ecting their combinatorial property of being an argument to relations
projected by predicates. Scopal relations are explicitly represented only once
the propositional structure has been established using a dependency predicate
x < y representing that the term involving x has scope over that involving
y (see chapter 7). Where this is not xed by rule, the choice of dependency
relations is assumed to be established broadly on-line as is anaphora construal,
through a choice of argument values to the dependency predicate. For instance,
the sentence Some man admires every woman interpreted with some having
scope over every will be represented by two pieces of information: rst the
6 The Epsilon Calculus was dened by Hilbert (see Hilbert and Bernays 1939) as a means of
providing an explicit calculus for the study of arbitrary names in natural deduction systems.
We will return to this calculus when discussing quantication (see chapter 7).
formula S : Admire(; y; W oman(y))(; x; Man(x)) and second the statement

that x < y; that is, in the resulting formula the quantier binding x precedes
the quantier binding y. The statement x < y, chosen as the tree and its
annotations unfold, gives the dependency relation between the two terms. This
statement sets an algorithmic process in motion that eventually results in a
formula with the right scope relations.
[h: : : ; x < y; : : : ti : Admire(; y; W oman(y))(; x; Man(x))]
[h: : : ei : ((; x; Man(x)) ] [h: : : (e ! t)i : Admire(; y; W oman(y))]
[h: : : (e ! (e ! t))i : Admire] [h: : : ei : (; y; W oman(y))]
Figure 2.7: Projecting quantied terms, with scope-dependency marking

In this tree, the dependency predicate indicates the scopal relations that must
be projected in order for the particular logical form required to be complete.
The details of this operation we return to in chapter 7.
2.2.2 Tree Modalities
The concept of tree growth requires a language to talk about trees. Decorated
trees will be represented by structured sets of declarative units in the following
way:
h: : : ei : John h: : : (e ! t)i : Admire(Mary)

Each of these three nodes can be described in the language DU by the atomic
propositions it satises.
h::::ti : Admire(Mary)(John) j= F o(Admire(Mary)(John)); T y(t)
h::::ei : John j= F o(John); T y(e)
h::::(e ! t)i : Admire(Mary) j= F o(Admire(Mary)); T y(e ! t)
The representation of these declarative units in a tree structure provides the
interpretation for the tree modalities of the Logic Of Finite Trees (LOFT {
Blackburn and Meyer-Viol 1994, Blackburn et al. 1996). In this logic, modal
statements allow the description of (part of) a tree from the perspective of
any arbitrary node within that tree. For instance, from the perspective of a
given node in a tree we can refer to facts which hold at its argument daughter,
its function daughter, or simply its daughter. Conversely, we can refer to its
mother, and nally we can address facts which hold at nodes it dominates.
For this purpose, the language DU will use the LOFT-modalities h#0i (at the
argument daughter), h#1i (at the function daughter), h#i (at some daughter),
h"i (at the mother), h# i (somewhere below) and h" i (somewhere above). The
dominance modalities h#i and h"i are interpreted along re exive transitive
closure of the basic tree relations daughter of and mother of respectively. So, for
instance, a formula h#0i holds at some node if its argument daughter satises
. If, in the following array, we let T be a shorthand representation of the nal
tree of example 2.1 (gure 2.2), then we have, for instance, the following facts:
T ; h: : : ti : Admire(Mary)(John) j= h#0 iF o(John); T y(e)
T ; h: : : ei : John j= h"0 i(T y(t))
T ; h: : : (e ! t)i : Admire(Mary) j= h"ih#i(T y(e))
T ; h: : : ti : Admire(Mary)(John) j= h# iF o(Mary)
Here h#0i(F o(John) ^ T y(e)) is satised at a node if there is an argument
daughter where F o(John) ^ T y(e) holds. And with respect to the top node of
the tree which we have represented as T this is indeed the case. On the other
hand, the formula h"ih#i(T y(e)) holds at a node if going up brings us to a node
which has a daughter satisfying T y(e). The meaning of the modalities can be
recovered from the abstract example given in gure 2.8.
1 ; h#0 i2 ; h# i5
2 ; h"0 i1 3 ; h#i4 ; h#i5
4 ; h" ih#0 i2 5 ; h"ih#0 i4
Figure 2.8: Interpretation of the existential tree modalities

As is traditional, the angled modalities, h#i, are interpreted as existential:
`there is a node : : : ', while square bracketed modalities, [#0]; [#1]; [#]; ["]; [#]; ["]
will represent their universal duals: `for all nodes : : : '. The exact denitions
follow in denition 6.
Some of the modalities are related as inverses to others. For instance, the
daughter relation is the inverse of the mother relation, and vice versa. For a
given modality h#i its inverse will be 1denoted by h# 1i. Now we have h#0 1i =
h"0 i, h"0 i = h#0 i, h#1 i = h"1 i, h"1 i= h#1 i, h# 1 i = h"i, h" 1i = h#i, h# 1 i
1 1
= h"i, and h" 1i = h#i.
2.2.2.1 Denition of the Language

At this point we have gathered enough information to give a more formal def-
inition of the language we will be using to represent the meanings of natural
language strings. These meanings will consist of unreduced terms of a typed
lambda calculus presented in the form of decorated trees. The structures this
language is meant to talk about are binary trees whose nodes are annotated
with declarative units. We will rst formally introduce the DU language and
then dene the data structures that are the intended models for the language.
The formulae of the language DU include, as is familiar, the set made up
of logical constants > `verum', ? `falsum',7 of atomic statements about Labels,
atomic statements about Formulae, Boolean combinations, and universally and
existentially quantied formulae. However, in addition, there are modal formu-
lae, where the modal operators are the LOFT operators denoting tree relations
like mother, daughter, etc.; and these will be central to the description of rela-
tions between decorated nodes.8
Denition 1 (The Language DU9 )

For
{ V AR = fx; x1; x2; : : : g a denumerable set of individual variables,
{ MV = fU; U1 ; : : : ; S; : : : g a denumerable set of meta-variables,
{ li 2 DLa [ VAR [ MV a Label value for each i: 1 i n,
{ 2 DF o [ VAR [ MV a Formula value,
i
{ t1; t2 2 DLa [ DF o [ V AR [ MV ,
{ # an element of f#0; #1; "0; "1; #; "; #; "g
i
A proposition A of the language DU is dened by

A ::= > j ? j La1 (l1 ) j : : : j Lan(ln ) j F o() j Eq(t1 ; t2 ) j A ^ A j A _ A j
j A ! A j 9xA j 8xA j h#iA j [#]A j?A j?;:
The individual variables are set in bold-face to distinguish them from the vari-
ables occurring in the elements of DF o, the formulae of the representation
language. The set of atomic formulae of DU { that is, formulae of the form
La1 (l1 ),: : : ,Lan (ln ), or F o() { will be denoted by AT , and the set of formulae
starting with a question mark will be denoted by REQ. The DU formula ?;,
where the symbol `;' denotes the empty set, is a special element of REQ that
will be given a xed interpretation in denition 3.
7 These two logical symbols stand for universal truth, holding at every tree node of every
tree, and the universal falsehood holding at no node of no tree, respectively. In combination
with the modal operators, they can be used to express the existence and non-existence of
daughter nodes, mother nodes, etc. Their uses here will become clear in the course of the
chapter.
8 We will also introduce in the language the operator ? for representing requirements (see
section 2.2.5), which we haven't yet discussed. This is done to avoid having to extend the
language. The interpretation of this symbol is given in denition 6.
9 See n.8.
2.2.3 Basic Tree Structures
The language DU is a language to talk about binary trees the nodes of which
are decorated with a declarative unit and a set of requirements. A basic tree
structure is a pair consisting of a tree domain, T rDom, a set of tree nodes, and
a set of tree relations, T rRel, between nodes.
Denition 2 (Basic Tree Structures) A basic tree structure T is an ordered
pair
T = hT rDom; T rReli
where
T rDom = f[du] j du is a declarative unit; g
T rRel = f0 ; 1; ; g.
For n; n0 2 T rDom, n 0 n0 holds if n0 is the argument daughter of n and
n 1 n0 holds if n0 is the function daughter of n. We dene `' as the
daughter relation, i.e. = 0 [ 1, and 10as the dominance relation,
i.e. is the re exive transitive closure of .
The relation between the basic tree structures and DU formulae is given by
the denition of satisfaction. This denition exhaustively xes the meaning of
the constructs of the DU language, and so is central to an understanding of
the concepts of decorated nodes in a tree and their relation to each other, and
hence, as we shall see in a while, the potential for tree growth.
Denition 3 (Satisfaction I) Given a Basic Tree Structure T ; n = [hl1 : : : ln i :
R] 2 T rDom, which is a pair of a tree node and a requirement set, and
D = DLa [ : : : DLa [ DF o , we say that node n of Basic Tree Structure T
satises DU formula , with the notation11
i n
T ; n j= ;
if
6= ?
=>
= Lai (a) and li = a
= F o() and =
= Eq(t1 ; t2 ) and t1 = t2
= ^ and T ; n j= and T ; n j=
= _ and T ; n j= or T ; n j=
= 9x and there is a t 2 D : T ; n j= [t=x]
10 The re exive transitive closure of the relation is the set of all ordered pairs hn; n0 i such
that there is a natural number k, possibly 0, such
n : : : n0 :
| {z }
k times
11 The Basic Tree Structures T interpret only the `static' logical connectives and operators:
`_', `^', and `9x', and `h#i'. The remainder of the standard connectives and operators, i.e.
`:', !', `8x' and `[#]' will be discussed shortly.
if i 2 f0; 1g and
= h#i i , and 9n0 2 T rDom : n i n0 and T ; n0 j= ,
= h"i i , and 9n0 2 T rDom : n0 i n and T ; n0 j= ,
= h#i , and 9n0 2 T rDom : n n0 and T ; n0 j= ,
= h"i , and 9n0 2 T rDom : n0 n and T ; n0 j= ,
= h# i , and 9n0 2 T rDom : n n0 and T ; n0 j= ,
= h" i , and 9n0 2 T rDom : n0 n and T ; n0 j= .
Notice that no decorated node n satises ?; > holds on every node, and the
atomic formulae hold if the declarative unit annotating the node has the right
label and formula values. The Boolean connectives and the modal operators
have their standard interpretation.
The meta-variables can be distinguished from the proper values by the fact
that only for a proper value li we have the satisfaction of 9xLai (x). 9x holds
at a node if [t=x] holds there for some t 2 D = DLa [ : : : DLa [ DF o. We
want the existential quantier to be able to express that some label or feature
i n
predicate has a value. And a meta-variable just won't do as a proper value.

For linguistic purposes, the language which describes such basic tree struc-
tures provides a very considerable richness in the way an individual node can be
characterized. An arbitrary node n in a tree may be described in the following
terms, with a Kleene star operator:
T ; n j= h# iX Some property X holds either at n or at some daugh-
ter, or at some daughter's daughter, etc. `Some prop-
erty X holds either here or at a node somewhere be-
low here.'
Or, for example, using the converse h"i modality, we can describe a node n as:
T ; n j= h" i(Cat(+Q) ^ T y(t)) Àt, or somewhere above, n is a

node bearing the feature +Q and
the attribute T y(t)'.
These descriptions are, notably, suitable for either correlating a node n high in
a tree with some arbitrary lower position in the tree, or, conversely, correlating
some node n with some arbitrarily higher node. However, the striking property
of such modal descriptions, and indeed the primary motivation for turning to
tree descriptions, is that they make available a form of underspecication for
tree descriptions, hence of structure. Conventionally, descriptions of tree struc-
ture as part of the syntactic vocabulary have been unlike semantic descriptions
in this respect. Semantic descriptions of expressions may be weak with respect
to some assigned content (see chapter 1), a partiality of content which our for-
mulae must be able to re ect.12 Tree descriptions, to date, have been allowed
12 Partiality of content is well recognized in the literature: see Barwise and Perry 1983,
Kamp 1984, Kamp and Reyle 1993, van Deemter 1996.
no such exibility in grammar formalisms. The only concept of underspeci-
cation available in the syntactic vocabulary has been of feature specication, a
re ection of redundancy. Tree structure, as part of syntactic description, has
been invariably assumed to be dictated by algorithm and fully xed. With the
use of Kleene star descriptions, we open up the possibility of syntactic relations
which, at a given point in the process of articulation for a given structure, are
as yet unxed, describable solely through modalities h"i, h#i. This possibil-
ity we have already introduced informally, and it will be further developed in
due course. Many consequences for natural language analysis follow from this
increased richness of descriptive potential. Here we just ag the potential that
the Kleene star operator in this modal tree logic provides for characterizing
long-distance dependency without having to invoke pairs of trees, movement or
even the transfer of features from one position in13 a tree to another. We shall
return to this in detail in the following chapters.
The operators we have dened above are called static because their evalua-
tion only refers to the current tree structure. In order to interpret the connec-
tives and operators with universal force { that is, `!', `8x', and `[#]' { we will
have to say some more about the tree structures that occur in the course of a
parsing process.
2.2.4 Partial Tree Structures

As so far characterized, the system is very close to Categorial Grammar, in
which interpretation and structure for the string can be compiled hand in hand
on a bottom-up basis. However, our aim, quite unlike that of most Categorial
Grammar formalisms, is to model the step-by-step nature of the process of
assigning a logical form to a natural language string, and to formally dene the
irreducibly partial nature of the descriptions involved at all non-nal stages.
Furthermore, in our system, the assignment of a logical form to a string is a goal-
directed task. The goal-directedness is realized by representing logical forms as
decorated trees and introducing every node in such a tree with a requirement
for a particular kind of sub-term, a requirement which is characteristically met
only at some subsequent point in the construction process. It is this shift to
introducing each node with a set of requirements that drives the tree growth
process, distinguishing this model from other more static systems.
13 The Kleene star operator is widely used in natural language parsers which use tree de-
scription languages, as already noted (see Marcus et al. 1983, etc.). However, its only other
use to our knowledge in a grammar formalism is in LFG (Lexical Functional Grammar: see
Kaplan and Zaenen 1989), where it is used to express uncertainty at the level of functional
structure. The primary dierence in the use of the Kleene star operator in LFG is that there
it is not associated with the dynamics of any tree update system (see section 3.3).
2.2.5 Requirements
At a particular state in the parsing process, a (partial) tree will have been
constructed, the nodes of which are decorated on the one hand by (partial)
declarative units and, on the other hand, by requirements to be satised. The
root node, for instance, will have a requirement that it be annotated by a lambda
term of type t, a sub-node of the root that it be annotated by a term of type
e, and so on. Each of these requirements is a proposition in the DU language.
This use of requirements on the development of a tree node has some
resemblance to the familiar concept of subcategorization, as a node decorated
by the labelled formula
h: : : ; (e ! (e ! t))i : Admire
within a tree may have a mother node which is decorated with a requirement
h#0 iT y(e) (i.e., a requirement for an internal argument for Admire). However,
the concept of requirement is much more general than subcategorization state-
ments: all nodes are introduced with requirements. Requirements form an
essential feature of the tree { they determine the set of successful extensions
of a given (partial) tree: namely, those in which all requirements are satised.
Consequently, our basic data structures are tree structures, the nodes of which
consist of (partial) declarative units paired with nite sets of requirements.
A requirement can be any formula in the DU language under development.
A tree node will now be decorated by a pair consisting of a partial declarative
unit du and a nite set of requirements R:
[du R]
Here the bullet separates the body of the declarative unit from the requirements.
In the language DU which describes these nodes, we will introduce a question
mark to represent requirements, that is:
[du R] j=? i 2 R.
Below is an example of a partially decorated tree T that occurs in the construc-
tion of John admires Mary, (2.1):
[ T y(t); h#1 i(T y(e ! t))]
[h:::ei : John ] [ T y(e ! t)]

The requirement for a type t at the top node and the presence of h:::ei : John at
an argument daughter go together with the requirement for a type e ! t to hold
at the function daughter. At the top node this is re ected by the requirement
for h#1i(T y(e ! t)).
T ; [ T y(t); h#1i(T y(e ! t))] j= ?h#1 i(T y(e ! t))
Moreover, because the function daughter has a requirement for type e ! t, at
the top node this shows by the satisfaction of h#1i?(T y(e ! t)).
T ; [ T y(t); h#1i(T y(e ! t))] j= h#1 i?T y(e ! t)
This use of requirements on nodes and the progressive eect they have on the
unfolding and subsequent annotation of the tree constitute the heart of the
transition rules which we will dene to capture tree growth.14
The logical formulae which are the output of the parsing process can be
represented as decorated basic tree structures of denition 2. However, the
structures that are constructed at intermediate stages of the parsing process
need not yet be basic tree structures. For instance, an intermediate structure
may have a node with only one daughter, or it may have two nodes related by
the daughter-of relation without being more specic about which daughter it is
(i.e., the relation is not (yet) of the form 0 [ 1), or, nally, it may have
one node dominating another node without there being a sequence of immediate
dominance steps witnessing this (i.e., the relation is not (yet) the re exive
and transitive closure of ). But all these intermediate structures have in
common that they can be completed to basic tree structures by simply adding
nodes, tree relations, and decorations. A structure which can be completed to
a basic tree structure (BTR) we will call a partial tree structure (PTR). We will
give an informal denition, anticipating the formal version in the nal chapter.
Denition 4 (General Tree Nodes) A general tree node is a pair
[du R]
where du is a declarative unit and R a nite set of DU formulae.
The set R stands here for a nite set of requirements (associated with that
node). A node [du] of a basic tree structure can be represented as a general tree
node in the form of [du ;], where the symbol `;' denotes the empty set (of
requirements). Requirements must be seen as tasks that have to be taken care
of, and thus no tree node should satisfy one of its requirements. So, in order for
the sets R to actually represent requirements, we must at least have
If T ; [du R] j= , then 62 R.
Denition 5 (Partial Tree Structures15 ) A partial tree structure T is a
tuple
T = hT rDom; T rReli
which can be completed to a basic tree structure. We say node [du R] in some
partial tree structure T , can be extended to node [du0 R0 ] in the partial tree
structure T 0 , with the notation,
T ; [du R] T 0 ; [du0 R0 ];
if
14 For examples of the uses of requirements, see section 9.2.2.3.
15 For more formal denitions, see section 9.2.1.1, denitions 2, 3 and 4.
structure T 0 can be constructed from T by adding nodes, tree relations

and decorations. Formally this will come down to stipulating that there is
a tree morphism (preserving tree relations) mapping structure T to T 0 .16
du0 is a consistent extension of node du: that is, du0 has updates or rewrites
on some label and/or formula co-ordinates of the corresponding values of
du. For each 1 i n there, the label values in DLa are partially ordered
by a relation i (DLa [ MV ) (DLa [ MV ). This relation embodies
i
an update relation on the labels, where l La l0 if l0 ìs more informative'

i i
than l. The specic content of this notion depends on the specic Label
i
dimension Lai. What is shared by all dimensions is that meta-variables

U lie below concrete values: that is, U La l for all l 2 DLa . Thus we
can represent the replacement of a meta-variable by a concrete term as an
i i
incremental change. If the label li is dened, then it can be one of two

things: it can be a meta-variable, li = U, or it can be a proper value,
li = a 2 DLa . And we have
i
U La a: i
Development along this relation is irreversible. In general we can see a

relation l La l0 as a re ection that l0 is an update of l or that there is
some nite rewrite sequence connecting0 l and l0. Terminal0 elements under
i
La (i.e. elements l such that l La l implies that l = l ) then represent
normal forms under this rewrite relation. For declarative units du; du0 we
i i
can now set du du0 if for all 0 < i < n + 1 : li La li0 and F o 0 .17
i
every requirement in R is either a member of R0 (it still is required)

or it is satised at du0, i.e. T 0; [du0 R0] j= . Here goal-directedness
is incorporated in the relation of extension: requirements may disappear,
but only by becoming facts. As a basic tree structure has an empty set of
requirements, if a structure T ; [du R] can be extended to a basic tree
structure T 0 ; [du0 ;], then all requirements in R must have become facts
at [du ].
0 18
The models we are going to dene will be essentially dynamic, in that they

represent `temporary' states, incomplete logical forms resulting from having only
partially parsed a natural language string. This is unlike the models of First-
order Predicate Logic which are `nished', `complete' or `static' structures; a
rst-order formula either holds on a model or its negation does. If a formula
holds, then that's it; and if it does not hold, then this can also not be changed,
for the model is given once and for all. The models we introduce in this chapter,
however, are not of this Platonic kind. Our models are, generally, momentary
entities that come into being at some point in a parse process only to be changed
at a next point. This change, however, is of a restricted kind: only increase
16 For a variety of tree extensions, see the remark of the nal chapter, 2.
17 For some examples of update relations, see the nal chapter, 4.
18 In the nal chapter, 17 gives a formal denition.
of structure is allowed. Once a momentary model has a certain feature, this is
maintained throughout the parse and can be recovered from the nal, `complete'
model. This feature of our models is captured under the name monotonic or
incremental. The notion of satisfaction on such models in ux requires some
rethinking of the meaning of the logical connectives. If we consider one of
our momentary models { that is, a partial tree structure { we can distinguish
between two kinds of connectives. On the one hand, there are the ones that
can be wholly evaluated on the local tree structure, the `static' connectives
as we have called them above. The meaning of these connectives is entirely
classical: they treat the momentary structure as a complete and nal one. On
the other hand, there are connectives which seem to need reference to future
developments of the current momentary model. A good example is the case of
negation. Suppose proposition T y(e) does not hold at some node n in structure
T . Classically we could state T ; n j= :T y(e). But now suppose that in some
development of model T , node n becomes annotated with T y(e). Then we do
have T 0; n0 j= T y(e), and the information :T y(e) has changed to T y(e). That
is, under a classical interpretation, whether a proposition holds at a node (e.g.
:T y(e)), in general, does not persist, and information may be lost. In our set-
up, we want information to persist. That is, we want :T y(e) to hold at a node,
if T y(e) does not hold there and never will. The presence of proposition
: excludes proposition forever from holding. But then, in the denition of
satisfaction of : at a node n of some tree T , reference has to be made to the
trees T 0 that model T can develop into.
Another example of a connective that we can only evaluate with respect to
development of models is implication. The proposition ! expresses an
invariant connection between satisfaction of and that of . Implication is
an essential connective to express such lawful connections. Under the classical
denition T ; n j= ! holds, for instance, if does not hold at n. But this
implies that, if comes to hold at some future development of T and n, it
may be that ! becomes false. Thus, a classical denition would allow a
change from truth to falsehood of an implication. Again, this is what we want
to prevent. When we have T ; n j= ! for some partial model T , then we
want it to mean that in any development T 0; n0 of T and n, if holds at this
development of n, then also will hold there. An implication does not just say
something about (a node in) the current, momentary tree: rather, it expresses
an invariance over possible developments.
The same holds, nally, for the universal quantier and the interpretation
of the universal tree modalities. For instance, the proposition [#] evaluated at
a node n in structure T expresses a generalization not only over the dominated
nodes in the T itself, but0 also over the nodes dominated by any development
of n in a development T . Again, the fact that holds at all nodes dominated
by n in the current structure T does not mean that it will hold at all nodes
in some development of T even if the annotations of the nodes in T do not
change, because in a development of T new nodes may have been added that
are not present in T . Consequently, if we want the propositions holding at a
node of a partial tree to represent information that persists all the way to the
nal complete tree structure, then we cannot let T ; n j= [#] only to mean that
holds at all nodes in T dominated by n.
The connectives and operators with universal force are necessary when we
want to express invariances, universal principles which our partial trees should
satisfy. These principles pertain to the axiomatization of the tree structures
themselves, but also to invariances pertaining to Label and Formula values,
their interactions and co-occurrences. With respect to the latter case, we will
have to introduce axioms regulating the behaviour of the F o and T y predicates
on the trees. For instance, a node may be annotated by at most one type. This
requires a principle of the form:
8x8y((T y(x) ^ T y(y)) ! Eq(x; y)):
Furthermore, the T y values at the daughters of some node (and also the F o
values) have to be related to the values on those predicates at the node itself
8x8y((h#0 iT y(x) ^ h#1 iT y(x ! y)) ! T y(y)):
The structure of the tree is regulated by LOFT principles of the form
{ h#0i ! :h#1i
{ h#0i ! h#i
{ h#0i ! [#0]
The rst of these principles states that an argument daughter is never a function
daughter. The second expresses that an argument daughter is always a domi-
nated node, and the third that there is essentially only one argument daughter.
Denition 6 (Satisfaction II) In terms of the relation of extension between
partial trees, we now reinterpret the evaluation of atomic formulae and give
the interpretation of the missing universal connectives and operators. So we
continue the denition of satisfaction. Node n = [hl1; : : : ; lni : R] of Partial
Tree Structure T satises DU formula , with the notation
T ; n j= ;
if
= Lai (a) and a La li
= F o() and F o
i
= ! and for all T 0 ; n0 : T ; n T 0 ; n0

if T 0; n0; 0j= 0 then T 0; n00 j=0
= 8x and for all T ; n : T ; n T ; n and all t 2 D
T 0 ; n0 j= [t=x]
if # 2 f#0; #1; #; #g
= [#] and for all T 0; n0 : T ; n T 0; n0 and all n00 2 T rDom0
if T 0n0 0# T0 0; n00 then 0T 0 ;0n00 j= 00
= [# ] and for all T ; n : T ; n T ; n and all n 2 T rDom0
1
if T 0n00 # T 0; n0 then T 0 ; n00 j=
=? and 2 R
=?; and R = ;
Notice that we have a new denition for satisfaction of atomic propositions.
The new denition takes account of the possible update of Label and Formula
values, for instance the substitution of a proper name for a meta-variable. Now,
for instance, we have
T ; [hl1 ; : : : ; ei : John R] j= F o(U)
because U La John. The shape of this denition guarantees persistence of
atomic propositions, even though updates of values may have taken place. That
i
is, for atomic , if T ; n j= and T ; n T 0n0, then also T 0 ; n0 j= . The

example given is typical for cases where a personal pronoun introduces a meta-
variable U at a node (which thus satises F o(U)) all occurrences of which are
to be replaced at some later stage by an antecedent, for example the formula
John. Now, although that updated node is annotated by a declarative unit with
Formula John (and no longer with Formula U), still the node satises F o(U).
A requirement ? holds if the formula occurs on the requirement list of that
node, and ?; is a constant which holds at a node if it has an empty requirement
list.
On the basis of the persistence of atomic propositions, it can be shown that
all question-mark-free propositions persist. Of course, propositions preceded
by a question mark are not persistent, that is exactly the point; if everything
goes as it should, then requirements disappear. However, if ? holds at some
T ; n then _? is persistent { requirements can, indeed must, be replaced by
(persistent) annotations.19
We can introduce negation by the following denition.
: df ! ?:
These operators and modalities quantify not only over (nodes of the) current
(partial) decorated trees, but over possible developments of the current struc-
ture. For instance, the top node of the current decorated partial tree need not
be the root node of the eventual tree. Given that we have falsum ? (satised
by no node) and verum > (satised by all nodes) in our language, we can `close
o' the top node by expressing facts like du j= ["] ?, meaning àt all mother
nodes of du the formula ? holds' i.e. du has no mother node (and will have
none). This is an operation that takes place on a tree the moment all words
of the natural language string have been processed: the node that happens to
be the top one at that moment is turned into a root node. At the other end of
the tree, we can declare bottom nodes to be terminal nodes by annotating them
with [#]?. This is a task of lexical entries associated with words: a word closes
19 Partial tree structures are not themselves models for the Logic of Finite Trees (they are
not basic tree structures), but they can always be extended to such models. This means that
a LOFT axiom does not in general hold at a node n of structure T . But, by denition, any
T can be extended to a binary tree structure which satises . Because : has been dened
as ! ? and by the clause for implication in denition 6, this comes down to:
if is a LOFT-axiom, then T ; n j= :: for every partial tree structure T .
o a branch downwards. This distinguishes an intermediate structure such as
b c
which can be extended both downwards and upwards, from the nal structure
["]?
a; [#]?
b; [#]? c; [#]?
which can no longer be developed to a structure in which the present frontier
nodes become internal ones.
2.2.5.1 Internal and External Modalities
The denition of the notation T ; n j= for the static Boolean and modal oper-
ators is identical to the notion T ; n j= as given by denition 3. But on these
partial structures there are in eect kinds of each underspecied modality. Sup-
pose, for instance,
T ; n j= h# i:
Thus, by denition, there is some n0 such that n n0 and T ; n0 j= . This
n , however, may be otherwise wholly unconnected to the daughters and grand-
0
daughters of n. We know that the partial tree can be extended to one where
n0 is dominated by n, but that is all. Let's call h# i the internal version of the
dominance modality.
On a partial tree we can, however, also interpret a `real' relation of dominance.
Let us introduce the symbols # and # with the meaning
T ; n j=# i n i n0 for i 2 f0; 1; #g and T ; n0 j= ,
T ; n j=# i there is a nite sequence n i1 : : : i n0 for
ij 2 f0; 1; #; g and T ; n0 j= .
n
We will call these the external variants of the modalities h#i and h#i respec-
tively. Figure 2.9 shows a situation where the two variants of h#i dier in their
interpretation. In the Partial Tree T of that gure, evaluation at the top node
gives T ; [ T y(t)] 6j= h#iF o(John) but T ; [ T y(t)] j= # F o(John). Similarly,
T ; [ T y(t)] j= # F o(xy[Admire(x)(y)]) but not necessarily T ; [ T y(t)] j=
h# iF o(xy[Admire(x)(y)]). However, T ; [ T y(t)] j= h# iF o(Mary) and
T ; [ T y(t)] j= # F o(Mary) By denition of a partial tree structure as extend-
able to a basic tree structure, we know that T ; n j= h#i implies that every
extension T 0 of T (T T 0) has an extension T 00 such that T 00 ; n00 j=# : the
internal variant can always be extended to the external one.
[h: : : i : T y(t)]
[h: : : ; ei : Mary ] [h: : : ; ei : John ] [h: : : i : T y(e ! t)]
[h: : : ; e ! (e ! t)i : [h: : : i : T y(e)]

xy [Admire(x)(y )] ]
Figure 2.9: Internal and external modalities
The advantage of having these variants of modalities is that we can distin-
guish between an orthodox concept of dominate, the external modality, denoted
by `#', which is dened over xed and completed tree relations and is, in fact,
an underspecied description of a fully specied situation, and a novel concept
of dominate, the internal one, which is essentially tied to the partial trees, where
the tree relations themselves may be underspecied and can, in fact, be seen
as a constraint on the possible developments to complete binary trees. This
distinction between internal and external modality will of course collapse in the
case of a fully specied modality such as h"0i.
2.2.6 Descriptions of Tree Structures
With the DU language and the concepts of basic and partial tree structures to
hand, we can now see how this language can be used to describe the develop-
ment of tree structures. As a computational device, the (theoretical) parser we
will construct can be seen as a transition system where the states are DU de-
scriptions of partial tree structures and the transitions map DU descriptions of
partial tree structures to DU descriptions of extended structures. Node descrip-
tions will be nite and consistent sets of DU formulae, and tree descriptions
will be coherent families of node descriptions (these concepts will be dened be-
low). The tree descriptions and node descriptions will be central to our account,
since the parsing process will consist of derivational activity on descriptions of
partial tree structures interspersed with incremental lexical transitions between
descriptions. This activity eventually leads to a tree description which is isomor-
phic in some very specic sense to the basic tree structure it is a description of.
That is, the descriptions themselves will `turn into trees'. This construction of a
model out of a description, out of syntactic material, is a well-known technique
in mathematical logic. Often, to show that a set of formulae in some (formal)
language is consistent, i.e. has some model, this model is actually created from
the syntactic entities of the language itself (consistent sets of formulae become
domain elements, names become the objects they stand for, etc.).
Every state of the parsing process is a tree description : that is, every state is
a consistent family of coherent sets. In general, this consistency is not something
that has to be shown or concluded; it is something that is preserved by the
transition rules of our system. Moreover, if the natural language string is a
grammatical sentence, then the nal description is isomorphic to a complete
binary tree without requirements { that is, a term in the logical representation
language serving as the interpretation of the natural language string.
In this section we will introduce rst node descriptions and then tree descrip-
tions as collections of descriptions connected by relations dened on the basis of
their ìnternal' structure. That is, here we deal with the `states' of the transition
system. The next chapter will then introduce the `transitions' between these
states that are induced by parsing a natural language string.
2.2.6.1 Descriptions of Nodes
A declarative unit plus requirements [du R] can be completely described by a
nite set of DU formulae: the union of the set of atomic DU formulae satised
by du and a set of requirements satised by that unit, i.e. the DU formulae ?
such that 2 R. For instance,
[h: : : ; ei : John h"iT y(t)]
can be represented by
f: : : ; T y(e); F o(John); ?h"iT y(t)g.
This means that the nodes of a decorated tree can be represented by a nite
family of nite sets of DU formulae.
Let us call a nite set of atomic DU formulae and requirements representing
a declarative unit plus requirements a DU set. Not every nite set of atomic
DU formulae and requirements is a DU set. For instance, if a set contains both
T y(e) and T y(t), then it cannot represent a declarative unit: a nite set of DU
formulae must be consistent with all information about the labels; in particular,
it must satisfy specic co-occurrence restrictions. The co-occurrence of two
values on the type label is excluded; another instance is the interdependence
of the content formulae and the type labels. On the other hand, if a set of
DU formulae is `consistent' with the principles of declarative units, then it may
contain more than atomic DU formula and requirements. So also then, it will
not (strictly) represent a declarative unit plus requirements. A description of
the declarative unit du may contain facts that can be computed or derived from
the elements of the DU set corresponding with du. That is, a description of du
may contain Boolean combinations of atomic DU formulae and requirements.
But there is more. A node description may also contain formulae headed by
a tree modality. Thus, a node description may represent a node within a
tree environment. For instance, the set Z = fF o(); T y (X ); ?h#1 iT y (Y );
h#0 iF o( )g describes a node in the context of an argument daughter which
satises F o( ):
[h: : : ; X i : h#1 iT y(Y )]
[h: : : i : ] [ T y(Y )]
A tree description containing only the set Z would not be coherent. It should
contain another description which is consistent with being an argument daugh-
ter. Consistency of descriptions must now take into account the LOFT principles
which determine the Logic of Finite Trees. Thus, if a description states that
holds at an argument daughter (h#0 i) and that holds at an argument daugh-
ter (h#0i ), then the description of that argument daughter must be consistent
with ^ , for it is a LOFT principle that there is at most one such daughter.
Now we can give an informal denition of a node description, as a nite set of
formulae that can be satised by some declarative unit in a partial tree structure
such that no formula is both established and required.
2.2.6.2 Descriptions of Trees
In order to code the structure of a decorated partial tree T ; n in descriptions,
we will introduce a monadic tree node predicate T n and a pointer }. A formula
of the form T n(a) will be called a tree node identier. Node descriptions can
be connected by a tree relation through the presence in them of tree node
formulae in combination with the tree modalities of the DU language; this gives
a description of T . The pointer } is an element of exactly one node description,
and this node describes n. For instance, the tree T ; [h: : : ei : John], where T is
the tree
h: : : ei : John h: : : (e ! t)i : Admire(Mary)

can be represented by the family of sets of DU propositions:
f fT n(a); F o(Admire(Mary)(John)); T y(t)g;
fh"0iT n(a); F o(John); T y(e); }g;
fh"1iT n(a); F o(Admire(Mary)); T y(e ! t)g g.
This will be called a tree description. Given that this is a tree with one mother
node and two daughters, there are two node descriptions in this tree descrip-
tion that essentially contain modal propositions. These describe the relation
of the nodes under description to the root node. As these two modal proposi-
tions demonstrate, descriptions containing modalities no longer purely describe
a node decoration; they represent a node within some context. The interpreta-
tion of the tree node addresses is straightforward and is best exemplied by the
following picture:
fT n(a); : : : g
fT n(a0); : : : g fT n(a1); : : : g
fT n(a10); : : : g fT n(a11); : : : g
In a sequence a0, 0 represents the argument daughter of a; in a1, 1 represents
the function daughter of a. Thus a10 indicates the argument daughter of a node
that is the function daughter of a node with address a.
A description containing T n(0) represents a top node of a tree that will
remain a top node. When trees are constructed in the parsing process, in
general it is not known whether a description that starts o as a top node will
remain so (and thus be the root node of the eventual tree). This is why a new
node description is invariably introduced containing a formula T n(a), where a
is a constant not yet occurring in the construction. In interaction with the tree
modalities, various constellations are expressible. So, given the formula T n(a)
expressing the location of a node description in the tree under construction, we
can set
T n(a0) $ h"0 iT n(a),
T n(0a) $ h#0 iT n(a),
T n(a1) $ h"1 iT n(a),
T n(1a) $ h#1 iT n(a),
and we can x the root node of a tree as follows:
T n(0) $ ["]?:
The tree node formula T n(0) holds at a node if it is a top node and remains
so throughout all developments. Note the use of the `falsum' { Àt every node
above the current one, ? holds'. As ? is satised by no node at all (denition
3), this means that there are no nodes above the current one.
Here are some examples of addressing in the node descriptions:
fT n(a); : : : g. This set need not be a top node, but it could be.
fT n(a); ["]?; : : : g. This must be a top node, so
fT n(a1); ["]?; : : : g is inconsistent (after all, T n(a1) $ h"1 iT n(a)). When
all words of a string have been processed, the the current top node can be
declared the root node of the tree. That is, the top node can be closed o
by adding ["]?. Then, by denition, we have T n(0), and this is an update
of the proposition T n(a) previously present in the same description (for
we have a T n 0).
fT n(a); [#]?; : : : g. This must be a terminal node. Bottom nodes are
closed o downwards { that is, declared terminal nodes by the words of a
natural language string. Again, note the use of the falsum.
f: : : h" iT n(a); : : : ; ?9xT n(x)g. In this node description, the requirement
can be fullled only when the description is connected to the description
with address T n(a) by a sequence of 0; 1 steps: only then is there
a t 2 DT n such that the description satises T n(t). This means that
we can combine a structurally underspecied address, h"iT n(a), with a
requirement such that the node can only rid itself of all requirements by
becoming structurally fully specied. It is this characterization, using the
Kleene star operator, which expresses the weak tree relation between a
left-dislocated constituent and the node at which it is introduced.
Now, by representing a decorated tree as a family of node descriptions struc-
tured by the T n predicate and the tree modalities, we create the possibility
for computational and deduction rules which take the content of more than
one description into account. First, we introduce the notion of the language of
addresses as a sub-language of DU .
Denition 7 (Addressing of Node Descriptions) The Address Language
ADR is the smallest set such that
>; ?; T n(t) 2 ADR for t 2 DT n
if 2 ADR and h#i, [#] are tree modalities, then h#i; [#] 2 ADR.
Now we use the language ADR to dene tree relations between node descrip-
tions.20
Denition 8 (Tree Relations between Node Descriptions) If ND and
ND0 are node descriptions and # 2 f#0; #1 ; #; #g, then
ND # ND0
if there is a 2 ADR such that either 2 ND0 and h#i 2 ND, or 2 ND
and h# 1 i 2 ND0 . We set ADR(ND) = ADR \ ND. This set gives all
formulae from ND that are involved in the determination of its location in the
tree.
On the basis of this system of relating node descriptions by the tree relations we
can formulate what it is to be a tree description. To describe a basic or partial
tree structure T , we have to give a description for every node; and if we want
to interpret the relations we have dened above as the real tree relations, then
20 For fully explicit denitions, see denitions 22{24 in chapter 9.
the organization of the node descriptions within a tree description T D must be

guided by the fact that they have to be LOFT-consistent and by the following
coherency demands.
Coherency:
there is exactly one ND 2 T D such that } 2 ND,
ND 2 T D and h#i 2 ND imply that there is a ND0 2 T D; ND #
ND0 and 2 ND,
ND 2 T D and [#] 2 ND imply that for all ND0 2 T D, if ND # ND0
then 2 ND.
Now, just as decorated partial trees can be extended (along the relation ) to
richer trees, so descriptions of partial trees can be extended along (essentially)
the subset relation, but accommodating the satisfaction of requirements.
Denition 9 (Extensions of Descriptions) For tree descriptions T D and
T D0, we say that T D0 extends T D, notation T D T D0, if for the pointer
constant } indicating the node currently under consideration,
for all ND 2 T D there is a ND0 2 T D0 such that ND=f}g ND0 [f? j
2 ND g.0
So growth of tree structure on the level of descriptions reduces to simple set
containment, except that the pointer may0 have moved and requirements ? 2
ND may have turned into facts 2 ND .
In the next chapter, we will discuss a variety of rules mapping one descrip-
tion into another. Our parsing process will always deal with descriptions of
trees and not with trees themselves. In general, if description T D is mapped
to description T D0 by such a rule, then T D T D0. The natural language
words will test whether the node description in T D containing the pointer con-
stant } satises some nite number of DU propositions. For instance, a word
may test whether the node description ND in tree description T D satises the
requirement ?, or whether ND has an argument daughter satisfying T y(e)
(i.e. whether h#0iT y(e) holds). Starting from some basic set of DU formulae,
the natural language string will construct a family of sets of DU formulae. If
everything proceeds as it should and the string is grammatical, then the result-
ing family of sets satises the denition of a tree description. This description
is then isomorphic to a basic tree structure in which all requirements have been
fullled. In other words, the resulting description is isomorphic to an unreduced
lambda term.
3
The Dynamics of Tree

Building
In the previous chapter we introduced the data structures we will be using
to represent the results of parsing a natural language string word by word {
partial trees decorated with declarative units and requirements { and we also
introduced the language we shall use to describe these structures. So far our
description has been static, merely setting out the declarative structure of our
parser, the data structures it employs. Yet the relation between the decorated
trees arising in the course of a left-to-right parse is incremental { trees grow.
The discussion was inevitably abstract, since any particular type of labelling
describes only one aspect of the structure (formula, tree relation, type and
so on) and furthermore, possibly only at one stage of development. And, as
we shall see in this and subsequent chapters, when we apply the model, the
process of tree growth will be central to every explanation. In this chapter
we will introduce the procedural structure of the space of decorated trees; that
is, we characterize possible tree developments, and formulate transition rules
which license such developments. The dynamics of the progressive build-up of
decorated trees guided by lexical input will be formulated in terms of a system
of transitions which map (the description of) a partial tree to a new partial
tree (description) in accordance with general rules and the lexical specication
projected by words. The starting point of the construction process is a partial
tree consisting solely of a root node with a requirement for an annotation of
type t: this is the formal re ection of the conclusion drawn in chapter 1 that
utterance interpretation is a goal-directed task of constructing some intended
proposition. This goal is achieved when a Type expression of the form T y(t)
annotates the root node and all words of the natural language string have been
processed. The construction process involves the creation of new nodes in the
tree. Each node is introduced with requirements, which we might refer to as the
aspects of the node description which are yet `to do'. Requirements are fullled
by appropriate annotations on that node, aspects of the node description which
55
56 The Dynamics of Tree Building
we might describe as `done'. The collection of annotations and requirements on

a node, we shall call its decorations. A node is completed when all requirements
are fullled. The characterization of the overall dynamics then determines how a
sequence of lexical items induces a succession of states starting from the initial
tree, consisting of a single node, through a number of partial trees in which
additional nodes are introduced, to the completed decorated tree, representing
a logical form.
3.1 The Parsing Process { A Sketch
Before formally dening the rules which license transitions from tree to tree, we
shall introduce them by setting out, step by step, three examples. The rst is a
basic SVO clausal sequence, the second an instance of left-dislocation, and the
third is from a verb-nal language in which a ` at' sequence of noun phrases pre-
cedes the verb. As we shall see, the rules will license the development of a logical
form represented by a tree. They include rules that introduce requirements for
appropriately annotated (new) nodes, rules that regulate the incorporation of
lexical information in the tree, and rules propagating and compiling the infor-
mation upward in a tree.
3.1.1 A Basic Example
We will display a sequence of partial trees arising in the course of parsing the
sentence (3.1):1
(3.1) John admires him.
For each tree arising in the course of the parsing process, we display a set of
node decorations ordered by tree relations. As we have discussed, we use `?'
as a prex to distinguish decorations on a node which are requirements from
those which are annotations. In each tree (description) there is one node, the
àctive' one, that is highlighted by a pointer. Formulae of the DU language will
be evaluated with respect to the pointed node. In tree descriptions, the pointer
is a DU constant and has the form `}'. We shall also give a tree display, but
this will be essentially informal { merely showing the minimum information to
indicate the tree growth { typically only with a set of requirements, a formula
annotation where appropriate, plus the pointer. Reference must be made to the
accompanying tree description to see the full information available at each stage.
Step 1: The outset
The starting point re ects the requirement to create a decorated tree the top
node of which is annotated with a formula of type t. A description of this sit-
uation has to introduce an arbitrary identier for the tree node predicate. As
1 We will ignore tense.
The Parsing Process { A Sketch 57
the initial tree has only one node, the position of the pointer is determined.
Tree Description I
fT n(a); ?T y(t); }g
The tree node identier a constitutes the anchoring point for the addresses which
the other nodes in the tree will receive; their addresses will be dened relative
to this starting point.
Step 2: Expansion of Requirements, Introduction
To get a formula of type t, one possibility is to obtain two other formulae, one
of type e and one of type (e ! t). Once we have those, we can create our type
t formula by function application. This abductive move is re ected in the
following extension of description I:2
Tree Description II
fT n(a); ?T y(t); ?h#0iT y(e); ?h#1iT y(e ! t); }g
Of course, this is only one of the possible developments. Notice that in the
move from description I to II no new node descriptions are added: in this case
the extension consists in expansion of the decoration of a node, but only by
extending the requirements on that node:
fT n(a); ?T y(t); }g fT n(a); ?T y(t); ?h#0iT y(e); ?h#1iT y(e ! t); }g:
This we call an Introduction rule.
In this use of the Introduction rule we expand a requirement with premisses for
a rule of Modus Ponens (Elimination) which would give us this requirement as
a conclusion, hence successfully leading to the required annotation. But the
rule itself is more general. Given an n-ary operation O, whenever we need
object o, then we look for objects o1 : : : on such that O(o1 ; : : : ; on) = o. So the
requirement for o introduces requirements for o1 : : : on.
The Introduction rule sets the stage for application of the Prediction rule.
Step 3: Introduction of new nodes, Prediction
The requirements for an argument daughter annotated with a formula of type e
and a function daughter annotated with one of type (e ! t) now lead to a rule
called Prediction to introduce descriptions of an argument or function daughter
containing the appropriate requirements. Applying the Prediction rule twice
to description II, once with respect to ?h#0iT y(e) and once with respect to
?h#1iT y(e ! t), results in:
2 Recall that h#0 i is interpreted along the argument daughter relation, h#1 i along the func-
tion daughter relation.
Tree Description III

fT n(a); ?T y(t)g
f?T y(e); }g f?T y(e ! t)g

f fT n(a); ?T y(t); ?h#0iT y(e); ?h#1 iT y(e ! t)g
fh"0 iT n(a); ?T y(e); }g
fh"1 iT n(a); ?T y(e ! t)g g
At this stage tree description III has tree relations as dened in section 2.2.6.2.
It is a general property of the parsing process that it creates descriptions which
eventually become isomorphic to the trees they describe. So the operations are
operations on tree descriptions; but the descriptions are dened in such a way
that they correspond directly to a tree structure.
Notice that in this example a decision has been made as to where to locate
the pointer in the new structure { here the subject node. This is one point
where language-specic dierentiation with respect to word order can occur.
Other languages may have variants of Prediction in which the pointer is located
at the predicate node. The position of the pointer is essential for the processing
of the words from the natural language string by Scanning.
Step 4: Incorporation of lexical information, Scanning
Now, for the rst time, the string to be parsed enters the construction process:
the occurrence of the word John leads, by information associated with it in the
lexicon, to the following description:
Tree Description IV
fT n(a); ?T y(t)g
fF o(John); }g f?T y(e ! t)g

f fT n(a); ?T y(t); ?h#0iT y(e); ?h#1iT y(e ! t)g
fh"0 iT n(a); F o(John); T y(e); ?T y(e); [#]?; }g
The word John has been scanned. Scanning of a word means applying the
`program', the sequence of instructions, projected by that word to the current
partial tree description oriented by the location of the pointer.
Tree Description III John
7 ! Tree Description IV.
The mapping associated with John supplies (among other things) the formula
T y(e) to the left daughter, and thus the node contains both T y(e) and ?T y(e).
By a rule called Thinning, requirement ? can be abolished from a node de-
scription, if that description also contains the formula . The structure of a
word is a conditional action package of the form
IF all of 1 : : : n hold at the node description identied by the pointer,
THEN execute action 1,
ELSE execute action 2.
The word John tests for the presence of ?T y(e) at the pointed node, then it adds
the propositions F o(John) and T y(e), and closes o the tree node downwards;
that is, it creates a terminal node. If it does not nd ?T y(e), it aborts.
John IF ?T y(e) `holds for' description fh"0iT n(a); ?T y(e); }g,
THEN add F o(John), T y(e), and [#]? to the description,
ELSE ABORT.
The actions that a word is allowed to take are discussed in section 3.2.2. They
include the construction of new nodes, their decoration with annotations and
requirements, and the movement of the pointer through the tree. As in step 2,
this application of Scanning does not involve changes in tree structure.3 In this
step, the change in node decoration does not merely consist in the addition of
formulae, for a requirement ?(T y(e) will be removed by Thinning :
fh"0 iT n(a); ?T y(e)g 6 fh"0iT n(a); F o(John); T y(e)g:
However, it is clear that the loss of a question mark should be considered as
an incremental change: after all, becoming fullled is what requirements are all
about. This notion of growth is incorporated in the denition of of incremental
change (denition 5), where
fh"0 iT n(a); ?T y(e)g fh"0iT n(a); F o(John); T y(e)g:
In the tree description we have created up to now, we need a rule that propa-
gates the information that F o(John) and T y(e) are elements of the argument
daughter of the top node. There, the proposition h#0iT y(e)) must be added,
to get rid of the requirement ?h#0iT y(e). This is what a transition rule called
Completion will accomplish.
Step 5: Evaluating the tree, Completion

The presence of the formulae F o(John) and T y(e) contained in the description
of the argument daughter now leads, by Completion, to the addition of the
formulae h#0iF o(John) and h#0iT y(e) at the mother node, thus getting rid of
the requirement for the latter formula there.
3 Other application of Scanning, however, may also involve change of tree structure. We
will see an example of this below when we model the processing of the word admire.
Tree Description V
fT n(a); h#0 iT y(e); ?h#1 iT y(e ! t); ?T y(t)g
fF o(John)g f?T y(e ! t); }g

f fT n(a); ?T y(t); h#0iT y(e); ?h#1iT y(e ! t)g
fh"0iT n(a); F o(John); T y(e); [#]?g
fh"1iT n(a); ?T y(e ! t); }g g
Step 6: Scanning again
The transitive verb admire now maps tree V to one with a fresh tree node and
a new requirement. The mapping of description V to VI works as follows. The
word admire tests for the satisfaction of ?T y(e ! t) at the pointed node. If
it nds this, then it adds h#0 iT y(e) as a requirement and constructs a function
daughter which it annotates with T y(e ! (e ! t)); F o(Admire). The lexical
specication of admire can be dened as:
IF ?T y(e ! t) holds,
THEN put ?h#0iT y(e); and make h#1i;
go there and put F o(Admire); T y(e ! (e ! t));
close o with [#]? and go up one node,
ELSE ABORT.
The tree resulting from executing the actions projected by this word is:
Tree Description VI
fT n(a); ?T y(t)g
fF o(John)g f?T y(e ! t);

?h#0 iT y(e); }g
fF o(Admire)g
f fT n(a); ?T y(t); h#0iT y(e); ?h#1 iT y(e ! t)g
fh"0 iT n(a); F o(John); T y(e); [#]?g
fh"1 iT n(a); ?T y(e ! t); ?h#0 iT y(e); }g
fh"1 ih"1 iT n(a); F o(Admire); T y(e ! (e !)); [#]?g g
In displaying the trees, we tend to mimic the relative order of words in the lan-
guage, but nothing turns on this. Word order is not re ected in the structure
itself, only in the order in which lexical actions take place.
Step 7: Prediction
Prediction can develop the requirement for an object h#0 iT y(e) of admire into
an argument daughter of the VP node with requirement ?T y(e).
It should be noted that these Scanning and Prediction steps, induced by the
lexicon, can be mimicked by rule applications: by a repetition of Introduction
and Prediction now with respect to the ?T y(e ! t) requirement at the predicate
node. First we apply Introduction to
f?T y(t); ?h#1 iT y(e ! t)g
fT y(e); F o(John)g f?T y(e ! t); }g

which gives
f?T y(t); ?h#1 iT y(e ! t)g
f?T y(e ! t);

fT y(e); F o(John)g ?h#1 iT y(e ! (e ! t));
?h#0 iT y(e)g
Then we unfold the rst modal requirement at the predicate node by Prediction,
which results in the same tree structure as description VI. So there are choices
involved in what to assign to the lexicon and to general principles. And we will
see that individual languages dier in the way they divide the labour between
the lexicon and general principles.
Step 8: Processing the pronoun
The object him is then scanned in the same way as the word John. The action
connected to this pronoun is of the form:
IF ?T y(e);
THEN add F o(U); T y(e); ?9xF o(x); ?h"iT y(e ! t);
ELSE ABORT.
The variable that this pronoun introduces is accompanied by the requirement
for 9xF o(x), which is satised only by a full formula value (not another meta-
variable), and the accusative case projects a requirement for h"iT y(e ! t) which
is fullled if the node is a daughter of a predicate-type node.4 Executing this
action results in the description:
4 We ignore here the problem that prepositions also require this accusative form.
Tree Description VII

fT n(a)g
fF o(John)g
fF o(U)g fF o(Admire); }g
f fT n(a); ?T y(t)g
fh"0iT n(a); F o(John); T y(e)g
fh"1iT n(a); ?T y(e ! t)g
fh"0ih"1 iT n(a); F o(U); T y(e); ?9xF o(x); ?h"iT y(e ! t); }g
fh"1ih"1 iT n(a); F o(Admire); T y(e ! (e ! t))g g
Step 9: Substituting the variable, Substitution
The string has supplied the variable U. In order to full the requirement (and
nish the parse without outstanding requirements), a value must be found for
U. We must substitute a proper formula of type e. The choice of such a value
is not determined by the natural language under consideration (in this case En-
glish), but requires other, external, pragmatic principles. Here is the interaction
site where the underspecied projection of a natural language string can be sup-
plemented. Although a choice of value for a pronoun is not regulated by the
language proper, language principles may restrict such pragmatic choices. In
structure VII, for instance, John is not a possible substituend for U. We will
suppose the value Bill is chosen, relative to whatever external considerations
ensure it is an appropriate choice, by the action
IF Bill is a co-argument,
THEN ABORT,
ELSE Substitute Bill for U:
The substitution then gives:
Tree Description VIII
T n(a)
fF o(John)g f?T y(e ! t)g

fF o(Bill); }gfF o(Admire)g
f fT n(a); ?T y(t)g
fh"1 iT n(a); ?T y(e ! t)g
fh"0 ih"1 iT n(a); F o(Bill); T y(e); ?h"iT y(e ! t); }g
fh"1 ih"1 iT n(a); F o(Admire); T y(e ! (e ! t)); [#]?g g
Step 10: Evaluating the tree, Completion and Elimination
With the terminal nodes of the tree annotated, we can evaluate the tree. From
the perspective of the predicate node, given the annotations on both daughters,
we can conclude h#0iF o(Bill) and h#1iF o(Admire). Now the annotations at the
root node can be subjected to (our version of) -reduction: this is a rule called
Elimination and is a form of labelled Modus Ponens dened over combined
Type:Formula information. This reduction is re ected in the implications
[h#0iF o(Bill) ^ h#1iF o(Admire)] ! F o(Admire(Bill))
and
[h#0iT y(e) ^ h#1iT y(e ! (e ! t))] ! T y(e ! t):
Thus, the result of applying the Formula value of one daughter to that of the
other (-reduction) is deposited at the mother node. There it satises the type
requirement decorating that node. This process of Completion and Elimination
enables us also to evaluate the root node as F o(Admire(Bill)(John)); T y(t).
Now, having no more words to process and having evaluated the tree to the
top node, this node is declared the root node of the tree by adding the annota-
tion T n(0) (equivalently ["]?), thus closing o the tree from above. By the tree
principles of section 2.2.6.2, the presence of this proposition at the top node al-
lows the addition of, for instance, h"0iT n(0) to the argument daughter, and this
is equivalent to T n(00) according to the principles of that section. Thus, the
addresses of the nodes in the description are propagated. So at this juncture,
all the tree node identiers can be updated, replacing the term a in T n(a) as
previously described:
Tree Description IX
fT n(0); T y(t); }g
fT n(00); F o(John)g fT n(01)g
fT n(010); fT n(011);
F o(Bill)g F o(Admire)g
f fT n(0); F o(Admire(Bill)(John)); T y(t); }g

fT n(00); F o(John); T y(e); [#]?g
fT n(01); F o(Admire(Bill)); T y(e ! t)g
fT n(010); F o(Bill); T y(e); [#]?g
fT n(011); F o(Admire); T y(e ! (e ! t)); [#]?g g
We turn to the details of these rules in the next section; but notice how the inter-
pretation process involves an assignment of a value to the formula projected by
the pronoun, and not simply a co-indexing restriction. Thus context-dependent
choices are modelled as part of the development of the tree.
3.1.2 A Left-Dislocation Example

The set of steps used in example 3.1 induces a simple { indeed trivial { tree and
its description. The description has shown how the tree can expand through
requirements for daughter nodes, which are introduced and then annotated. At
each non-nal stage, a partial tree is described; but, so far, the partiality of
the tree description has resided simply in the fact that nodes are not completely
annotated; that is, until the nal step, requirements decorate at least one node in
the tree. However the Logic of Finite Trees (LOFT) allows us a richer freedom in
setting out partial tree descriptions, for we can introduce nodes, annotate them
and full requirements on them, while nevertheless describing their relation to
the remainder of the tree as unxed { for instance, as dominated by some node
within that tree without any full specication. We shall use this freedom to
characterize long-distance dependency by introducing a tree node in the tree
merely as dominated by the top node, leaving a fuller specication until later
in the tree construction process. We take a simple example of dislocation, for
example (3.2):
(3.2) Bill, John admires.
In parsing a string such as (3.2), the information presented by Bill at the rst
position in the string is not suÆcient to x how it is to contribute to the overall
logical form. So, though the starting point of the parse for (3.1) and (3.2) is
identical, the subsequent steps are dierent.5
Step 1: The outset
As before, we have tree description I, the starting point of the parse. In fact,
all parse representations will have the same starting point.
Tree Description I
Step 2: Adjunction of a dominated node
Now we allow the introduction of an unxed node, as follows:
Tree Description II0
fT n(a)g
f?T y(e); }g
f fT n(a); ?T y(t)g
fh" iT n(a); ?T y(e); ?9xT n(x); }g g
This describes a truly partial tree, because it contains an connection that
cannot be analysed in a sequence of 0, 1 or # connections. This move is an
5 Shortly, we shall return to the issue of whether these two modes of interpretation should
be distinguished at this early point in the interpretation process.
example of the rule of *Adjunction. We have created a location for an argument
in the term under construction, without specifying where this argument is to be
located within the sub-term structure. This underspecication with respect to
its location is witnessed by the requirement ?9xT n(x), which is only satised
once the node has a fully specied location { that is, for T n(x), where xf0; 1g.
This extension of tree description I, the starting point, is an alternative to
the development shown in tree description II of the previous example. The
alternatives present for our parser are determined by the natural language under
observation
tree description I
tree description II or tree description II0 or : : : or : : :
where tree description II results from description I by an application of the In-
troduction rule, whereas tree description II0 results from the same description I
by applying the *Adjunction Rule.
Step 3: Scanning
Now the word Bill maps this structure onto one where the unxed node receives
F o(Bill) and T y(e) and the pointer has returned to the top.
Tree Description III
fT n(a); }g
fh" iT n(a); F o(Bill)g

f fT n(a); ?T y(t); }g
fh"iT n(a); F o(Bill); T y(e); ?9xT n(x)g g
Step 4: Prediction
Having processed Bill, the parse of the remainder John admires proceeds as if
the complete string starts with John. We take the present tree to be the outset
and proceed as in the previous example. That is, we introduce a subject node
and a predicate node by the rules of Introduction and Prediction:6
Tree Description IV
fT n(a)g
fh" iT n(a); F o(Bill)g f?T y(e); }g f?T y(e ! t)g

f fT n(a); ?T y(t); ?h#0iT y(e); ?h#1 iT y(e ! t)g
fh" iT n(a); F o(Bill); T y(e); ?9xT n(x); [#]?g
fh"0 iT n(a); ?T y(e); }g
6 We retain the tree node identier of the unxed node in the tree display for clarity.
and we proceed exactly as in the previous example, processing consecutively the

words John and admires. This results in the structure
Tree Description V
fT n(a)g
fh" iT n(a); fF o(John)g
F o(Bill)g
fF o(Admire)g f?T y(e); }g
f fT n(a); ?T y(t)g
fh" iT n(a); F o(Bill); T y(e); ?9xT n(x); [#]?g
fh"1 ih"1 iT n(a); F o(Admire); T y(e ! (e ! t)); [#]?g
fh"0 ih"1 iT n(a); ?T y(e); }g g
Now the string is nished, but the structure still has a requirement outstanding
at a terminal node (the last node description). On the other hand, there is
surplus material in the unxed node (the second node of the description above).
Now, merging the unxed node with the pointed terminal node gives the nal
structure. In a tree description, merging of two nodes simply comes down to
taking the union of the corresponding descriptions:
fT n(a)g
fF o(John)g
fF o(Admire)g fh" 0 ih"1 iT n(a);

h" iT n(a); F o(Bill); }g
f fT n(a); ?T y(t)g
fh"0 iT n(a); F o(Bill); T y(e); [#]?g
fh"0 ih"1 iT n(a); F o(Admire); T y(e ! (e ! t)); [#]?g
fh"1 ih"1 iT n(a); h"iT n(a); F o(Bill); T y(e); ?T y(e); ?9xT n(x); [#]?; }g g
Notice how this tree has one terminal node, the object node, with two address
formulae, h"1ih"1iT n(a) and h"iT n(a). However, because the principle
h"1 ih"1 iT n(a) ! h" iT n(a)
is a LOFT theorem, the proposition h"iT n(a) from the last node description
does not carry any new information
(h"1 ih"1iT n(a) ^ h"iT n(a)) $ h"1ih"1iT n(a):
The address h"1ih"1iT n(a) is equivalent to T n(a11), and this allows us to con-
clude 9xT n(x); so the requirement for a xed tree position is satised. Thus
the description of the unxed node disappears while injecting its content in an
appropriate terminal node.
Having annotated all the terminal nodes, steps of Completion and Elimina-
tion apply successively enabling annotations to be added to non-terminal nodes
as before, so that the whole tree is evaluated.
The tree description resulting after the merge operation is identical to that
assigned to (3.1), showing how two dierent strings can lead to one and the
same logical form representation. The use of the Kleene star operator makes
the incremental construction of a logical form essentially dynamic, for a node
in the tree at one point may get some more precise tree relation later in the
parse. It is this dynamic use of the Kleene star operator which is novel to our
framework and which we will develop in the characterization of long-distance
dependency phenomena in general.
3.1.3 Verb-nal Languages and the Grammar-parser Prob-
lem
It might seem that this model of processing, even as so far set out, is doomed to
failure as dening a general grammar formalism, for the types of rules indicated
apply in a top-down way; and this, though possibly suitable for projecting
structure for an SVO language such as English, while following the left{right
dynamics of parsing, would seem to be entirely unsuitable for languages with
other constituent orders. In particular, verb-nal languages, with their sequence
of noun phrases all preceding some verbal complex containing possibly more
than one verb might seem to pose severe problems for the general research
strategy being advocated here:
(3.3) Hiroto ga Masa ga ringo o tabeta to itta [Japanese]
HirotoNOM MasaNOM appleACC ate comp said
`Hiroto said Masa ate an apple.'
By consensus in recent years, it is assumed that verb-nal languages require
assignment of congurational structure along the same lines as all other lan-
guages;7 yet they seem to require a means of parsing strikingly dierent from
head-initial languages such as English, hence a radical dierence between gram-
mar and parser formalisms (see Inoue and Fodor 1995). There is a further
problem posed by Japanese, from the parsing perspective, which appears to
7 Languages which display relative freedom of major constituent order have been subject
to ongoing debate as to whether or not they require structural analysis in terms which re ect
the at sequence of noun phrases, or, contrarily, the congurations that need to be postulated
to re ect the mode of combination of the constituent parts of a sentence. However agreement
(at least within the Minimalist/GB framework) has settled on the conclusion that at any
syntactic level dened as expressing structural generalizations about language, all languages
display the same structural properties. See Hale 1983, Speas 1990, Inoue and Fodor 1995,
Gorrell 1995, Kiss 1991.
buttress this conclusion. A parser of a Japanese string faces a notorious prob-

lem of ambiguity. There is no certain identication of a sequence of NPs until
the full sequence of verbs is parsed as to which of them belong to the main
clause, or a subordinate clause, or a relative clause; and even at the end of
the parsing task ambiguity may still remain. Even in an example such as (3.4)
in which the objects described to a large extent disambiguate the object and
subject relations depicted, there remains considerable ambiguity, as it may be
unclear whether Hiroto is subject of the verb tabeta in the subordinate clause,
subject of the main verb itta, or both:
(3.4) Hiroto ga ringo o tabeta to itta
HirotoNOM appleACC ate comp said
either `Tom/Dick/Hiroto/Akiko ... said that Hiroto ate the apple.'
or `Hiroto said that Tom/Dick/Hiroto/Akiko ... ate the apple.'
Consideration of this ambiguity problem suggests one of two conclusions. Either
parsing in Japanese has to involve back-tracking if decisions made early on in the
parsing process con ict with incoming lexical material (as advocated by Inoue
and Fodor 1995); or the structure to be assigned, hence the interpretation, is
projected relatively late, with successive constituents parsed as a at sequence
to be mapped onto congurational structures determined by the grammar at
some late stage in the interpretation process (see Gorrell 1995). Either way, we
seem confronted with the conclusion that the parsing and grammar formalisms
are clearly disjoint.
In fact, however, we shall see that the emphasis on underspecication of tree
relations allows us to describe languages such as Japanese surprisingly straight-
forwardly, providing a basis for showing the commonalities and dierences be-
tween head-initial and head-nal structures. Furthermore, it will account for
the fact that languages such as Japanese provide a at sequence of noun phrases
as input to the structure-building process of interpretation, from which familiar
congurational representations are nonetheless projected. This approach will
thus provide a new slant on the debate as to whether such languages are or are
not congurational. We assume that the parsing device itself is uniform and
incremental, and that the resulting tree congurations are identical.
One thing that is certain, if we are to characterize languages such as Japanese
and English in similar terms, is that in Japanese we must allow the tree struc-
ture projected by the expressions to be fundamentally underspecied in all early
stages of the parsing process. And this is what we shall do. We shall assume
that interpretation in Japanese is a top-down process only very shallowly. Con-
stituents will project unxed nodes, and associated case specications serve
to narrow down the tree positions they can occupy. We shall further assume
that interpretation is driven by indicators of structure such as verbs (which oc-
cur nally in any clausal sequence) and morphological indicators of boundaries
between one structural domain and another (such as clausal boundaries).
In this sense, the projection of structure, i.e. parsing, is head-driven.8
We assume, as before, the Axiom which seeks to establish a root node dec-
orated by a formula of type t; but now we have no a priori development of
structure. The responsibility for tree construction (starting from the Axiom)
is squarely placed on the side of the lexicon. The lexical actions projected by
nouns, noun phrases and verbs will all look for nodes requiring type t, and induce
structure from the root directly. So the lexical actions do not need structural
tree context to be expressed, but they may impose contextual requirements on
the nodes they annotate { indeed, case requirements are dened precisely to
narrow down the space within which nodes may be identied. In any comple-
tion to a structure which fullls all such requirements, the nodes will have found
their intended location.
The lexical entries for verbs have conditions of the form:9
IF ?T y(t) holds,
THEN make(h# i), put there ?T y(t) and construct `verb frame',
ELSE : : :
Here, the location of the new node constructed is dened by means of the tree
relation . This relation is weak, because w i w0 implies w w0 whenever
i 2 f0; 1; #; g. The actions undertaken by the (projection of the) verb then
set out a tree skeleton in which the DU propositions projected by NPs have to
nd a location.
Nouns and noun phrases project lexical actions of the form
IF ?T y(t) holds,
THEN make(h# i) and put there F o(); T y(e) and : : : ,
ELSE IF ?T y(e) holds,
THEN : : :
ELSE
So for example, the proper name Hiroto:
Hiroto IF f?T y(t)g,
THEN make(h# i); put(F o(Hiroto); T y(e)),
ELSE IF ?T y(e), .
THEN put(F o(Hiroto));
ELSE ABORT
8 Inoue and Fodor (1995) argue that interpretation cannot be treated as head-driven, be-
cause of multiple ambiguities internal to noun phrases. If, however, we assume that genitives
are indicators of relative-like structures (what in chapter 4 we shall call LINK ed structures),
the no genitive marker being an indicator of such a LINK ed structure, then the critical ex-
amples which led them to conclude that interpretation cannot be head-driven and furthermore
has to allow back-tracking are no longer problematic.
9 We assume that the process of building an unxed node providing the local root of a tree
is also induced by the lexicon in order to keep the spread of alternative parse trees as small as
possible. make(h#i) is an instruction to build a tree-node along the tree-relation # relative
to the current node.
In other words, noun phrases may either occur with an (as yet) completely
unspecied role in the developing logical form, or they may occur at positions
where their (combinatorial) role is strictly circumscribed.
We will give an example involving one level of subordination to show that
even the structure projected from verbs must be taken to be initially unxed
with respect to the ultimate root:
(3.5) Hiroto ga ringo o tabeta to itta
HirotoNOM appleACC ate COMP said
either `Tom/Dick/Hiroto/Akiko ... said that Hiroto ate the apple.'
or `Hiroto said that Tom/Dick/Hiroto/Akiko ... ate the apple.'
The rst step is common to all parse sequences: we start again from the Axiom,
but this time we x from the beginning that this will be the root node of the tree.
Tree Description I
fT n(0); ?T y(t); }g
The subsequent step is then an execution of the actions associated with the NP
Hiroto. This results in the following tree:
fT n(0)g
fh" iT n(0); F o(Hiroto); }g

f fT n(0); ?T y(t)g
fh" iT n(0); F o(Hiroto); T y(e); ?9xT n(x); }g g
The marker ga, indicating nominative case then adds to this annotated node a
requirement ?h"0iT y(t).10
IF fT y(e)g,
THEN put(?h"0 iT y(t)); gofirst"(T n(0));
ELSE ABORT.
Notice that this case marker needs a finished node to apply to (i.e., its condition
is T y(e) instead of ?T y(e)).
fT n(0); }g
fF o(Hiroto); h"0 iT y(t)g

f fT n(0); ?T y(t); }g
fh" iT n(0); F o(Hiroto); T y(e); ?9xT n(x); ?h"0 iT y(t)g g
10 gofirst" is a macro dened as a search upwards for the rst occurrence of some given
label or formula value. Once found, the pointer rests there (see section 3.2.2).
With the node completely annotated, the projection of ga now returns the
pointer to the root node. Notice that the case marker itself is not an instruction
to construct a node; it is merely a restriction on the relative position of the node
in some tree structure.
The next word, the common noun ringo (àpple'), again annotates a node
with Fo(Ringo) and T y(e) related to the root by , and builds a sub-tree:
fT n(0)g
fF o(Hiroto)g f?T y(e); }g
fF o(P (; P ))g
fF o(U)g fF o(Ringo)g
f fT n(0); ?T y(t)g
fh"iT n(0); F o(Hiroto); T y(e); ?h"0 iT y(t); ?9xT n(x)g
fh"iT n(0); ?T y(e); ?9xT n(x); }g
fh"1ih" iT n(0); T y(cn ! e); F o(P (; P ))g
fh"0ih" iT n(0); ?T y(cn)g
fh"0ih"0 ih" iT n(0); T y(e); F o(U)g
fh"1ih"0 ih" iT n(0); T y(e ! cn); F o(Ringo)g g
The building of this local sub-tree with root node requiring T y(e) is induced by
actions associated with the common noun itself, because Japanese is a language
in which nouns need no morphological determiner. The resulting structure is
nevertheless, in default of any quantier, an existentially quantied expression.
As with the rst noun phrase sequence, the case suÆx o adds a requirement
to the root node of this locally constructed tree, here 11h"0iT y(e ! t), and the
case marker then returns the pointer to the root node.
fT n(0)g
F o(Hiroto) f?h"0 iT y(e ! t); }g
fF o(P (; P ))g
fF o(U)g fF o(Ringo)g
11 Here we ignore the dependency predicate and the selection of one of its arguments to
x the scope of the quantied term, though it should be noted that this sequence of unxed
nodes provides a good base for modelling the observation (Hoji 1986) that in Japanese scopal
choices are very largely driven by linear order, indenites apart (see chapter 7).
f ffT n(0); ?T y(t)g

fh" iT n(0); F o(Hiroto); T y(e); ?h"0iT y(t); ?9xT n(x)g
fh" iT n(0); ?T y(e); ?9xT n(x); ?h"0iT y(e ! t); }g
fh"1 ih" iT n(0); T y(cn ! e); F o(P (; P ))g
fh"0 ih" iT n(0); ?T y(cn)g
fh"0 ih"0 ih" iT n(0); T y(e); F o(U)g
fh"1 ih"0 ih" iT n(0); T y(e ! cn); F o(Ringo)g g
We dene verbs in essentially the same way. They are taken to construct a node
decorated with a requirement ?T y(t) dominated by the root node, and induce
a full `verb-frame', with argument places decorated by meta-variables, exactly
as though a sequence of pronouns had been supplied along with the verb. The
verb tabeta (actually tabe plus ta, the past tense marker meaning àte') rst
constructs a node requiring type t; and, having constructed such a node, denes
the verb frame:
IF f?T y(t)g
THEN make(h#i); put(?T y(t); ?9xT n(x));
make(h#0 i); put(F o(V); T y (e); ?9xF o(x)); go(h"i)
make(h#1 i); put(?T y (e ! t));
make(h#0 i); put(F o(W); T y (e); ?9xF o(x)); go(h"i);
make(h#1 i); put(F o(T abeta); T y (e ! (e ! t))); gofirst"(?T y (t))
ELSE ABORT
As with the common noun, the motivation for dening this projection of struc-
ture from the lexicon is that Japanese is a free pro-drop language in which sen-
tences containing just a verb are well-formed sentences, with subject and object
construed from the context. Notice how, even though these actions determine
a parsing process which is lexically driven (and in this sense bottom-up), local
structure is nevertheless built by a sequence of actions dened in the lexicon on
a top-down basis:
fT n(0)g
fF o(Hiroto)g fF o(; x; Ringo(x))g f?T y(t); }g
fF o(V)g fh"1 ih" iT n(0)g
fF o(W)g fF o(T abeta)g
f fT n(0); ?T y(t)g
fh"iT n(0); F o(Hiroto); T y(e); ?h"0 iT y(t); ?9xT n(x)g
fh"iT n(0); F o(; x; Ringo(x)); T y(e); ?h"0iT y(e ! t)g
fh"iT n(0); ?T y(t); ?9xT n(x); }g
fh"1ih" iT n(0); T y(e ! t)g
fh"0ih" iT n(0); T y(e); F o(V); ?9xF o(x)g
fh"0ih"1 ih" iT n(0); T y(e); F o(W); ?9yF o(y)g
fh"1ih"1 ih" iT n(0); T y(e ! (e ! t)); F o(T abeta)g g
(In this tree display, we assume for simplicity that steps of Completion and
Elimination have taken place fully annotating the sub-trees projected by the
words Hiroto and ringo.) At this point, enough material has been collected
to create a complete decorated binary tree, i.e. a logical form. The meta-
variables which function as arguments for the predicate tabeta can be assigned
a value exactly as would a meta-variable projected from a morphological pro-
noun (as in the rst example). In particular, we can merge the nodes dec-
orated by F o(Hiroto) and F o(; x; Ringo(x)) with those decorated by F o(V)
and F o(W) respectively. We might also allow the nodes decorated by ?T y(t)
to be merged, yielding a fully specied structure with a root annotated by
F o(T abeta(; x; Ringo(x))(Hiroto)) (= `Hiroto ate an apple'). But, as the string
is not yet nished, this last step may well turn out to be premature.
The suÆx to now adds to the root node of this sub-structure a requirement
for a mother of type e ! t (i.e. its position in the eventual tree must be that
projected by a subordinate clause), and returns the pointer to the root. Now
the verb itta (`said') is dened as building a sequence of nodes in like manner
to tabeta, except that the argument daughter of its predicate node is required
to be of ?T y(t):
IF f?T y(t)g
THEN make(h#i); put(9xT n(x));
make(h#0 i); go(h#0 i); put(F o(X); T y (e); ?9xF o(x)); go(h"i);
make(h#1 i); go(h#1 i); put(?T y (e ! t));
make(h#0 i); go(h#0 i); put(F o(Y); T y (t); ?9xF o(x)); go(h"i);
make(h#1 i); go(h#1 i); put(F o(Itta); T y (t ! (e ! t)));
gofirst"(?T y (t))
ELSE ABORT
fT n(0); ?T y(t)g
fF o(H )g fF o((; x; R(x)))g f?T y(t)g f?T y(t); }g
fF o(V)g fh"1 ih" iT n(0)g fF o(X)g fh"1 ih" iT n(0)g
fF o(W)gfF o(T abeta)g fF o(Y)g fF o(Itta)g
f fT n(0); ?T y(t)g
fh" iT n(0); F o(Hiroto); T y(e); ?h"0iT y(t); ?9xT n(x)g
fh" iT n(0); F o(; x; Ringo(x)); T y(e); ?h"0iT y(e ! t)g
fh" iT n(0); ?T y(t); ?9xT n(x)g
fh"1 ih" iT n(0); T y(e ! t)g
fh"0 ih" iT n(0); T y(e); F o(V); ?9xF o(x)g
fh"0 ih"1 ih" iT n(0); T y(e); F o(W); ?9yF o(y)g
fh"1 ih"1 ih" iT n(0); T y(e ! (e ! t)); F o(T abeta)g
fh" iT n(0); ?T y(t); ?9xT n(x); ?h"0iT y(e ! t)g
fh" iT n(0); ?T y(t); ?9xT n(x); }g
fh"1 ih" iT n(0); T y(e ! t)g
fh"0 ih" iT n(0); T y(e); F o(X); ?9xF o(x)g
fh"0 ih"1 ih" iT n(0); T y(e); F o(Y); ?9yF o(y)g
fh"1 ih"1 ih" iT n(0); T y(e ! (e ! t)); F o(Itta)g g
Note: F o(H ) abbreviates F o(Hiroto)
F o(R) abbreviates F o(Ringo)
Having processed the entire string, the resulting tree structure must now be
turned into a complete binary tree without outstanding requirements. This
comes down to supplying the meta-variables with proper values (getting rid of
the requirements of the form ?9xF o(x)) and nding a xed tree location for all
nodes with underspecied addresses (getting rid of the requirements ?9xT n(x)).
The meta-variables providing the decorations can be replaced either through
anaphora resolution or by merging their nodes with those projected by ringo or
Hiroto; in all cases these choices may be driven largely by externally imposed,
pragmatic considerations. We have here already assumed that the nodes pro-
jected by ringo or Hiroto do indeed merge with the subject and object nodes
projected from tabeta, though this is by no means the only option. The in-
terpretation `Hiroto said that he ate the apple' can then result from anaphora
resolution, which replaces the meta-variable X annotating the subject node as-
sociated with structure projected by itta with a copy of the Formula Hiroto.
Then a completed annotation of the node decorated by F o(Y); T y(t), is pro-
vided by merging it with the sub-tree projected by Hiroto ga ringo o tabeta
to. Finally, the node requiring type t introduced by itta is xed, and steps of
Completion and Elimination then lead to all non-terminal nodes duly receiving
an annotation. One last step is the merging of the root node, T n(0) and the
node:
h" iT n(0); F o(Itta(T abeta((; x; Ringo(x))(Hiroto))(Hiroto))); T y(t)
thus nally xing all node relations in the tree. Other choices of construal of
the meta-variables could have been made to full the requirements, giving rise
to alternative interpretations.
The advantage of using the concept of (sequences of) unxed nodes in charac-
terizing a language such as Japanese is that it allows a straightforward re ection
of the observation that such languages, in one sense, display a at structure and,
in another, a congurational structure as languages such as English.12
In these examples we have introduced the construction of a logical form at
a leisurely pace. As the informal account of English and Japanese sentences
has indicated, though the specic lexical characterizations dier from language
to language { with dierent load carried by general rules and the lexicon { the
causality of the parsing process is essentially the same. The initial state is a
root node, with a requirement for a type t object. Each stage of the construc-
tion process corresponds to a decorated partial tree. As the tree structure is
induced, its nodes are progressively decorated with labelled terms, fullling im-
posed requirements as the words are successively scanned. Any change to a
given partial tree is an enrichment; nothing is ever cancelled en route. Only
requirements can disappear, and these only by being fullled. The nal state is
the return to the root node with no requirements left at any node. Furthermore,
the resulting structures are identical across the dierent types of language { bi-
nary trees whose types of annotation and modes of composition are common to
all languages.
It is worth noting, even at this early stage, the general way in which this
system diers from other, more familiar systems. Modelling the parsing process
as an evolving tree is not in itself new { indeed it is denitional of a parsing
system that it must dene transitions across partial descriptions of analyses of
a string (see Marcus et al. 1983). What is novel about this system, as indicated
in chapter 1, is that the system denes possible transitions from one tree to
another without reference to any independent set of axioms that constitute
the grammar of the language. To the contrary, the transition rules themselves
incorporate the grammar. They make available a concept of well-formedness for
a given language through the requirement that the words of a string contribute
in sequence to the construction of a logical form. The system will thus provide
a putative grammar formalism, our major task in this book is to explore the
viability of this idea.
12 See Hale 1983, Kiss 1991 and Diesing 1992 for the debate on whether these languages are
congurational.
3.2 The Parsing Process Dened

We are now in a position to address the dynamics of the individual transitions
in a more formal way. The object of the parsing process is to construct a binary
tree without requirements, the top node of which is decorated by a formula
of type t by using all information in a natural language string. The set of all
these complete binary tree structures we will denote by Goal. This construction
process is controlled by tasks. Every task is centred around the description of a
particular node with a requirement to cull from the string some object of a given
type. Tasks may generate sub-tasks (located at daughter nodes), the overall task
(associated with the top node) being the construction of an object of type t. In
the course of a parse, a task goes through a sequence of task states: the various
stages, descriptions, of the node in the course of the parse. Every task state
consists of two kinds of formulae. First, there are DU formulae without question
marks, representing the information that has been constructed, achieved, the
`done' part of the task state. Secondly, it contains DU formulae headed by a
question mark, standing for the requirements that are still to be fullled, the
part of the task state as yet `to do'. The initial state of a task has in the
`done' part an address formula, the tree location of that task.13 The `to do'
part contains the single requirement for a proposition of the form T y(t); that is,
for the construction of an object of type t. A nal task state is a node without
question marks: all requirements have been fullled. For instance, the minimal
element, the starting point of every parse is the Axiom
Axiom: fT n(a); ?T y(t); }g.
Here T n(a) gives the location of the node in the tree, and ?T y(t) stands for
the requirement for an object of type t at this node, and } represents that this
node is the location of the pointer. The Axiom is the initial state of the initial
task. The initial task is associated with the description of a (putative) top node
as the only node of a tree. As the only description present, it is the location of
the pointer. In the course of a parse starting from the Axiom, (the description
of) a tree is created. A parse ends essentially after the last word of a natural
language string has been processed. A successful parse of a grammatical natural
language string must result in (a description of) a logical form. That is, it must
end with a tree description T D where all task states are in their nal state, i.e.
T D 2 Goal. Moreover, if a node description ND is a terminal node in T D, then
[#]? 2 ND (`there are no daughters and there never will be'). Together with
the fact that T N (0) entails ["]? (`there is no mother, and there will never be'),
we have a tree description that can no longer be extended with descriptions of
new nodes. Thus an extension of this tree description can only consist in the
addition of information.
Thus, a goal T D 2 Goal has the form:
Goal: f fT n(0); : : : ; T y(t); F o(); }g; fT n(00) : : : g; fT n(01) : : : g : : : : g
13 Recall that ADR is the Address Language for describing tree node identiers (see chapter
2, denition 7).
The Parsing Process Dened 77
This is a tree description all node descriptions of which are nal task states and
with the pointer at the root node.
If the procedure that leads from Axiom to Goal is sound, then the nal
description represents { is isomorphic to { an unreduced term in the language
of our logical forms.
By parsing a grammatical natural language string, the Axiom is connected
to a goal in Goal by a nite sequence of tree descriptions. Each description
T Di+1 in this sequence is a development of the previous description T Di; that
is, T Di T Di+1. So the construction of a complete binary tree (2 Goal) can
be seen as a process moving along a path of the kind
Axiom = T D0 T D1 : : : T Dn 2 Goal:
In the previous chapter we discussed the goal-directed space of decorated par-
tial tree structures, P T Rs, in which such sequences are located. We will use
the acronym PPT to denote this space. In this chapter we will introduce the
second half of the picture, the actions which drive the Axiom towards a Logical
Form in PPT . For the course from Axiom to Goal cannot follow just any path
available; it must be a path licensed by the natural language under considera-
tion within the context of some language-external mechanisms. In the extended
examples we have seen that three kinds of actions contribute to a sequence of
parse states. Firstly, lexical actions, projected by the words of the language;
secondly, computational rules, determined by general principles of the language;
and thirdly language-external actions, like substitution, the resolution of scope
ambiguity, or inferential activity on the basis of the F o values of the (partially
created) tree description.14 That is, a transition hT Di; T Di+1i must be licensed
either by a word of the natural language under consideration, by general prin-
ciples of that language, or by general external principles. We will add to PPT
a set ACT of actions determined by some natural language. Actions are partial
multi-valued functions which map the tree descriptions in their domain to en-
riched tree descriptions.
{ Computational rules are actions which bring out information contained in
the current tree description and make hypotheses about structure.
{ Lexical actions map one tree description to a next one, adding information
in the process. These actions are projected by the words of a natural language.
Lexical actions are dened as conditional actions of the IF THEN ELSE variety.
A lexical transition tests IF some nite set of formulae (the condition) holds at
the node description where the pointer is located. This may include the presence
(or absence) of labels at that node, modal statements about decorations located
at descriptions related to the pointed one, and may also involve requirements. If
the condition holds there, THEN a (sequence of) actions is undertaken resulting
in a new tree description, ELSE ÀBORT'.
14 Of this last, we shall have nothing to say, since this concerns general context-dependent
inferences. However, we shall take up in subsequent chapters the processes of substitution
and resolution of scope ambiguity.
{ Pragmatic actions add information to a tree description which is neither

contained in the natural language string nor follows from language-specic prin-
ciples. These are actions of two sorts: (i) substitution processes which ènrich'
aspects of the tree description as it is projected, primarily by processes of substi-
tution replacing some lexically projected meta-variable with a completed term
taken either from other annotated nodes in the tree under description or from
externally provided information; (ii) steps of inference dened over formulae an-
notating nodes of the tree in conjunction with other externally provided informa-
tion. The implementation of both types of process is subject to system-external
constraints on their application. But the actions themselves are unrestricted
{ substitution, for example, allowing arbitrary terms of appropriate type to
replace a given projected meta-variable.
Given a set L of lexical actions, a set C of computational rules, and a set P
of pragmatic actions jointly determining a language L(L; C; P ), we will call an
LCP conguration a pair
(T D; w)
of a tree description T D and a string w of lexical actions; that is, w 2 L. The
set of all LCP congurations will be denoted by CONF (L; C; P ). The basic
rewrite relation of our parsing model is the binary relation on LCP congura-
tions,
(T D; w) )LCP (T D0; w0 );
where )LCP is dened to be )L [ )C [ )P , and the components of this
union are dened by
(T D; w) )L (T D0; w0 )
if w = w0 for 2 L and (T D) = T D0.
That is, )L transitions consume the rst word of the string and execute the
action it projects.
A transition dened by a computational rule or a pragmatic action has the
form
(T D; w) )x (T D0; w0 )
if x = C (or is the pragmatic case, x = P ), w = w0 , and there is some rule
2 C ( 2 P ) such that (T D) = T D0 .
Here no word is consumed; the action undertaken is determined by the set
C of rules or the set P of pragmatic actions.
The balance between the sets L, C, P may shift both between languages and
within a language, giving rise to diering reliance on contextual provision of
node annotations, and diering word-order eects. With respect to the relative
reliance on lexical or pragmatic actions, languages may dier as to whether
all argument nodes for a given predicate need lexical input, as in English, or
whether such arguments may be projected by a verb as meta-variables, hence
identied by pragmatic action, as in so-called pro-drop structures. Exemplifying
dierent weighting of the sets L and C , for English it is the Introduction and
Prediction rules that introduce subject and predicate nodes, whereas, in subject-
pro-drop languages, for example Arabic, it is the actions dened by the verb
that must induce the subject and predicate nodes (see chapter 4). In German,
we assume that Introduction and Prediction introduce not only the subject
and predicate nodes, but also the nodes requiring object and verb (see chapter
5). In Japanese, by way of yet further contrast, all nodes are introduced by
lexical action (though there may be locally dened sub-trees projected e.g. from
a verb for a predicate plus its attendant arguments). These alternatives for
constructing a common set of structures have the eect of imposing dierent
limits on word-order variation in the various languages.
The actions in C , P and L are all composed of elements of the set ACT of
actions. This set has the following kinds of actions:
1. creation of a new tree node connected by a tree relation to a given node
2. moving from a given node to another node along a tree relation
3. adding atomic annotations or (arbitrary) requirements to a given node
4. merging of a node with a dominated one.
Notice that the four kinds of actions are incremental on tree descriptions: if
is such an action, then T D (T D) for every T D in the domain of .15
These basic actions can now be combined in ACT by executing one after the
other (sequential composition `;'), by nitely iterating them (`*'), by choosing
indeterministically between a number of them (choice `+') and, nally, by con-
ditionalization of the form
IF DU formula holds at a given node,
THEN do action 1 ,
ELSE do action 2 .
Now we can restrict the general notion of development of tree structure as
formalized in the relation between partial tree structures P T R to those de-
velopments that are licensed by the lexical actions L and the computational
rules C . For instance, CP considers only developments without lexical con-
tributions. This is the relation along which the implication !CP is evaluated:
we consider invariances over developments that are possible without referring
to the next word. A connective of this sort that will be used in section 3.2.3
is the negation :C (an abbreviation of !C ?). This proposition excludes
from all computational developments of the current tree node (but lexical or
pragmatic actions may still add ).
It is now straightforward to dene a concept of well-formedness over the
sequence of transitions which a string of words must dene if it is to successfully
lead to a logical form as an interpretation assigned to that string. The sequence
15 Pointer movements may change the location of the `}' constant, but this symbol was
explicitly excluded in the denition of for tree descriptions (denition 9).
of words must lead from the Axiom to the nal state using each action dened
by the words in the order given, computational rules or pragmatic actions, so
that a well-formed logical form results, with no requirements on any tree node
left outstanding and no words left unprocessed.
(Axiom; w) = (T D0; w) )LCP (T D1; w0 ) )LCP : : : )LCP (T Dn; e) 2 Goal
where the constant e, as usual, denotes the empty string. As is common, we let
)LCP be the re exive and transitive closure of )LCP .
The set of grammatical strings, given C , L and P , can then be dened as the
strings that, starting from the Axiom, will, at the last word, reach a decorated
partial tree that is either already an element of Goal or that can be moved into
Goal by applying some computational and/or pragmatic rules:
L(L; C; P ) = fw 2 L j 9T D 2 Goal : (Axiom; w) )LCP (T D; e)g.
3.2.1 Computational Rules
By denition, whenever we have (T D; w) )C (T D0; w), then we have 8w 2
L : (T D; w) )C (T D ; w), because the string is not involved in computational
0
rules. So the notion T D )C T D0 makes sense. For visual convenience, the
computational rules 2 C will be displayed as transition rules of the form
1
2
where 1 is a (schematic) tree description to which the rule is applied, and 2
is the (schematic) tree description resulting from that application: whenever
T D )fg T D0, then 1 matches tree description T D, and 2 matches T D0.16
The transitions can be grouped into a number of discrete rewrite rules. Fig-
ure 3.1 gives an overview of the rules we use in this book. The basic rules come
in pairs, one element of the pair being in a sense the converse of the other one:
one element of the pair will use the requirements as a basis for computation,
whereas the other element will generally use annotations to compute other facts.
The rst pair which we consider will make this clear.
3.2.1.1 Introduction and Elimination
The Introduction rule licenses the expansion of a requirement for an annotation
of a given type into (a subdivision) of requirements for objects of other types: it
considers a requirement for a given type and generates sub-tasks to create two
other types the fullment of which would allow us to deduce the original type
by Modus Ponens. The fact that there is a pointer involved turns this rule into
a schema whereby every choice of pointer position determines a distinct rule.
16 In any application of these rules we allow variation as to whether their eect is achieved
through lexical or computational actions.
Introduction
f: : : f: : :?T y(Y ) : : : ; }g; : : : g
f: : : f: : :?T y(Y ); ?h#0 iZ0 ; ?h#1 iZ1 ; : : : ; }g; : : : g
where either Z0 = (T y(X ) ^ }) and Z1 = T y(X ! Y );
or Z0 = T y(X ) and Z1 = (T y(X ! Y ) ^ }):
Elimination f : : : fh#0 i(F o(); T y(X )); h#1 i(F o( ); T y(X ! Y )) : : : ; }gg
f: : : fF o(APL(; )); T y(Y ); h#0 i(F o(); T y(X )),
h#1 i(F o( ); T y(X ! Y )) : : : ; }gg
Prediction y
f: : : fX; : : :?h#iZ; }gg
f: : : fX : : :?h#iZ 0 g [ fY g; fh# 1 iX; : : :?Z g [ fY 0 g:::g
where X 2 ADR and Z = ( ^ }); Y = ;, Z 0 = ; Y 0 = f}g;
or Z = Z 0 = ; Y = f}g, Y 0 = ;
Completion
f: : : fX; : : : g; fX 0 ; : : : ; }g : : : g
f: : : fX; : : : h#i; }g; fX 0 : : : gg
where X 2 ADR; X = h#iX 0 or X 0 = h# 1 iX
and if ? 2 fX 0 : : : g then = ;:
Thinning f: : : f: : : : : :?; }g : : : g
f: : : f: : : : : : ; }g : : : g
*Adjunction ffT n(a); : : :?T y(t); }gg

ffT n(a); : : : ; ?T y(t)g fh" iT n(a); : : : ; ?9xT n(x); ?T y(e); }gg
Adjunction f: : : fX; : : : ; ?T y(x); }g : : : g
f:::fX; : : : ; ?T y(x)g : : : fh" iX; : : : ; ?9xT n(x); ?T y(x); }g:::g
X 2 ADR
Basic Merge
f: : : ND; ND0 : : : g
f: : : ND [ ND0 : : : g
} 2 ND0 ; (T D0 ; ND0 ) `PR V ADR(ND)
Gap Merge
( for type e) f: : : ND; ND0 : : : g
f: : : ND [ ND0 : : : g
} 2 ND00 ; T y(e0 ) 2 ND;
V ?T y(e) 2 ND0 ;
(T D ; ND ) `PR ADR(ND)
y For the meaning of ADR, see section 7.
Figure 3.1: The computational rules
The Elimination rule, conversely, considers the satised sub-tasks with their
types and deduces the required type.17 The Elimination rule incorporates a form
of Modus Ponens dened under the Type label (T y), with function application
dened under the Formula predicate label (F o).
Introduction
f : : : f: : :?T y(Y ) : : : ; }g; : : : g
f: : : f: : :?T y(Y ); ?h#0 iZ0 ; ?h#1 iZ1 ; : : : ; }g; : : : g
where either Z0 = (T y(X ) ^ }) and Z1 = T y(X ! Y ),
or Z0 = T y(X ) and Z1 = (T y(X ! Y ) ^ }).
This rule sets the stage for an application of Modus Ponens under the T y pred-
icate which would give the required type.
An instance of this rule, that has been used in steps 2 and 3 of the basic
example, is the following:
f: : : f: : :?T y(t) : : : ; }g; : : : g
f: : : f: : :?T y(t); ?h#0 i(T y(e) ^ }); ?h#1 iT y(e ! t) : : : ; }g; : : : g
An application of this rule analyses a requirement for a type t object into two
new requirements, one for a type e and one for a type e ! t object. Moreover,
the pointer follows the type e requirement.
Elimination
f : : : fh#0 i(F o(); T y(X )); h#1 i(F o( ); T y(X ! Y )) : : : ; }gg
f: : : fF o(APL(; )); T y(Y ); h#0 i(F o(); T y(X )),
h#1 i(F o( ); T y(X ! Y )) : : : ; }gg
This rule applies Modus Ponens under the Type label and, in tandem, executes
function application under the Formula predicate.18
The rules Introduction and Elimination work together with the following
pair: Introduction sets the stage for applications of the Prediction rule, and the
Completion rule precedes application of Elimination.
3.2.1.2 Prediction and Completion

Now we have two rules which license the creation of new tree node descriptions
and the transfer of information from one node description to another. Given,
17 The rule of Introduction as it stands is much too general, allowing the introduction of
pairs of nodes annotated with arbitrary type as long as they meet the specication of being
of the form X , X ! Y , respectively. We leave it in this general form here, but a more precise
characterization can be found in the axiom section of chapter 9.
18 The asymmetry between the Introduction and the Elimination rule derives from the fact
that we do not prune the trees. Given a normalized formula , the Introduction rule chooses an
unreduced term ( ; ), which reduces to . Two daughters are created at locations for the sub-
terms and . Conversely, given two daughters with annotations and , the Elimination
rule does not prune them when it annotates the mother with the reduct. Symmetry would
arise with an Elimination rule which prunes the daughter nodes.
for instance, ?h#1iT y(e ! t) in a node description ND
fT n(a) : : :?h#1 iT y(e ! t)g;
the Prediction rule allows the introduction of a new description with address
h"1 iT n(a) and the (simpler) requirement ?T y(e ! t). So this rule allows the
unfolding of modal requirements into tree structure and simpler requirements.
Prediction
f : : : fX; : : :?h#iZ; }gg
f: : : fX : : :?h#iZ 0 g [ fY g; fh# 1 iX; : : :?Z g [ fY 0 g : : : g
where X 2 ADR and Z = ( ^ }); Y = ;, Z 0 = ; Y 0 = f}g,
or Z = Z 0 = ; Y = f}g, Y 0 = ;
The complications arise because of the choice of pointer position; in applying
the rule, the pointer only moves if a new pointer position is created.
We can apply this rule to the output of the application of the Introduction rule
we have given as an example:
f : : : fX; : : :?h#0 i(T y(e) ^ }); h#1 iT y(e ! t); }gg
f: : : fX : : :?h#0 iT y(e); h#1iT y(e ! t)g; fhh#0 1iX; : : :?T y(e); }gg
Conversely, given the creation of a formula statement F o() with associated
type statement T y(e ! t) at the function daughter of a node description, the
Completion rule licenses the annotation of that (mother) node with h#1 i(F o( );
T y(e ! t)).
Completion
f : : : fX; : : : g; fX 0; : : : ; }g : : : g
f: : : fX; : : : h#i; }g; fX 0 : : : gg
where X 2 ADR; X = h#iX 0 or X 0 = h# 1 iX
and if ? 2 fX 0 : : : g, then = ;.
It will be a matter of choice when to apply the Completion rule (followed by
Elimination ) in the course of a parse. In general, we will apply the rules as late
as possible. We will restrict applications of the Completion rule to situations
where all requirements in the source description have been satised. That is,
we demand that the node description containing the pointer } above the line in
the rule formulation should not contain question marks: information may only
be propagated from one node description to another if no requirements in the
source are outstanding. After all, it may turn out that some requirements are
unfulllable and thus come from a description which is not a viable description
of a tree node.
3.2.1.3 Thinning
In the course of a parse, annotations and requirements accumulate within node
descriptions. A requirement within a node description ND is declared fullled
if ND contains both ? and . At this point the Thinning rule allows us to get
rid of the question mark.
Thinning
f: : : f: : : : : :?; }g : : : g
f: : : f: : : : : : ; }g : : : g
The application of some rules may require that a description no longer has
any requirements. Application of the Thinning rule must thus precede the
application of such a rule. Notice that this rule requires the presence of the exact
formula that is required. That is, we cannot apply the rule to, for instance,
f: : : ; ; : : : ; ?( ^ )g
directly. We need a derivation step which concludes ^ from these premisses
and adds this proposition to the description.
3.2.1.4 Adjunction and Merge
So far we have set out a system of rules which enables tree descriptions to be
successively enriched, starting from the initial node description. Central to this
system is the concept of partial specication: at any stage but the last, the
description of both tree nodes and tree relations may be partial. We have now
to add the rules which enable us to add node descriptions with an underspecied
tree relation to the tree description so far constructed. Recall that this, with
its use of the Kleene star operator, is the means of describing the dynamics
involved in parsing left-dislocation structures.
In order to allow underspecied tree descriptions, we need rules for their
introduction and elimination. The introduction will be taken care of by Ad-
junction rules. These introduce node descriptions which still have to nd a
proper location in the tree under development. The Elimination rule merges,
i.e. unies, two nodes in a given tree, in general a node with another whose
address is fully specied.
The general form of an adjunction rule is:
f: : : fT n(a); : : : ; }g : : : g
ffT n(a); : : : g : : : fh" iT n(a); : : : ; }gg
Addition of an unxed node.
We will use two forms of this rule. The rst form, which we call *Adjunction,
is used in the treatment of long-distance dependencies.
*Adjunction
ffT n(a); : : :?T y(t); }gg
ffT n(a); : : : ; ?T y(t)g fh" iT n(a); : : : ; ?9xT n(x); ?T y(e); }gg
In this version a node requiring a type e is introduced, dominated by a node
with a requirement for an object of type t. This rule sets the stage for the
processing of a (sentence-initial) noun phrase which has an as yet unspecied
argument role in the structure to be projected. With this transition so dened,
left-dislocated expressions can be characterized as projecting their content in
the normal way, but now with respect to a node with an underspecied address.
Adjunction
f: : : fX; : : : ; ?T y(x); }g : : : g
f: : : fX; : : : ; ?T y(x)g fh" iX; : : : ; ?9xT n(x); ?T y(x); }g : : : g
X 2 ADR
The rule introduces a duplicate of the node currently under the pointer and
moves the pointer to this copy. The address of this new node is underspecied19
in that it is only determined that the copy lies somewhere `below' its original.
This rule sets the scene for a type x nested within a type x. That is, tree growth
by Introduction and Prediction gives now the auxiliary trees from tree-adjoining
grammar (TAG) formalisms (Joshi et al 1975), which share a label between the
root and some terminal node. If we expand the tree below the pointed node,
then we can eventually reunite the two versions of the same node by a merge,
but we might as well not have split the node. On the other hand, if we build
on top of the pointed node, then we have to construct some top node of the
same type to merge with the original one. For this we need some tree modier
of type x ! x as the function daughter of that top node. So this method of
construction allows for adjunct-like operations. In this type of construction, it is
essential that the node has been split up, because, in eect, we have interpolated
tree structure, and real interpolation is not an incremental process, and so is
debarred.
A = fT n(a); : : : ; ?T y(x); }g
B = fh"iT n(a); : : : ; ?9xT n(x); ?T y(x); }g
...
C = fh#0ih" iT n(a); : : : T y(x); }g
Here node A has generated the copy B, and C is a node which has B as an
argument daughter. Eventually C will be positioned between nodes A and
B , or could merge with A, for it is a LOFT theorem that (T n(a) ^ h#0 i> !
h#0 ih" iT n(a)).
A node description with an underspecied address has to nd a denite
location at some point in the course of the parse. For this purpose we introduce
19 In chapter 4, this will include both daughter and link relations.
a general rule to combine information, that we call Merge. This allows node
descriptions with underspecied addresses to nd their xed location in a tree
description by merging with such nodes.
Basic Merge
f: : : ND; ND0 : : : g
f: : : ND [ ND0 :V: : g
} 2 ND ; T D0; ND0 `P R ADR(ND)
0
In this rule two node descriptions are combined in case the address { that is all
propositions in the ADR language of one node (ND0 above, see denition 7)
{ can be `derived' or `computed' from the pointed node using the principles in
P R. These principles allow the use of standard propositional and modal logic,
the tree-specic LOFT principles and various Vprinciples regulating the Label
and Formula predicates.20 If (T D; ND0) `P R ADR(ND), then the address
of ND is an underspecication of the address of ND0: for instance, ND has
address h"iT n(a) while ND0 has address h"1 ih"0iT n(a), or ND has address
h"1 ih" ih"0 iT n(a) and ND0 address h"1 ih"1 ih"1 ih"0 iT n(a). As an application
of the general rule we will merge an underspecied node annotated by a formula
John with a node with a fully specied location but annotated by a meta-variable
of type e (this situation will occur in chapter 6).
f: : : fh"iT n(0); F o(John); T y(e); ?9xT n(x)g
fh"0iT n(0); F o(U); T y(e); }g; : : : g
f: : : fh" iT n(0); h"0iT n(0); F o(U); F o(John); T y(e); ?9xT n(x); }g; : : : g
f: : : fh"iT n(0); h"0 iT n(0); F o(John); T y(e); }g; : : : g
To take these steps in sequence, the rst transition step is an application of the
Merge rule. By LOFT we can show,
fh"0iT n(0) : : : g `P R h" iT n(0);
so the address of one node description can be derived from the other description.
The second transition then takes account of the fact that
fh"0iT n(0)g `P R 9xT n(x) & fF o(John)g `P R F o(U);
so two applications of Thinning will give the bottom line of the derivation.
A more specialized form of this rule can be dened that requires for a merge
of two nodes that they supply complementary type information. That is, one
node should supply the type required by the other. This is a typical situation in
the characterization of left-dislocation sequences.21 We call this rule Gap Merge
because it allows a node (giving type information) to ll a gap (a node requiring
20 See section 9.2.4.
21 We retain here the familiar terminology of gap, but there are no such nodes in this system.
A gap is nothing more than a xed node with a requirement for a certain type value which in
successful derivations merges with a node with a hitherto unxed tree-node address.
a certain type):
Gap Merge (for type e)
f: : : ND; ND0 : : : g
f: : : ND [ ND0 : : : g
} 2 ND ; T y(e) 2 ND; ?T y(e) 2 ND0 ; (T D0; ND0 ) `P R V ADR(ND)
0
Gap Merge is applied when the pointer is at a node requiring an object of type
e and the underspecied node can supply such an object. Thus this form of the
rule allows only merging of information with a gap, merging in order to full
requirements. As an example, we show how an underspecied node annotated
by a formula John may merge with a node with a fully specied location but
requiring a formula of type e.
f: : : fh" iT n(0); F o(John); T y(e); ?9xT n(x)g fh"0 iT n(0); ?T y(e)}g; : : : g
f: : : fh" iT n(0); h"0iT n(0); F o(John); T y(e); ?T y(e); ?9xT n(x); }g; : : : g
f: : : fh"iT n(0); h"0iT n(0); F o(John); T y(e); }g; : : : g
The rst transition step is an application of the Gap Merge rule as before. The
second transition is then an application of the Thinning rule with respect to the
propositions ?9xT n(x) and ?T y(e).
So Merge is the rule which has the eect of xing what has hitherto in
the parsing process been described as unxed within a tree. It is a rule that
unies two descriptions for a given node. Primary amongst the cases to which it
applies is the case when a xed node is introduced with a requirement which the
annotations of the unxed node would satisfy and there is no other candidate
for its completion. In such a case the two node descriptions which the node
may have,collapse into a description of a single node. The result will be that
the node has then a xed tree position, while nevertheless remaining consistent
with the earlier description. This rule applies in the construal of wh questions,
in topicalization structures, and in relative clauses:
(3.6) Who do you think will succeed?
(3.7) My mother, I think will succeed.
(3.8) The man that I think will succeed has already applied.
However, the rule may apply whenever there is in the current tree description
a description of both a xed node and an unxed node which could potentially
be identied; but, in most cases, this unication process will immediately lead
to inconsistency.22
The possibility of unifying a node described as unxed with one that is xed
will also allow us to unify meta-variables with their values, for a meta-variable is
22 It should be noted that a node may only be annotated with more than one Formula value
if the values stand in an update relation Fo. Recall that the Formula values re ect the
content of the natural language expressions, so gender and number information do not project
Formula values, but values for some discrete Labels.
consistent with any xed term of the same type, so nothing prevents an identical
interpretation of (3.9) and (3.10):
(3.9) My poor mother, I can just see she's going to suer.
(3.10) My poor mother, I can just see is going to suer.
Indeed, the interpretation of the pronoun in (3.9) must be construed as identical
to the annotation of the unxed node in order to lead to a well-formed outcome.
As we shall see in chapters 5 and 6, this will have considerable ramications, for
it will provide the basis for an analysis of resumptive pronouns, the distinction
between topic and focus structures, and a solution to the puzzles posed by
crossover phenomena.
3.2.1.5 Pointer Movements
The dynamics of tree constructions require actions to be taken at specic nodes
in a partial tree. These nodes are highlighted in the tree by the presence of a
pointer. Structurally this pointer is without meaning (that is, the pointer has
no role to play in the logical form under construction), but dynamically the
pointer and its location in the tree form an essential feature of a partial tree
structure. The same partial tree with the pointer at two distinct locations may
behave completely dierently with respect to actions projected by the words
of a string, or actions induced by general computational principles. In other
words, for a lexical or computational action to be eective, it must come at
the right moment.
Pointer movement can be induced in a variety of ways:
First and foremost, pointer movement is induced by the computational
rules like Prediction and Completion (see section 3.1.1, step 3).
A second source of pointer movement is given by the lexical entries. A
word projects a lexical entry, a lexical action (section 3.2.2). This action is
executed depending on a condition which is evaluated with respect to the
current pointer position, and it leaves the pointer at some new location in
the extended structure.
A last source of pointer movement is part of a general strategy on what
to do when some construction is completed; that is, if all requirements
at some node have been fullled. For instance, the pointer may then
move to some c-commanded terminal node which still has requirements
outstanding, if there is such a node.
Pointer movements are guided through the tree along tree relations. That is,
we exclude àbsolute' addressing. In this way these movements can readily be
interpreted on the familiar partial tree structures. The set of movements is
constituted by the following actions. There are the basic movements, taking
one step from tree node to tree node along one of the tree relations. These basic
movements can now take part in complex ones through the regular operations
of sequential composition `;', indeterministic choice `+', and nite iteration ` '.
Movements can also be made conditional on decorations of the starting point.
3.2.1.6 Alternative Parse Courses
The progress from Axiom to Goal is non-deterministic: at every state of the
parse, the word currently under consideration can generally be assigned more
than one structural role in the tree description by principles in C [ P . For
instance, an NP heading a natural language string may end up as subject John
admires Mary; it may be a fronted object, Mary, John admires; or it may be
a topic constituent, (As for) Mary, John admires her. The set C must contain
actions to set up these three structures. The transitions determined by the
computational rules, i.e.
(T D; w) )C (T D0; w);
are non-deterministic ()C is a multi-valued function). But, as is evident, for
the parsing process to get o the ground at all, for any given tree description T D
the set C should contain only a nite number of (alternative) actions which have
T D in their domain: only a nite number of developments can be computed.
That is, we stipulate:
ACT (T D) = f 2 C j T D 2 dom()g
is a nite set. Moreover, we should not be able to apply rules indenitely: after
a certain number of applications, we must have reached a tree description that
is no0 longer in the domain of any rule in C . Thus, for a given T D, not only is
fT D j 9w : (T D; w) )C (T D0 ; w)g nite, but also fT D0 j 9w : (T D; w) )C
(T D0; w)g.
3.2.2 Lexical Transitions

The transitions we have discussed in the previous sections all proceed without
reference to the natural language string. They are all formal, or computational,
rules of the system. In this section we introduce the actions projected by the
words of a natural language. We will call these projections Lexical Actions, and
the transitions induced by them, Lexical Transitions.
Lexical actions increase the informational content of a tree description: they
map a tree to another tree in a way that no computation can achieve. For
example, we might dene a verb as inducing the node which it then annotates
with a predicate, adding a requirement for its sister node on the mother. In
so doing, we must allow a word to build nodes and to decorate nodes with
requirements as well as annotations. This involves moving around a tree, so
we shall need an action which moves the pointer. In this way, the actions
projected by a single verb carry out a whole range of dierent updates. Other
words may project only an annotation as a decoration (proper names), or only
a requirement (case particles), or an annotation/requirement on the mother

node (tense specications), or even only movement of the pointer (some forms
of expletive). Examples of these will follow shortly.
Each lexical action involves a triple of a condition and two actions:
IF all formulae in hold at the pointed node description,
THEN do action 1,
ELSE do action 2.
where ÈLSE' is commonly an instruction to abort. This is because the action
of interpreting a lexical item in a string is not something which is optional. The
action is obligatory, and its eects unavoidable.
The lexical actions we will introduce are composed of a set of basic or atomic
actions. The basic actions consist of creation of new nodes relative to old ones,
decoration of nodes, and moving to nodes (which includes staying put). This
set of basic actions is dened as:
Denition 10 (Actions) For T rDes the set of all tree descriptions, # 2 f#0
; #1 ; "0; "1 ; #; "; #; " g and a DU proposition, the Basic Actions are the partial
functions:
1. 1 = fhT D; T Di j T D 2 T rDesg, ABORT = ;. These represent the
halting action and the abort action respectively.
2. make(#) : T D 7! T D. This action creates a node ND0 such that ND #
ND0 , where } 2 ND 2 T D.
3. go(#) : T D 7! T D0. Here T D = T D0 except possibly for the location of
the pointer. If } 2 ND 2 T D, and } 2 ND0 2 T D0, then ND # ND0,
4. put() : 0T D 7! T D0. Here, if } 2 ND 2 T D, then T D = T D0 except
that ND = ND [ fg,
5. merg (ND1; ND2) : T D 7! T D0. Here, if } 2 ND1 and (T D1; ND1) `P R
ADR(ND2 ), then T D0 = T D except that T D0 = T D fND2g and
V
ND10 = ND1 [ ND2 .23

We can put actions together to form complex actions, by executing one ac-
tion after another (sequential composition `;') and doing that any nite number
of times (nite iteration `'), or by indeterministically choosing between them
(indeterministic choice `+').
6. if ; 0 are actions, then so are ; 0 ; + 0 ; .
Finally, we can put actions together in a conditional IF , THEN 1, ELSE 2
statement. The conditional action that we dene is a little more complex than
this. It allows the condition to contain free variables which can be bound to
values at the pointed node such that these values can be used in the action part
23 In chapter 9 a slightly dierent denition will be given (denition 28).
of the conditional action: IF (x), THEN put((x)), ELSE ABORT is then
interpreted as follows: if at the pointed node (t) holds (i.e. [t=x]), then this
value t is passed on to the action, and the rst action has the form put([t=x).
7. If is a set of DU propositions all free variables of which occur in x,
and ; 0 are actions, then h(x); ; 0 i is a conditional action with the
denition:24
h(x); ; 0 i =
fhT D; T D0i 2 [t=x] j t 2 (D [ MV ) ; T D `P R [t=x]g
[
fhT D; T D0i 2 0 j :9t 2 (D [ MV ) ; T D `P R [t=x]g:
(Recall that the notion ``P R' was discussed in section 3.2.1.4. Here, the notion
T D `P R means (T D; ND) `P R for the unique ND 2 T D such that } 2
ND.) That is, if `holds at' (T D; ND), where } 2 ND, for some substitution
of t for variable y 2 x, then action is executed, where, in the body of this
action, the variable y is replaced by t. If does not `hold at' (T D; ND) for any
substitution for y, then action 0 is executed.
This construction enables composite actions to be dened which manipulate
values that are determined by the pointed node. For instance, we can dene
actions which transport features from one node to another.
hfLai (x)g; go(#0 ); put(Lai (x)); ABORTi
maps the value of the Lai feature at the current node, if there is such a
value, to the Lai feature at the argument daughter, otherwise it aborts.
We might also dene an action that enables annotations to be added to the
closest clausal node (e.g. for tense suÆxes):
gofirst"(?T y(t)) = hf?T y(t)g; 1; go(")i ; hf?T y(t)g; 1; ABORTi.
The action gofirst"(X ) moves the pointer upwards to the rst higher
node with annotation X and requirement T y(t), and then stops. The
action gofirst#(X ) is the corresponding action going downwards.
We can also dene a macro that involves starting at some node, going somewhere
to execute some actions, and then returning to the point of departure.
hfT n(x)g; gofirst"(T y(e)); put(); gofirst#(T n(x)); ABORTi.
The return to the starting point is eected by binding the variable x to
the current value of the tree node predicate. This value can then be used
to return by one of the gofirst actions. Of course, this presupposes that
the current node does indeed have such a value.
24 We use h(x); ; 0 i in the denition as the three-fold sequence of hTEST, ACTION 1,
ACTION 2i, but in general we depict such conditional actions in the form:
IF (x)
THEN
ELSE 0
A further action that we will describe deals with the introduction of a fresh
logical name, either a constant or some variable, to be used in some formula .
The notion of `freshness' is here dened with respect to the tree structure under
development.
freshput(a; )=
IF f"# F o(a1 )g
THEN IF f" # F o(a2 )g
THEN : : :
: : : IF f"# F o(an )g
THEN ABORT
ELSE put(F o([an =a]))
:::
ELSE put(F o([a2 =a]))
The action freshput(a; ) generates a xed nite number of instances
of the term a and uses the rst one that is fresh in . So, to actually
contribute a fresh variable to some tree, this number has to exceed the
number of a occurrences in that tree.
This macro is used in the actions projected by proper names and pronouns.
We will now exemplify some lexical actions in detail. The name John will
have the lexical entry25
John IF f?T y(e)g
THEN put(T y(e); [#]?); freshput(John; F o(John))
ELSE ABORT
The action projected by `John' tests whether the pointer is located at a node
requiring an object of type e. If this is the case, then it annotates the node
with F o(Johni ), the rst John-copy not occurring in the tree, with the type
declaration T y(e), and with [#]?, which declares this node to be a terminal one.
If the test fails, then (this strain of) the parsing process aborts. So distinct
occurrence of the name John in one sentence will give rise to distinct logical
constants annotating tree nodes. Dierent tree nodes may end up with instances
of the same logical name, but only through a process of copying.
The English transitive verb like will have the entry:
like IF f?T y(e ! t)g
THEN make(h#1i); go(h#1i);
put(F o(Like); T y (e ! (e ! t)); [#]?);
go(h"1 i); put(?h#0 i(T y (e)))
ELSE ABORT
25 As is standard in the literature, we side-step problems in dening linguistic properties
of proper names and assume that they are logical constants. In a fuller characterization, a
linguistic name should be analysed as projecting a meta-variable in the manner of pronouns,
with restrictions on the substituend that it be a term denoting an individual suitably named.
See section 7.3.1.1 for a preliminary formulation of the analogous presuppositional properties
of denite noun phrases.
Notice how like, at the pointed node decorated with a requirement for T y(e ! t),
releases a sequence of actions: creation of a rst daughter node, annotation of
that node with a type and a formula and the imposition on the mother node,
the predicate node, of a requirement for a second daughter of type e. Actions
decorating a node with annotations are the analogue of lexical substitution
processes in other frameworks (though it should be remembered that it is not
lexical items which are inserted in the tree in this framework, but their logi-
cal correlate). Actions decorating a node with requirements are analogous to
subcategorization statements. However, in this framework, recall, all nodes are
introduced with requirements, so unlike standard subcategorization statements
which are imposed on sisters to terminal nodes as a requirement on lexical in-
sertion in a tree, in this framework the use of requirements is more widespread.
The characterization of like illustrates how actions induced by a lexical item
may include decorations of a node other than the one at which the pointer sits.
The tense specication on a verb is another example involving annotation of a
mother node; and arguably agreement can also be expressed as an addition to
the mother node of a modal requirement. We presume this takes place from the
node requiring T y(e ! t),26 following phonological evidence27that the suÆx is
identied as entirely separate from its host (see Kaye 1996):
suÆx -s (tense marker) IF f?T y(e ! t)g
THEN put(Agr(sing)); go(h"1i);
put(T en(Si Se ); ?h#0 iAgr(sing))
ELSE ABORT
Lexical specications may also indicate the relative position in the tree of the
node currently being decorated. Thus case information, which indicates what
kind of node in the tree the formula currently being projected should be taken
as decorating, can be expressed as a modal requirement on the node under de-
scription. In Japanese, for example, in which case particles are suÆxed to an
NP constituent, the particle -o projects a requirement on the relative position
of the node currently being annotated that its mother node must be a formula
of type e ! t:28
o IF fF o(); T y(e)g
THEN put(?h"0 iT y(e ! t))); gofirst"(T n(0))
ELSE ABORT
The projection of wh expressions illustrates the need for nested specications
for a lexical item. Such nested characterizations arise in all cases where a lexical
26 We have made several assumptions about tense. First, we have assumed here that tense
projects a relation between time points. The time point relative to which the proposition
expressed is evaluated is itself a label to the propositional form in the resulting formula.
Secondly, we have assumed that tense is a decoration on a node, rather than projecting a
node of its own (see section 2.1.1.1).
27 In this book we have nothing to say about the phonology/syntax interface. So we leave
on one side the clitic-like status of -s in likes and the morphological characterization of phono-
logically reduced forms such as clitic and resumptive pronouns.
28 The suÆxes also add the information that the pointer is returned to the root.
expression has distinct but overlapping conditions of use. For example, we take
wh-initial expressions in questions to project a categorial feature Q, wh-in-situ
questions to simply project the meta-variable:29
what IF f? T y ( e ) g
THEN IF fh" i?T y(t)g
THEN put(F o(W H ); T y(e); [#]?);
go(h" i); put(Cat(Q))
ELSE put(F o(W H ); T y (e); [#]?)
ELSE ABORT
Such item-internal disjunctions should not be confused with discrete lexical
entries such as might be proposed for bank1 and bank2. The item-internal dis-
junctions are for cases in which an individual word projects a number of discrete
actions, which share a core subset of properties. For example, a word in the
language may project a single Formula value with an accompanying sequence of
actions which are particular to individual structural contexts. In the example de-
ned here, wh expressions invariably project a distinguished WH meta-variable;
but in addition, when initial in a clausal sequence and annotating an unxed
node, an additional clause-typing feature Q is added to the node from which
the unxed node was introduced. Notice that this meta-variable WH is not ac-
companied by a substitution requirement ?9xF o(x). That is, the WH variables
may remain as variables in the nal structure, and wh questions are accordingly
taken to express an incomplete propositional formula, lacking a substituend for
(at least) one argument. This distinguishes them from all other lexically pro-
jected meta-variables which have to be substituted as part of the construction
process (see section 3.2.3).
The examples listed so far have primarily been English, but these should
not be taken as illustration of an invariant cross-language pattern. The details
of the actions may dier from language to language, with variation in both the
condition and the actions. We give here the projection of Arabic abil (= `like'),
showing how in subject-pro-drop languages the actions initiated by the verb
lead to the construction and annotation of the subject position and the projec-
tion of the Japanese tabeta (= èat'),30 where the verb induces both an unxed
node requiring type t and the construction of a sub-tree all of whose nodes are
annotated:
29 Cat is a sentence-typing predicate. In general, for convenience, we shall simply use the
short-form Q (see chapter 5).
30 As elsewhere, we are here ignoring the tense suÆx -ta, taking it, for simplicity, to be part
of the stem.
abil [Arabic] IF f?T y(t)g
THEN make(h#0i); go(h#0i); put(F o(U); T y(e); ?9x(T n(x)));
go(h"0 i); make(h#1 i); go(h#1 i); put(?T y (e ! t));
make(h#1 i); go(h#1 i); put(F o(Abil); T y (e ! (e ! t)));
go(h"1 i); put(?h#0 i)
ELSE ABORT
tabeta [Japanese] IF f?T y(t)g
THEN make(h# i); go(h# i); put(?T y(t); ?9xT n(x));
make(h#0 i); put(F o(U ); T y (e); ?9xF o(x)); go(h"i);
make(h#0 i); go(h#0 i); put(F o(V ); T y (e); ?9xF o(x));
go(h"i); make(h#1 i); go(h#1 i);
put(F o(T abeta); T y (e ! (e ! t))); gofirst" (?T y (t))
ELSE ABORT
The details of each such lexical entry in part re ects language-particular word
order, since the specics of pointer movement determine what subsequent ac-
tions may be available. The Arabic verb abil moves the pointer back to the
predicate node awaiting a step of Prediction, hence the presence of a following
object expression, whereas the Japanese verb tabeta moves the pointer back to
the root.
3.2.3 Pragmatic Actions and Lexical Constraints
Of the types of lexical action considered so far, one important sub-type has
been omitted. There is a set of lexical actions which update the tree description
with node decorations which do not in themselves full all requirements on that
node, but serve rather as constraints on a third type of action involved in the
construction process { pragmatic actions.
There are two principal sorts of constraints on pragmatic actions:
(i) constraints on the pragmatic process of Substitution
(ii) constraints on the form of inferences derivable from the logical form once
constructed.
We have nothing to say here about the latter, apart from noting the existence
of particles such as even and only, and the connectives but, however and af-
ter all, whose distinctive property is the contribution they make not to the
truth-functional content of logical form but rather to the non-logical inferences
derivable from that form.31
Under heading (i) fall all expressions whose lexical specications systemat-
ically underdetermine their interpretation in context { i.e. anaphoric expres-
31 Recall from chapter 1 that the construction of some distinguished logical form is assumed
to be combined with some set of premisses constituting `the context', the combination yielding
some suitable array of contextually derived implications (Sperber and Wilson 1995).
sions, quantifying expressions, temporal auxiliaries, etc.32 It is characteristic

of anaphoric expressions that they do not themselves project a complete anno-
tated formula on a node, but rely on some substitution process to provide that
completion. All the system-internal characterization provides is a place-holding
device and a locality constraint on the substitution process. The intrinsic con-
tent of the pronoun he, for example, is a place-holder:
he IF f?T y(e)g
THEN put(F o(U); ?h"0 iT y(t); ?9xF o(x))
ELSE ABORT
Note the projection of a requirement for a F o value, whose eect is to ensure
that no requirements remain in the nal state of the annotated node. No re-
striction is imposed, however, distinguishing selection of a variable, a quantied
term or a constant: construal of a pronoun as bound-variable, referential or
E-type is merely a consequence of dierent selected forms of substituend (cf.
chapter 7 for a discussion of how E-type anaphora might be characterizable in
this framework).
The pragmatic action of Substitution which the lexical characterization of
a pronoun presumes on is external to the computational/lexical system. It is,
however, constrained by this system (through locality) and by more general
considerations pertaining to the additional pragmatic goal of constructing some
appropriate context with which to draw inferences from the given string.33
The intuition behind the notion of locality we use is that of a domain of co-
arguments to a predicate. In a fully annotated binary tree representing a logical
form, the predicate annotates a terminal node that is a function daughter (1).
In a tree without adjunct structures, this terminal node is related to the top
node, annotated by a type t formula, by a sequence of function relations.
T y (t)
arg
arg pred
We can describe such a sequence with the transitive and re exive closure of the
function daughter relation: 1 (which we can refer to in the language by h#1 i).
The arguments of the predicates can be found annotating argument daughters
along this function spine. From the perspective of an argument node (which
32 In a more complete characterization, names would also fall in this class. Unlike Pinkal
(1995) and Poesio (1996), we assume that lexical expressions with quite disjoint interpretations
are characterized as discrete lexical items. However, for present purposes, we assume that
verbs express xed concepts. See Marten (1999) for an exploration within this framework
of the view that verbs systematically underspecify their type assignment, the extractability
of NPs from within PPs being a re ex of their optional argument status { with verbs said
to lexically project a partial representation of a predicate whose type specication, hence
content, is xed only by the structural context.
33 Constraining actions such as Substitution are where relevance and other pragmatic con-
siderations have a role to play.
itself may dominate structure), its co-arguments can be found by moving to
the mother (upward along 0), then at least one step upward or at least one
downward along the function relation (1), and then considering the formula at
the argument daughter. This route upwards is described by the sequence CA
of modalities
CA = h"0 ih"1 ih"1 ih#0 i;
the route downwards analogously by h"0ih#1 ih#1 ih#0i.
These (sequences of) modalities allow us to formulate locality constraints on
substitution for pronouns. A substitution action SUBST (John) 2 P for the
proper name John, for instance, can now be given the (provisory) form:
SUBST (John) IF fCA(F o(John)); F o(U); ?9xF o(x)g
THEN ABORT
ELSE put(F o(John))
This substitution action checks, when the pointer is at a node annotated by
a place-holder projected by a pronoun, whether F o(John) annotates a co-
argument node. If it does, then no further action is taken; if it does not, then
F o(John) is added to the annotations of the current node.
When we turn from fully annotated binary trees to partial ones, the notion
of locality becomes somewhat more complicated. A verb projects a predicate
which is surrounded by its arguments either as already constructed or as still
required. Now the notion of a co-argument of a node includes not only the
co-arguments in the current partial tree, but also the nodes in that tree that
can become co-arguments { that is, structurally underspecied nodes.
T y (t)
John arg
U pred
In this tree, the node annotated by John is not reachable by CA from the one
annotated by U, but it could merge with the node annotated by àrg' and thus
become a co-argument. Merge is one of the actions in C , thus there is a C
successor of the above tree where John and U are co-arguments.
T y (t) T y (t)
John arg C John
Upro pred Upro pred

Substitution of John for U should thus be excluded. We can express this by
using the negation :C discussed in section 3.2.
SUBST (John) IF f:C CA F o(John)); F o(U); ?9xF o(xg

THEN put(F o(John))
ELSE ABORT
This action tests that no node that can merge with a co-argument is annotated
by John. If this is the case, then F o(John) is added; otherwise no action is
taken. The use of :C is computationally safe here, because there are only a
nite number of computational developments to check.
This use of negation gives us a means of expressing a familiar concept of
locality. When we need to distinguish this concept from a weaker relation hold-
ing between two nodes within a single tree (see chapter 4), we shall refer to
two argument daughter nodes along a function spine as being most-local to each
other.
Pronouns are by no means the only form of anaphoric expression. There are
the denite and demonstrative determiners, and there are also tense particles,
which we might dene using temporally restricted variables, Si, Se:
didAUX IF ?T y(t)
THEN put(T en(Si < Se ); F o(Si : U))
ELSE ABORT
The meta-variable U here is a place-holder for some formula of type t.
As these various lexical statements have illustrated, the lexicon is the repos-
itory of the idiosyncratic, with individual lexical items decorating one or more
nodes with annotations/requirements, possibly also inducing new nodes, impos-
ing requirements on them, or imposing constraints on pragmatic actions (e.g.
Substitution ).
3.3 Summary
In this chapter we have laid out a set of rules for building up annotated trees
with a propositional formula labelling its root node. In chapter 4 we shall see
how trees can be put together to yield sets of linked tree structures.
In the mean time, we should not lose sight of the underlying motivation in
developing the model. We are modelling the process of building up an inter-
pretation of a natural language string, developing a structural representation of
content by building up a tree over which denotational content can be dened.
The model re ects the partiality of information at every stage, explicitly char-
acterizing the gap between input specications of individual lexical items and
the output content as assigned to the string. The representation will not be
established until the tree description is completed. Even the relations between
individual nodes may be relatively unxed at intermediate stages of interpreta-
tion.
The concept of interpretation that this model provides is that of an emergent
logical form induced by the progressive mapping of a natural language string
Summary 99
onto a tree structure. We don't model natural languages as inference systems,
but as devices for creating expressions over which inference can be dened.
Conceived of as the development of a grammar formalism, the current model
is also a departure from assumptions of other formalisms. The particular gram-
mar formalism set out here is not constructed according to the formal language
model. Rather, we are building in an incremental projection of structure re ect-
ing the left-to-right sequence of words as the mode of interpretation. This makes
it very dierent from categorial formalisms along the lines of Morrill (1994) for
example. We are not dening structure for sentences by a set of rules that
merely express structural properties of strings and their interpretation; we are
modelling the building of structure as a step-by-step process that re ects the
procedure of using the language.34 This is distinct even from those categorial
frameworks which incorporate some sensitivity to linearity into the formalism
for example Steedman 1996, Milward 1994; for in these the assumption that
natural languages are formal systems of inference is preserved as the central
methodological assumption; and the process of interpretation is, accordingly,
an incremental process of projecting model-theoretic content. In our system,
to the contrary, the incremental process is one of building the structure over
which a particular model-theoretic content is to be dened; in consequence, the
invocation of structure is ineliminable.
The present model is unlike most orthodox formalisms, in making left-to-
right projection of structure central to the explanation. The importance of this
will emerge in the course of the next chapters. Unlike most other formalisms,
moreover, specically HPSG (see Pollard and Sag 1994) and LFG (see Bresnan
1982), there is no structure dened over the string itself. A logical form paired
with a string is, rather, a progressively projected representation of a content for
that string.
In dening the interpretation process in irreducibly structural terms, the
model is similar to Discourse Representation Theory (DRT; see Kamp and
Reyle 1993). Conceptually, it is very close to DRT, sharing with that paradigm
a commitment to modelling the process of building up interpretations. How-
ever, unlike DRT, the sequencing intrinsic to the process is not merely sentence
by sentence, but word by word. The interpretations, once completed, have a
bottom-up compositional semantics, as does a DRS (discourse representation
structure). The dierence between the two formalisms is that in our formalism
it is the left-to-right process of building up the structure which itself provides the
basis for syntactic explanation. And the partially top-down process of structure-
building, far from being problematic for the preservation of compositionality,35
is fully consonant with the tree growth concept of compositionality advocated
here.
In dening the central role of tree growth, there are some similarities with
the Minimalist programme (in particular, given the Minimalist Programme rule
of Merge that induces growth of tree structure, see Chomsky 1995); but in the
34 For a discussion of this distinction in comparing logic formalisms, see Gabbay and Hod-
kinson 1992.
35 For discussion of compositionality in DRT see van Eijck and Kamp 1997.
present system, there is no movement, and furthermore, tree projection is at

least in part a top-down process with anticipation of structure. This integration
of linearity dierentiates it markedly from the strictly hierarchical restriction
imposed on syntactic forms of explanation within that system and its conse-
quent bottom-up projection of structure (see Chomsky 1995). Moreover, in the
model outlined here, there is no requirement of a one-to-one correspondence
between lexical elements and nodes in a tree. Thus, to take an arbitrary exam-
ple, where Minimalist accounts might project a node both for a clitic pronoun
and, separately, for the item it is a copy of, in this system we can formulate an
account of anticipatory clitic pronouns as projecting a requirement which the
full form of NP fulls; so there is no need to multiply functional projections
to accommodate the various clitic and full-form positions in a sequence (see
Sportiche 1992, Manzini and Savoia 1998). Quite generally, discontinuous de-
pendencies can be expressed in terms of an initially underspecied node and its
later specication. And binding of discrete nodes in a tree structure is replaced
by the various tree update mechanisms.
In its projection of tree descriptions that are simply sets of attributes real-
izable as annotated trees, the system has obvious aÆnities with HPSG. HPSG
is a system in which all concepts of structure are restated as tree descriptions,
comprising a system of constraints on natural language strings. Within this,
features are dened as percolating up and down a tree relative to specied con-
straints, thus supercially corresponding to the concept of an unxed tree node
and its subsequent resolution as articulated here. Two design properties distin-
guish the two formalisms. First, though an HPSG grammar could be dened to
express left-to-right tree growth, this would be at best peripheral. There is no
dynamics within the system other than unication of feature complexes. There
is, accordingly, no concept analogous to the dynamics of creating a node and
imposing on it requirements for subsequent processes to satisfy. Indeed there is
no notion of `subsequent' with which to express this. There is no concept of a
fully annotated node lacking a xed position and awaiting update at some later
point. Secondly, there is in HPSG no necessary correspondence between syn-
tactic and semantic vocabulary; all correspondences require explicit denition;
and there is no restriction on types of syntactic features. In consequence, there
is no principled way to restrict the typing of syntactic features. The present
system, to the contrary, rests on the assumption that the structures induced are
semantically transparent and subject to orthodox semantic evaluation, with the
bonus of capturing the partiality of specication integral to natural language
expressions.
Of the major frameworks, it is Lexical Functional Grammar (LFG) which
shares with the framework proposed here the characterization of long-distance
dependency as a form of underspecication of syntactic description representable
using the Kleene star operator. Kaplan and Zaenen's (1989) concept of func-
tional uncertainty is in this respect a precursor to this analysis. LFG articulates
a number of levels of representation, each with its own formal primitives and
geometry with rules of structural correspondence. There is c-structure, which
is the phrase-structural characterization of the surface string; f-structure, which
Summary 101
is a level of structure representing the abstract syntactic functions of the sur-
face forms; a-structure, which is a level of structure representing thematic roles
in a thematic hierarchy (see Bresnan and Moshi 1990); and a level of semantic
structure which represents semantic content directly. It is f-structure over which
the concept of functional uncertainty is dened. An f-structure is a hierarchy of
attribute{value pairs in which each attribute, being a syntactic primitive (e.g.
SUBJECT, OBJECT, COMPLEMENT), is associated with a single value, that
value being either another f-structure or an atom (the lexical items of the lan-
guage). A distinguishing property of such f-structures is that nothing precludes
multi-dominance, all that is required being a single value for each attribute. So
two discrete f-structures within some overall f-structure may be identied { for
example, the f-structure assigned to some attribute TOPIC and that assigned
to some discrete attribute COMPLEMENT OBJECT. Hence the identity of
topic constituent and the complement object in John, Bill criticized at the level
of f-structure. Functional uncertainty is then dened as an indenite sequence
of attributes between one f-structure and another, characterized through the
Kleene star operator as a variable over indenite sequences of attributes. What
notably distinguishes this concept of functional uncertainty from the concept of
merging an unxed node with a xed node, as proposed in this framework, is the
lack of dynamics in the Kaplan and Zaenen account, with no corresponding con-
cept of uncertainty at any level representing interpretation. To the contrary, as
an abstraction from the surface sequence of expressions, f-structures must con-
tain lexical items of the language under description amongst the values assigned
to its attributes. In this, the assigned f-structure is a level of structural analysis
for strings, albeit not in their surface sequence. Anaphoric expressions, for ex-
ample, are retained at f-structure as values of a PRED attribute for SUBJECT,
OBJECT, etc., with no replacement of him by some other expression.36 What
LFG has instead of a concept of tree construction is a number of distinct levels
of representation. But the correspondence rules between the discrete levels do
not incorporate the dynamics of interpretation either, because they are dened
as relations between complete annotated congurations of quite dierent formal
type.
The only framework we know of which incorporates left-right processing
into the grammar formalism is the Left-Associative Grammar of Hausser (1989,
1999), in which branching structure is built up on a word-by-word basis, such
structures re ecting simply the relation between each word and the following
remainder. Rules dene transitions in terms of some assignment of word cate-
gory plus a rule package which determines possible completions for the string,
with additional restrictions imposed by valency statements associated with, for
example, verbs. This grammar, which, according to Hausser, can be manipu-
36 The structural nature of the constraints on anaphoric interpretation of Dalrymple (1993)
are equivalent to the constraints articulated here, also dened in terms of functional uncer-
tainty (what she calls ìnside-out functional uncertainty'). It is relative to these constraints
that evaluation of the anaphoric expressions is determined, but these constraints do not, as
articulated here, serve as input to any structural replacement of the anaphoric expression by
the expression which constitutes its antecedent.
lated both as a parser and as a generator, denes a mapping between a string

and a data structure representing a predicate and its arguments, and these data
structures constitute the interpretation made relative to context by a matching
process between a data structure constructed from the sentence and a (richer)
data structure corresponding to the context. There are two formal anticipatory
devices which are the analogue to our concept of requirement { the concept of
rule package and the concept of valency specications (that need to be checked).
However, because the rules are dened to yield strings of words, with the re-
sulting data structures as the only concept of structure, the system is not well
set up to express syntactic generalizations other than those which are a re-
ection of such predicate argument arrays. In consequence, though Hausser's
left-associative grammar is dynamic, it is nevertheless essentially a-syntactic.
To the contrary, in our framework, the explicit use of modal logic in combina-
tion with the notion of transitions across partial trees provides us with a tool
for constructing structure on a top-down basis in a general way.
As we shall see in the following chapters, the attention to the detailed dy-
namics of how structure is built up not only provides a cross-linguistically
valid framework for syntactic analysis; it also provides a means of capturing
cross-linguistic variation.
4
Linked Tree Structures

4.1 Relative Clauses { Preliminaries
We have so far only looked at relatively straightforward constructions { binary
branching structures and no more than the simplest examples of long-distance
dependency. Nevertheless, we now have in place a framework that in princi-
ple enables us to describe the development of all relations within a single tree.
We have dened tree development in terms of introducing a node with cer-
tain requirements, and, subsequently, fullling those requirements. Within this
framework, we have focused on concepts of partiality and development integral
to tree growth. At the level of tree structure, we have dened several forms
of incomplete tree description, with trees being incomplete because some nodes
retain unfullled requirements, or because not all tree relations are fully iden-
tied, in each case subject to later development. At the level of the individual
formulae, we have introduced a concept of an incomplete formula, with a meta-
variable and its associated substitution. And we have a sketch of a lexicon in
which lexical information is dened in terms of idiosyncratic sequences of tree
constructions and updates.
This picture is very far from complete, however. A dening property of
natural languages is their varied potential for building up complex concepts from
simple ones. From the concept `man', we can construct the concept `man that
I like', `man that my father thinks I will like'. Furthermore, having established
some referent, we can also build structures that add information { extending
`John' to `John, who I like', or to `John, who my father thinks I will like'. And
so on. The development of a formula by introducing meta-variables is thus only
the tip of an iceberg. Complex concepts are built up through the development of
tree structures, and we need to dene ways in which the information annotating
(the root node of) a tree structure can modify the information at a distinct tree
node. Thus, we have to incorporate the concept of families of trees.
The interest of an extension to families of trees is that it is at this level that
the interaction between the various rules and actions is most in evidence. A
103
104 Linked Tree Structures
prime example is the interaction between anaphora resolution and the process
of establishing a xed position in a tree for an initially unxed node in the
interpretation of relative clause sequences. The substantive interest of such an
extension is that it involves cross-language variation. Despite the universality
of concepts of long-distance dependency and anaphora resolution, we get con-
siderable variation across languages as to how these two processes interact. As
we shall see shortly, it is by paying close attention to what information is made
available at each step of the tree development that we shall have a systematic
basis for characterizing the range of cross-linguistic variation. In this chapter we
set out an account of how relative clauses are parsed, showing how the account
to be developed for one language can be extended, with minor variations, to
characterize the multiplicity of forms displayed across languages.
Our next formal move, then, is to extend the tree description language so as
to allow multiple trees associated with a single sentential sequence { linked tree
structures. The concept of linking tree structures involves the projection of two
tree structures in tandem and a ow of information from one tree to another.
To take the most transparent case by way of example, consider informally the
step-by-step process involved in processing:
(4.1) John, who I like, criticized Hegel.
The building up of an interpretation for this sequence involves building a pair of
structures { one for John criticized Hegel, a second for who I like. This process
is engineered by the relativizing element who, which uses the representation of
John in the rst structure to construct a copy of this representation in the sec-
ond structure, thus securing a `link' between the two structures in the form of a
shared formula. This quasi-anaphoric process associated with an initial wh ele-
ment leads to the presence of an unxed node in the second tree whose position
then has to be resolved { exactly as in left-dislocation sequences considered in
chapter 2. This sequence of actions apparently required for English is, however,
not the only way of securing appropriately linked tree structures, and it is the
dynamics of how such linked tree structures are incrementally projected from
which the dierent types of relative structure and their associated restrictions
will emerge. As we shall see, the initial underspecication of the tree relation
between the node projected from a wh expression and those of its contiguous
expressions in the string remains central to the account. It is just that there is
something else going on in addition { the process of incrementally building up
linked tree structures.
4.1.1 The LINK Relation

The tree-building process that we shall be dening will involve a `sideways'
move, attaching to a node annotated by some Formula value the root-node of
a new tree. This second tree is stipulated to have the same as an annotation
of one of its nodes. That is, given a node description with address X , the head
node, at which the pointer, }, is located,
f::::fX; :::; F o(); T y(x); }gg
Relative Clauses { Preliminaries 105
the root node of a second tree will be introduced by a LINK relation. The
second tree will eventually have some node annotated by . That is, when the
process of interpreting the relative sequence is complete, we will have a structure
satisfying
f::::fX; :::; F o(); T y(x)gg fhL 1iX; T y(t); # F o(); }gg:
Here hL 1iX establishes that a node with address X can be reached along a
LINK relation (to be dened). We say that the root node of this new tree, hence
the tree itself, is linked to its head in the rst tree. The introduction of the root
node of the second tree linked to the head node we call LINK Introduction.
In the course of this chapter we will introduce a variety of LINK Introduction
rules, all diering in how they eect the common feature of creating a linked
tree which somewhere shares a Formula value with the head node.
The shared element can be carried over embedded in a requirement on the
root node of the new linked tree, or a copy can be constructed as an annotation
on some node constructed as part of the LINK transition itself. This copying
process may or may not be subject to locality restrictions on where the copied
formula can occur in the tree. Moreover, the main structure containing the
head node may be initiated before the LINK operation (as in English), or the
process can be dened in reverse. All these variations are expressible within the
framework, and lead to a view of relatives which provides a basis both for char-
acterizing the array of possibilities and for itemizing idiosyncratic subvariants
within dierent languages.
Because we wish to take the phenomenon of relatives as a means of exploring
the range of cross-language variation, in this chapter we shift to a more linguis-
tic form of presentation. Now that we have delineated a framework to provide
a background set of assumptions, we can start from the data with certain ex-
pectations as to the type of solution to be envisaged. Having surveyed a span
of data in a single language (English), we shall dene rules of transition and
lexical specications necessary to characterize these (section 4.2), and then use
this account as a point of departure for setting out the range of cross-language
variation observable across languages (section 4.3). We discuss relative struc-
tures which invariably make use of resumptive pronouns, head-nal relatives
and the so-called head-internal relatives. Following this route, we aim not only
to set out the beginnings of a formal framework, but also to give a taste of how
it can address the empirical test of characterizing a variety of cross-linguistic
data in a principled and revealing way.
4.1.2 The Data Reviewed
The rst step in characterizing relative constructions is to remind oneself of the
wealth of variation in the phenomenon of relative clause construction. Rela-
tive clauses constitute the primary means in natural languages for elaborating a
given expression by the addition of a modifying structure, of arbitrary complex-
ity. In some languages, relatives occur with the modifying structure following
the head (as in English), in some with the modifying structure preceding the
head (Japanese, Hindi, Chinese), in yet others with part of the head preceding
the modifying structure, part of it following (Chinese):
(4.2) Hanako ga kaita hon [Japanese]
HanakoNOM writeP AST book
`the book that Hanako wrote'
(4.3) ai ta muqin de nei ge ren [Chinese]
loves his mother rel that classifier man
`that man that loves his mother'
(4.4) nei ge ai ta muqin de ren
that classifier loves his mother REL man
`that man who loves his mother'
In some languages (German and optionally English), the relative structure is ex-
plicitly introduced by an anaphoric-like expression, either wh (English), demon-
strative (Dutch and German), or denite (Arabic) in form, carrying over infor-
mation from the head node into the modifying structure:
(4.5) l- mudarris illi mabsu:t [Egyptian Arabic]
the teacher that happy
`the man that is happy'
As we shall see in due course, these languages vary as to how strong this
anaphoric property is, this correlating with obligatory presence or absence of a
resumptive pronoun. Thus in Egyptian (and other forms of) Arabic, all non-
subject positions in a relative require a resumptive pronoun:
(4.6) l- mudarris illi Magdi darab-u
the teacher that Magdi hit him
`the teacher that Magdi hit'
In other languages (Hebrew) the complementizer may not carry any structure-
specic encoding { it is the general subordinating device that is used to mark
the relative (note the two occurrences of se in (4.7)):
(4.7) ha-?-is se xasavt
p se Dani pagas (?oto)
p
the-man that thought2:Sing:F em that Dani met (him)

[Hebrew]
`the man that you thought Dani met'
In some languages there may not even be any expression marking the relative
boundary. In English, the relativizer is optional (except in subject position):
(4.8) the man I like
Relative Clauses { Preliminaries 107
In Japanese, regular relative constructions never use a relativizing particle (see
example (4.2) above). In Arabic, relatives with an indenite head never use a
relativizing particle:
(4.9) *mudarris illi Magdi darab-u [Egyptian Arabic]
teacher that Magdi hit him
à man that Magdi hit'
(4.10) mudarris Magdi darab-u
teacher Magdi hit him
à man that Magdi hit'
In many languages, relative structures may be expressed apparently without
any head:1
(4.11) John wa Mary ga muita no o tabeta [Japanese]
JohnT OP MaryNOM peeledNom:ACC ate
`John ate what Mary peeled.'
In yet others, the head may apparently occur within the relative clause rather
than within the main clause (Korean, Japanese):2
(4.12) John wa Mary ga ringo o muita no o tabeta
JohnT OP MaryNOM appleACC peeledNom:ACC ate
`John ate the apple Mary peeled.'
Even in a single language, there may be a wide variety of possibilities. In
English, there are the so-called non-restrictive relatives which appear to be a
form of modication of the noun phrase as a whole:
(4.13) John, who I like, smokes heavily.
and the so-called restrictive relatives, which are a modication of some part:
(4.14) The man who I admire least smokes heavily.
Relative clause modication may not always require explicit introduction:
(4.15) The man I admire least smokes heavily.
There is more than one form of relative pronoun, the set of wh relatives {
who, which (homophonous with the wh question words) and the subordinating
particle that (homophonous with the demonstrative that ):
(4.16) The man who I met yesterday is a nurse.
1 The nominalizer in Japanese required for such `headless' relatives takes the form of the
genitive. Note the use of `Nom' to describe a nominalizer particle, `NOM' to describe nomi-
native case.
2 There is dispute over whether these structures are relative clause structures at all. See
section 4.3.4.
(4.17) The man that I met yesterday is a nurse.

These forms may or may not be case-marked. In English, this is vestigial,
restricted to the use of whom, itself increasingly restricted to contexts in which
other forms are debarred:
(4.18) ?The man whom I met yesterday is a nurse.
(4.19) The teacher to whom the letter had been addressed was not available.
And, like Romance languages, English displays the complexity of pied-piping in
which a wh expression occurs as part of an arbitrarily complex NP:
(4.20) Minerva, under the wheels of whose chariots many men have fallen, is
the most powerful goddess of our time.
What all such structures display, across English and other languages, is that at
the level at which the interpretation is represented, the interpretation of both
main clause and relative clause must have one term in common, whose presence
in the two structures is dictated by some grammar-internal process.
Analyses of relative clauses do not present a homogeneous account of this
range of phenomena, leaving its diversity as a puzzle. The restrictive and non-
restrictive forms of construal (see Chomsky 1981, Sar 1986, 1996, Fabb 1990,
Borsley 1992, Espinal 1991) are generally treated as quite dierent in structure.
The problem, given that constituency is dened over the string, is that such
non-restrictive relatives provide no evidence of being adjoined to any of their
surrounding constituents.3 In movement accounts, the non-restrictive relative
constituent is said by some to be attached at a level of LF0 (see Sar 1986), by
some at a level of discourse structure (see Fabb 1990), and by Espinal (1991)
even to be represented by a system of multi-dimensional trees. Given that LF
is taken to be the interface level between syntax and semantic evaluation, it
is hard to know what status to assign to LF'; and some analyses are openly
agnostic about the status of the level required (see Fabb 1990 in particular).
Despite the widespread informal evocation of a concept of discourse grammar
for structures such as these, the concept remains quite unclear, both from the
perspective of a grammar whose remit is to characterize sentence structure, and
3 As Fabb (1990) points out, the determiner-NP sequence which is visibly the head of the
relative clause does not, in a non-restrictive construal, appear to c-command the following
relative and so form a constituent with it, and such relatives do not seem to stand in any
regular relation to other, following constituents. Unlike (i), in which an additional constituent
leads to ill-formedness, (ii) is well-formed:
(i) *I gave my brother my sister a lift to the station the garage.
(ii) I gave my brother, who had a heavy suitcase, a lift to the station, which was closed.
A relative, non-restrictively construed, cannot even be analysed as c-commanded by the top-
most node, as such a structure in (iii) would by standard assumptions preclude the indicated
interpretation by principle C, given that the antecedent clause and relative pronoun would be
co-indexed and would mutually c-command each other:
(iii) I saw my brother, which was a shock.
The Analysis { A Sketch for English 109
from the perspective of a pragmatics whose remit is to explain all other in-
terpretation phenomena as the product of a generalized reasoning system (see
Sperber and Wilson 1995). The only analysis within movement frameworks
which approaches the goal of providing a treatment for relative structures that
draws together head-initial and head-nal structures as minor variants of a sin-
gle pattern is that of Kayne (1994), with its concomitant disadvantage of highly
abstract structures and very rich processes of movement. In HPSG accounts
on the other hand (see Sag 1997), there are detailed accounts of individual lan-
guages, but no attempt to provide an overall typology of the range of structures
made available in dierent languages. And in categorial grammar, the range of
coverage is restricted to English nite relative constructions (see Morrill 1994),
with, equally, no obvious way of extending the accounts provided to a broader
range of data. The challenge is to provide an account which provides both a
general framework of possibilities and a principled account of why and how the
diversities arise.
4.2 The Analysis { A Sketch for English

Our starting point is, as before, to view the phenomenon within the context of
the construction of a logical form, a build-up of information on a left-to-right
basis. Viewed from this perspective, a single thread runs through the disparate
English forms of relative clause. The construction process involves taking a node
in some partial tree structure as the head, and building from it the top node of
a new tree which has imposed on it a requirement for some dominated node to
be annotated with a copy of the formula annotating the head. The initial wh
expression, by projecting a copy of the formula annotating the head, guarantees
the presence of this formula in the new tree. Exceptionally in English, the copy
of the formula of the head may be directly constructed without any relativizing
particle as trigger (the phenomenon very generally analysed by invoking an
empty complementizer). Once the second tree is completed, the construction of
the `host' tree proceeds as before. The result is a pair of linked trees, in which
a formula annotating a node in one tree is identied with a formula annotating
a node in the other.
To characterize these processes, we extend the language DU with two fur-
ther modal operators, hLi and its converse hL 1i. hLi is an operator which
is interpreted along a relation connecting an arbitrary completed node in one
tree to the top node of a new tree: the LINK relation. Also dened are two
modal operators, hDi and hU i, which represent the union of the dominance and
LINK relations and the union of the converse of the dominance and converse of
the LINK relations, respectively. These underspecied operators generalize the
dominance modalities. For instance, the proposition hDi evaluated at some
tree node not only has the dominated nodes as its area of evaluation, but may
cross over into a second (and third) linked tree to continue its evaluation there.
h# i , on the other hand, cannot cross a LINK relation, as h# i is interpreted
over the re exive transitive closure of the daughter relation (#) alone.
For the dynamics of the tree construction, we dene a variety of rules which
all have in common that, at some point in the construction process, a second tree
is initiated by introducing a top node with requirement ?T y(t), which cannot be
completed without there being some node annotated by a given formula from
the rst tree. A rule of this type is called a LINK Introduction rule. As we
shall see, languages vary in the way the constraint of a shared formula in a
pair of trees is implemented; indeed, we shall also see variation within a single
language.
4.2.1 Dening Linked Tree Structures
The rst step is to formulate the extensions to the structures and the language
we have dened in chapter 2, which are required to deal with the notion of linked
trees.
Denition 11 (Link Relations and Link Modalities) A Linked Basic Tree
structure LBT R T is a nite set of partial trees T1; : : : ; Tn with disjoint Tree
Domains and a Link relation L S1in T rDomi S1in T rDomi such
that, if n L n0 holds between n 2 T rDomi ; n0 2 T rDomj , then i 6= j . More-
over, there is no loop Ti L : : : L Ti in T (where Ti L Tj , if there are
n 2 T rDOMi ; n0 2 T rDomj : n L n0 ).
The language DU has modalities hLi and hL 1i with interpretation
T ; n j= hLi i 9n0 : n L n0 & T ; n0 j=
T ; n j= hL 1 i i 9n0 : n0 L n & T ; n0 j=
On the basis of the Link relation we dene the relation D as the re exive and
transitive closure of [ L and introduce the modalities hDi and hU i with
the denition
T ; n j= hDi i 9n0 : n D n0 & T ; n0 j=
T ; n j= hU i i 9n0 : n0 D n & T ; n0 j=
Now, a Linked Partial Tree Structure is a structure that can be completed to
a Linked Basic Tree Structure. Because the Link relation L is a fully speci-
ed tree relation, we will expand the domain of the tree node predicate T n to
include values including a link step.
hL 1 iT n(a) () T n(aL),
so when a LINK structure is built from some head node with address T n(a),
the root node of that LINK structure can be identied as T n(aL).
Given a node n in a Linked Partial Tree Structure T , we call the local tree
with respect to n the set of nodes fn0 2 T j n i n0; i 2 f#0; #1; #; #gg. That
is, the local tree with respect to node n is the set of nodes that can be reached
without crossing a Link connection. On the other hand, the global tree with
respect to a node n 2 T is the set of nodes that can be reached from n by i,
where i may be any sequence of indices.
4.2.2 Relativizers Annotating Unxed Nodes
Of the analyses we give in terms of linked tree structures, the rst is English,
in which the relative element is a pronominal-like device, annotating an unxed
node constructed within the newly induced LINK structure. We take non-
restrictive relatives (4.21) as providing the clearest form of the pattern we want
to analyse. In (4.21), the formula projected from John is copied over into the
tree associated with who I much admired, leading to an interpretation of (4.21)
as equivalent to (4.22):
(4.21) John, who I admire, has left MIT.
(4.22) John has left MIT, and I much admire John.
The copying process is ensured by the relative complementizer. These com-
plementizers, appropriately called relative pronouns by Jespersen (1961), are
dened to transfer information from one structure to another. The rule of
LINK Introduction applies to a node in a tree decorated by some declarative
unit h: : : ei : and induces the root node of a new tree in whose subsequent
development a copy of is required to occur. The resulting relation between
this node and the new root node is the LINK relation, the tree so introduced
the LINKed tree.4 This new tree has, by denition, a requirement imposed on
it not merely to derive a formula of type t, but more specically to establish
within it a node at which a copy of occurs.
In English we have to ensure the initial position of the wh expression in the
clausal sequence, albeit possibly embedded within a containing constituent, so
we dene the LINK Introduction rule for this language as creating both the
root node of the LINK ed tree and an unxed node requiring a copy of the head
formula, so that ensuing development of the LINK ed tree may follow the same
sequence of transitions as displayed in left-dislocation structures. On this rst
formulation, we assume that the unxed node, when the requirement of con-
structing a copy of the head is fullled, will be fully annotated:
LINK Introduction (Preliminary)
z
head
}| {
f: : : fX; F o(); T y(e); }gg
f: : : f| X; F o({z); T y(e)g}; fh
|
L 1iX; ?T y(t)g; fh"ihL 1 iX; ?F o(); ?T y(e); }gg
{z } | {z }
head linked node unxed node
The eect of this LINK Introduction rule applied in the analysis of (4.21)
is as follows. It takes the partial tree resulting from having parsed John and
creates the structure in gure 4.1.5 Note the imposition of the requirement
4 We shall refer to the resulting pair of tree structures as a pair of linked structures.
5 As before, throughout this chapter, the tree displays are essentially informal, and the
complete description is given only in the list below.
fT n(a); ?T y(t)g
fh"0 iT n(a); f?T y(e ! t)g

F o(John)g
fhL 1 ih"0 iT n(a); ?T y(t)g
fh" ihL 1 ih"0 iT n(a); ?F o(John); }g

f fT n(a); ?T y(t)g
fh"0 iT n(a); F o(John); T y(e)g
f fhL 1ih"0 iT n(a); ?T y(t)g
fh" ihL 1 ih"0 iT n(a); ?F o(John)g g:
Figure 4.1: Building a LINK transition with LINK Introduction
?F o(John) on the unxed node in the newly induced LINK ed structure. In

this simple case, the English wh relativizer, a wh relative pronoun, projects a
copy of F6o(John) as an annotation on the unxed node, securing a copy in the
new tree. The result of processing John, who in (4.21) then gives:
f fT n(a); ?T y(t)g
f fhL 1 ih"0 iT n(a); ?T y(t)g
fh" ihL 1 ih"0 iT n(a); F o(John); ?F o(John); T y(e); [#]?; }g g:
Now Thinning removes the proposition ?F o(John) from the bottom node de-
scription. This sequence of actions updates the tree description, leaving the
task of establishing a more precise tree node description for this node exactly
as in left-dislocation structures or wh questions. As in those constructions, the
resolution of the precise relation between the as yet unxed node and the emer-
6 The anaphoric role of wh expressions as relative pronouns is discrete from their role in
questions, where they are dened as meta-variables to be retained in the resulting formula
(see chapter 5).
gent tree is constrained by the fact that the relation can only connect nodes
within the local tree (it cannot traverse a L connection between nodes).
Re ecting this sequence of actions, the lexical actions projected by the En-
glish relativizer who can be dened as follows:7
whorel IF f?T y(e); h"ihL 1 iF o(x)g
THEN put(F o(x); T y(e); [#]?)
ELSE ABORT
For the more complex case, such as whom and which, which can occur nested
within a fronted (pied-piping) construction,
(4.23) The oor, the surface of which was scratched, is due for renovation.
(4.24) The political prisoner, on behalf of whom large numbers of letters have
been written, is being released today.
the characterization of the relative position of the node which the variable to
be bound to the head annotates will need the more complex modality:
" h" ihL 1 iF o(x);
this being a requirement that at some unxed node (indicated by h"i), and pos-
sibly embedded within it (indicated by "), constructed from a node LINK ed
(indicated by hL 1i) to some head annotated with some formula value, there
be the appropriate copy:8
LINK Introduction (nal version)
z
head
}| {
f: : : fX; F o(); T y(e); }gg;
f: : : f| X; F o({z); T y(e)g}; fh
|
L 1iX; ?T y(t)g; fh"ihL 1 iX; ?T y(e); ? # F o(); }gg
{z } | {z }
According to this specication, LINK Introduction induces an unxed node
with a requirement for a copy of the head formula to occur in a dominated
position. The projection of the relative pronouns which and whom which oc-9
cur in pied-piping constructions will, accordingly, have to be more complicated:
7 Notice that the value of the Formula predicate is bound to the variable x, as dened in
chapter 3, denition 10.
8 Recall the distinction between internal and external uses of `" ', in which the latter is
interpreted over the dominance relation, and the former over the relation which need not
(yet) be the re exive transitive closure of the immediate dominance relation.
9 We ignore here the peripheral use of which as a relativizing determiner, for simplicity:
(i) ??I read the book, which monstrosity I refused to circulate.
However, in principle, a lexical characterization of which as a determiner is unproblematic.
The complication which its denition presents is analogous to the determiner the, which is
dened to take the nominal meta-variable and lift it to become a formula annotating the root
node of the local tree of type e (see section 7.3.1.1).
whichrel IF f?T y(e); " h" ihL 1 iF o(x)g

THEN put(F o(x); T y(e); [#]?)
ELSE ABORT
whomrel IF f?T y(e); " h" ihL 1 iF o(x)g
THEN IF ?h"0iT y(t)
THEN ABORT
ELSE put(F o(x); T y(e); [#]?)
ELSE ABORT
Whom is precluded from occurring in subject position, but may otherwise occur
freely. So in (4.25),
(4.25) John, the mother of whom I distrust, has arrived.
the LINK Introduction (nal version) rule induces the root node of a LINK ed
structure:
fhL 1iX; ?T y(t)g
and an unxed node with the requirement that it dominates an occurrence of
the formula John :
fh"ihL 1 iX; ? # F o(John); ?T y(e)g
The projection of structure by the determiner and the common noun Mother
(the details of which we will return to immediately) then induce a sub-tree
containing a node for object of Mother (much like a transitive verb), as in
gure 4.2.10
fhL 1 iX; ?T y(e)g
fF o(P (the P )); f?T y(cn)g
T y (cn ! e)g
fF o(U); T y(e)g f?T y(e ! cn)g
f?T y(e ! cn)g fTFyo((eMother );
! (e ! cn))g
Figure 4.2: The internal structure of the mother of whom
The object node now provides the conditions for the relative whom to enter
F o(John) as an annotation.
10 As elsewhere, this tree display presents only those aspects of the tree description that
are relevant, suppressing all others { e.g. tree node identier, scope relations, and so on.
We characterize mother as projecting here a two-place term Mother. See section 4 for a
characterization of genitives.
According to this analysis, the essence of English relative complementizers is
their encoded anaphoric property constructing a copy of some antecedent head
formula { rather than any parallel with quantier binding, as is standard in
categorial, GB and HPSG formalisms.
The application of the LINK Introduction rule has been exemplied by
non-restrictive relatives, but restrictive relative clauses are also projected as
LINK ed structures, and by exactly the same Introduction process. In this
case, the LINK relation is initiated from a node annotated by a variable pro-
jected from a common noun.11 In the analysis of quantication, we use variable
binding term operators of type cn ! e (projected from determiners), which bind
nominal variables occurring within a nominal formula of type cn. The specics
of quantication will be taken up in chapter 7, but in gure 4.3 we display
the internal structure of cn representations and a lexical entry for man. Nodes
annotated by T y(cn) immediately dominate a node annotated with a nominal
variable of type e, and a node with the restrictor formula, of type e ! cn, the
entire construct projected by the common noun.12 In restrictive relative con-
fT y(cn); F o(U; Man(U))g
fF o(U); T y(e)g fF o(X (X; Man(X )));
T y (e ! cn)g
The top node is annotated by the nominal.
The left daughter is annotated by the nominal variable.
The right daughter is annotated by the restrictor.
man IF f?T y(cn)g
THEN make(h#1i); go(h#1i); put(F o(Man); T y(e ! cn); [#]?);
go(h"1 i); make(h#0 i); go(h#0 i); put(F o(U); T y (e))
ELSE ABORT
Figure 4.3: Internal nominal structure projected by man
struals, it is the nominal variable, i.e. U in gure 4.3, which is taken as the
head formula of the LINK ed structure, and which is copied over as a require-
ment (see gure 4.4 for a display of the processing of the LINK transition in
(4.26)):13
11 To describe the type cn as the `common noun' type is misleading, given the internal
structure of nodes so annotated. We shall restrict the term common noun to nouns in the
object language string. The formula of type e ! cn which such common nouns project, we
shall refer to as the restrictor; the variable with which the restrictor combines we shall call
the nominal variable; and the formula annotating the mother node of type cn we shall refer
to as the nominal.
12 In the main, in English, the trigger for this sequence of actions is ?T y(cn). In structures
lacking a determiner, which is the more common cross-language pattern, the trigger would be
?T y(e), and in such cases the sequence of actions would include the building of a determiner
node annotated by the lambda term F o(P (; P )) involving an epsilon operator.
13 Lexical items may distinguish between such discrete forms of relative construal. The
relative that, for example, can only induce restrictive relative construals, and may not occur
f?T y(e)g
fF o(P (the P )); f?T y(cn)g

T y (cn ! e)g
fF o(U); T y(e)g fTFy(oe(Y (Y; Man(Y )));

! cn)g
fhL 1 iX;
?T y(t); }g
fh" ihL 1 iX;

F o(U); T y (e)g
Figure 4.4: Structure resulting from parsing the mani whoi in (4.26)
(4.26) John ignored the mani whoi Sue likes ei.

The relative pronoun in such structures causes the formula of the head, here
a nominal variable, to be copied as an annotation on the unxed node; the
subsequent process is then as for non-restrictive relatives. The interpretation
of these types of LINK ed structures, however, once set out, diers. The tree
structures projected by non-restrictive relative clauses are pairs of trees whose
root nodes are decorated with propositional formulae sharing a common term.
In the case of restrictive relatives, by contrast, where the shared term is a
nominal variable, the completed LINK ed structure combines with that of its
head to give a complex restrictor. In gure 4.5 we display this process for the
sequence man who Sue likes. This compound formula, (x; (x) ^ (x)), of type
cn, combines with some lambda term (of type cn ! e) to yield a term of type e
of the form (; x; (x) ^ (x)), where is a variable binding term operator (like
), which remains in this skeletal form until scopal dependencies are resolved at
the root node (see chapter 7). Notice the function here of the LINK ed tree in
extending the formula at some antecedently constructed node.
Thus, for the construal of the relative in (4.26), with the pointer at a node
requiring type e, the determiner imposes a requirement for a daughter requiring
T y(cn) and creates a daughter which it annotates with a formula of type cn ! e.
The common noun man then induces a pair of daughter nodes, one annotated
in pied-piping constructions.
fT y(cn); LINK fLike(Sue;U)g
F o(U; Man(U))g
fF o(U)g fF o(X (X; Man(X )))g

fT y(cn); LINK fLike(Sue; U)g
F o(U; Man ^ Like(Sue; U))g
fF o(U)g fF o(X (X; Man(X )))g

Figure 4.5: Combining a linked structure with the head
with the restrictor X (X; Man(X )) and one annotated by some fresh nominal
variable U. This second node is LINK ed to the top node of a second tree, and
the nominal variable is transferred as usual. With due processing of the following
relative clause, the formula F o(Like(U)(Sue)) annotates the root node of the
LINK ed structure. Now, the conditions for combining the nominal content with
that of the LINK ed structure are met, yielding a formula of type cn:
F o(U; Man(U) ^ Like(U)(Sue))
annotating the head node. This formula will then combine with the formula
projected by the determiner to yield a term of type e (see chapter 7 for more
details of our account of quantication); and the nal tree is as schematically
displayed in gure 4.6 (in which all details other than F ormula values are
suppressed).
The account of English relatives proposed allows a very natural, though
language-idiosyncratic, extension to complementizer-less relative forms. All that
is needed is the option of allowing a copy of a formula from the host node to be
transferred over not as a requirement, but as an annotation:
LINK-COPY Introduction
z
head
}| {
f: : : fX; F o(); T y(e); }gg
f: : : f| X; F o({z); T y(e)g}; fh
|
L 1iX; ?T y(t); }g; fh" ihL 1 iX; F o(); T y(e); gg
{z } | {z }
F o(Ignore(the,x; Man(x) ^
Like(x)(Sue))(John))
F o(John) F o(Ignore(the,x; Man(x) ^

Like(x)(Sue)))
F o(the,x; Man(x) ^ F o(Ignore)

Like(x)(Sue))
o((U; Man(U))^
F o(P (the,P )) FLike (U)(Sue))
F o(U) F o(Man)
F o(Like(U)(Sue))
F o(Sue) F o(Like(U))
F o(U) F o(Like)
Figure 4.6: The nal result of processing (4.26) with restrictive relative
Notice that with the process of LINK dened as annotating the unxed node
with a copy of the formula of the host node, the eect in these structures is
exactly that of an empty complementizer.14
This analysis of relative clauses has two distinctive properties. Firstly, the
tree structure corresponding to the relative sequence is projected externally to
14 This characterization of complementizer-less relatives makes no reference to whether or
not the head node is part of a quantied phrase. It might be argued that such complementizer-
less forms are exclusively restrictive in construal. However, Swinburne (1999) argues (from
a slightly dierent formulation of the LINK transition rule) that examples such as (i){(ii)
involve the projection of a LINK ed structure for the small-clause predicate:
(i) A man came in late.
(ii) Everyone went home tired.
In particular, he points out that these have an E -type form of interpretation, requiring that
the entire term annotating the head is copied onto a LINK ed structure as the subject of the
small-clause predicate. Such examples also provide evidence against the observation of Fabb
(1990) that non-restrictive relatives do not modify quantied noun phrases.
the tree projected by the main clause.15 Secondly, the relativizing element is
pronominal in nature rather than quanticational; and the characterization of
relative pronouns as quasi-anaphoric captures both the similarities between rel-
atives and wh questions, and their dierences. Like wh expressions in questions,
relative pronouns project an annotation for a node whose nal position in the
tree is not yet xed, but unlike wh expressions in questions, which project a
meta-variable (W H ), relative pronouns serve as a channel, providing a copy of
the formula annotating the node from which the LINK ed structure is initiated.
Furthermore, unlike wh questions, there is no wh-in-situ relativizing pronouns
{ all relative pronouns must project a formula onto an unxed node. Notice,
incidentally, that there is no discrete logical type corresponding to the category
of complementizer. A further advantage of this analysis is its articulation of
the dierence between restrictive and non-restrictive relatives as simply the dif-
ference between constructing a copy within the independent structure of either
the formula annotating the root node of the sub-tree projected by some noun
phrase sequence or a nominal variable which is contained within it.16 The ar-
ticulation of this dierence does not require any invocation of structures licensed
15 There are a number of ways in which this LINK relation might be used for more general
application. First, given that the modal operator hLi is dened merely in terms of a relation
between nodes in a tree, we would expect also to dene LINK operations which involve no
transfer of information from the one tree to the other. Though we shall not pursue this
here, this provides a natural starting point for an account of co-ordination, which would have
the benet of debarring Merge applying to unify unxed nodes with some node in a second
conjunct of a noun phrase as in:
(i) *Who did John see Mary and?
Second, we might extend the present denition of the LINK Introduction process to allow
transfer of formulae of arbitrary type, imposing a requirement for a second occurrence within
that newly induced LINK ed tree of formulae of various types. Such an extension would allow
the following data to fall within the general rule:
(ii) The man was slovenly, which his brother was not.
(iii) John gave a book to Mary, to whom we hadn't expected him to give anything.
(iv) The man failed to clean up, which was annoying.
The generality of this phenomenon turns on a combination of factors { the process of LINK
itself; what types of expression are licensed in an unxed position at the projection of a
root node in a tree; the type of the expression triggering the transfer; and the allowed type
ambiguity of the relative pronoun, which in English can be one of several types { e, cn ! e,
e ! t, or even type t. Though we do not explore this in detail here, it is arguably the
intersection of these which gives rise to the apparent restriction on what can be transferred
across a LINK relation, not any restrictiveness intrinsic to the LINK process itself.
16 Sag (1997) argues that while complementizer-less relatives are N0 adjuncts, relatives with
an explicit complementizer are instances of NP adjunction (and hence requiring the analysis
we have given here to non-restrictive relatives), on the basis that restrictive relatives with
a complementizer can be constructed on nominals that have no obvious internal analysis in
specier and N0 :
(i) All who lost money were reimbursed.
However, nominal ellipsis is independently warranted for these constructions, given the well-
formedness of (ii), suggesting that at the level of interpretation the LINK relation is never-
theless dened over nominal variables:
(ii) All were reimbursed.
at dierent levels of the grammar, as in other frameworks (see section 4.1.2).17

Of the island constraints central to GB and Minimalist paradigms, this anal-
ysis expresses so-called strong island restrictions, through the characterization
of all left-dislocation structures as unxed within a single tree. As so charac-
terized, an unxed node (positioned by means of the relation) cannot be
merged across a LINK relation, precluding so-called extraction out of relative
clauses.18 The present analysis has nothing to say, however, about weak island
phenomena in which the resolution of one unxed node may con ict with the
resolution of another, and here we side with Pollard and Sag (1994) and consider
the supposed unacceptability of the data not to be clearcut.19 Acceptability of
such weak island data is known to be enhanced by a range of non-structural
restrictions (e.g. whether the initial wh form is `discourse-linked' (= D-linked;
see Pesetsky 1987), and in the light of this, we take (4.27){(4.28) not to be
structurally excluded:
(4.27) Who does John wonder what Mary gave?
(4.28) Which newspaper did John ask who Mary recommended to buy?
In chapter 6 we shall consider in much more detail a phenomenon which poses
the same puzzle of unclear data (weak crossover). Here we simply note that the
decision not to give a structural account of such phenomena does not preclude
cross-linguistic variation, as the translation-equivalent lexical items in dierent
languages may nevertheless project discrete sets of actions.
17 English also contains supposedly headless relatives, structures which { if this informal
characterization that they lack a head is correct { should be precluded on the LINK account
of English relative construal:
(i) What I learned from Anne was very valuable.
There are two possible analyses, either of which is compatible with our framework. On the
one hand, we could take the wh form as a specialized form of indenite head with a xed
LINK connection, following Bresnan and Grimshaw (1978). Or we could take these structures
as projecting exceptionally an empty head for which a particular form of what as a relative
pronoun needs to be made available, following Grosu (1996). Both of these analyses involve
some structure-specic stipulation.
18 Pollard and Sag (1994) argue that even pairings of a wh expression into some gap position
in a relative clause do not constitute an absolute constraint in virtue of examples such as (i):
(i) Pethadrine is the drug which I know someone who died of an overdose of.
We take these not to be well-formed. The judgement of (i) as acceptable appears to be due
to the lexical property of the indenite as obligatorily taking narrow scope with some discrete
term (see chapter 7), which in this interpretation is identied as the variable projected by the
preceding denite (in the primary LINK ed structure annotating an unxed node at the outset
of building that structure). To model such an interpretation, we would have to loosen Merge
to allow an unxed node to be updated in virtue of some annotation on a node which was
only a sub-part of the whole annotation, an extension which has no independent motivation.
19 The framework as set out so far provides no explanation of the sentential subject restric-
tion, which constraint is standardly taken as holding in all languages:
(i) *What is that Mary met unlikely?
The only grounds we would have for precluding (i) would be on the basis of the agreement
form on the auxiliary, requiring either that the pointer move to the subject node, or that it
move to some predicate node, neither of which restrictions is met by (ii), an analysis which
would require modication of the lexical specication of auxiliaries indicated in chapter 3. We
leave this possibility open.
Relatives: Towards a Dynamic Typology 121
4.3 Relatives: Towards a Dynamic Typology
As indicated at the outset of the chapter, the particular set of rules underpinning
English is only one of a number of ways in which a linked tree conguration can
be established. Stepping back from the specics of the rules as dened for
English, a head-initial SVO language, we might expect variations as to:
(i) how the sharing of a formula between the two trees is achieved,
(ii) what restrictions there are on the tree positions at which the shared for-
mula may occur,
(iii) the relative order in which the pair of linked tree structures is projected,
(iv) whether or not the LINK Introduction rule is encoded in actions of specic
lexical items.
4.3.1 Relativizers Projecting a Requirement
We have so far seen that the introduction of a LINK ed structure realizes the
LINK constraint for a common formula by the introduction and annotation of
an unxed node. But this is a considerable enrichment of the minimal char-
acterization denitive of the transition itself; and should a language have no
construction-specic device with which to ensure the presence of the required
second copy, then there is only one way to satisfy the imposed constraint,
and that is to use the general copy device available in all languages, namely
anaphora. In many languages there is no specialized copy device particular to
relative clauses. There is simply a regular pronominal copy at an appropriate
xed position in the tree.
Take for example the case of (Egyptian) Arabic { a language which uses
resumptive20 pronouns in all restrictive relative clauses (apart from the subject
position); and these clauses lack any relative complementizer if the nominal
head is interpreted as indenite:
(4.29) za?al mudarris Magdi darab-u
p [Egyptian Arabic]
angry teacher Magdi hit him
À teacher who Magdi hit was angry.'
The process of projecting the construal of relatives in Arabic is displayed graph-
ically in gure 4.7.
20 In all its forms, Arabic is a verb-initial subject-pro-drop language whose predicates are
invariably associated with contextual reconstruction of the subject position in the absence
of any expression indicating such information, exactly as though a subject expression had
been lexically projected. Recall from section 3.2.2 that we characterize this by dening the
condition for all verbs in Arabic as ?T y(t), and assigning verbs a sequence of actions which
includes projecting the subject node and annotating it with a pronominal meta-variable as
a trigger for a process of anaphoric resolution. The consequence of this analysis is that all
noun phrases preceding the verb have to be analysed as projecting either an unxed node or a
LINK ed structure (see Edwards et al. forthcoming for arguments in defence of this analysis).
The examples given in this section are from Egyptian Arabic, provided by Malcolm Edwards.
HOST TREE f?T y(t)g
fF o(Za?al); f?T y(e)g

p
T y (e ! t)g
fF o(P (; P )), f?T y(cn)g

T y (cn ! e)g
fF o(U), fF o(Mudarris);
T y (e)g T y (e ! cn)g
LINKED TREE f?T y(t);

?hDi(F o(U))g
fF o(Magdi), f?T y(e ! t)g
T y (e)g
fF o(Darab), fF o(U),
T y (e ! (e ! t))g T y (e); }g
Figure 4.7: The relativizer as projecting a requirement, satised by a subsequent
pronominal
The operation of building a LINK ed structure is a freely available option,

but this process merely constructs a LINK ed root node with a requirement for
some copy of the formula annotating the head in the LINK ed tree. There is
no relative pronoun ensuring this copy at some unxed node and no LINK-
Copy Introduction rule, unlike in English. Nevertheless, there is no need for any
specic rule in Arabic to full this requirement. The clitic pronoun is a regular
weak form of pronoun; and clitic pronouns can be interpreted indexically as
freely as strong forms. It is simply that at least one pronoun in a relative
sequence must have the formula annotating the head as an antecedent if the
requirement decorating the top node is to be fullled. In the absence of any
distinguished relative pronoun, some clitic pronoun has to be interpreted as
identical to the head.
In our analysis up to now, the wh relativizer has obviated the need for a re-
sumptive pronoun by annotating some unxed node. But, in Egyptian Arabic,
relative clauses with a head noun marked for deniteness, though obligatorily
associated with a complementizer illi,21 nevertheless require the presence of a
resumptive pronoun in all positions other than subject, as though no comple-
mentizer were present. In fact, we have to ensure that illi and non-denite
relatives are in complementary distribution { if the head is marked as denite
by the determiner, illi is required; if the head is not so marked, illi is precluded:
(4.30) *l-mudarris Magdi darab-u

the-teacher Magdi hit him
`the teacher Magdi hit'
(4.31) l-mudarris illi Magdi darab-u
the teacher that Magdi hit him
`the teacher Magdi hit'
(4.32) mudarris Magdi darab-u
teacher Magdi hit him
à teacher that Magdi hit'
(4.33) *mudarris illi Magdi darab-u
teacher that Magdi hit him
à teacher that Magdi hit'
To ensure this distribution, we dene illi as inducing a LINK Introduction spe-

cic to the denite article, dening the general LINK introduction, complemen-
tarily, as operating only in the absence of this trigger. We itemize the relative
clause rule for Egyptian Arabic, illi, the denite article prex l-, the nominal
mudarris (`teacher') as follows:22
LINK-Introduction (Egyptian Arabic)

fX; F o(); T y(e); Def ( ); }g
fX; F o(); T y(e); Def ( )g, fhL 1iX; ?T y(t); ?hDi(F o(); T y(e)); }g
XADR
21 Or yalli in other forms of Arabic { a form morphologically related to the demonstrative

pronoun.
22 We side-step all issues related to the anaphoric use of the denite determiner, simply
indicating the operation involved as P (the P ). See section 7.3.1.1 for a formal statement
in which the denite determiner is dened as projecting the meta-variable introduced by the
nominal onto the mother node of type e, this meta-variable then being available for some
(later) pragmatic process of Substitution.
mudarris
IF f?T y(e)g
THEN IF f#1 9xF o(x)g
THEN go(h#0 i); make(h#1 i); go(h#1 i);
put(F o(Mudarris); T y (e ! cn); [#]?);
go(h"1 i); make(h#0 i); go(h#0 i); put(F o(V); T y (e))
ELSE make(h#1 i); go(h#1 i);
put(F o(X (; x; X )); T y (cn ! e));
go(h"1 i); make(h#0 i); go(h#0 i); put(?T y (cn));
make(h#1 i); go(h#1 i); put(F o(Mudarris)),
T y(e ! cn); [#]?);
go(h"1 i); make(h#0 i); go(h#0 i); put(F o(V),
T y(e); Def ( ))
ELSE ABORT
illi
IF fF o(x); T y(e); Def (+); [#]? g,
THEN make(hLi); go(hLi); put(?T y(t); ?hDi(F o(x); T y(e)))
ELSE ABORT
l
IF f?T y(e)g
THEN make(h#1i); go(h#1i); put(F o(P (the(x); P (x))); T y(cn ! e));
go(h"1 i); make(h#0 i); go(h#0 i); put(?T y (cn));
make(h#0 i); go(h#0 i); put(?T y (e); Def (+)); gofirst" (?T y (e))
ELSE ABORT
Note the lack of an unxed node in the LINK Introduction rule for Arabic and
the use of the hDi modality, interpreted over the re exive transitive closure of
L [ . With ?hDiF o(U) on the root node of the linked tree, the LINK
Introduction displays the second respect in which languages such as Arabic
dier from Romance and Germanic languages: the copy of the head formula
may occur anywhere in the global tree, not just in the tree local to the top
node.
Notice, next, the characterization of mudarris (= `teacher'). Item-internal
disjunctions of this sort are used to express dierent eects which a lexical item
may project in dierent structural contexts.
Then, with the lexical specication of illi (and the determiner l ), the dif-
ference between English and Arabic resumptive use of pronouns is ensured.
Locality apart, the primary dierence between the two types of language is that
in English the relative pronouns are properly anaphoric in nature, securing a
copy of the head formula within the LINK ed structure. Illi, by contrast, does
not { it merely initiates a LINK introduction, imposing a requirement on the
LINK ed tree that it contain such a copy.
This cross-linguistic dierence in relativizers might seem surprising, particu-
larly given that illi is a depleted form of demonstrative. However, the intrinsic
weakness of illi { merely inducing the construction of a LINK ed structure in
which a copy of the formula annotating the head should occur { is not such a
surprising property if we recall that there are two dierent types of anaphoric
devices in language, the regular anaphoric form and more specialized expletive
forms. Expletive anaphoric expressions have a range of eects which mark them
out as distinct from their non-expletive counterpart. Their interpretation is pro-
vided not from what is already made available within the discourse or23structural
context, but from some expression which occurs only subsequently:
(4.34) Iti is likely [that John will come]i.
(4.35) Therei arose [a storm]i.
Expletive uses of pronouns are not an option freely available for all pronomi-
nal forms, but have to be lexically dened as a separate use of each appropri-
ate pronoun. They are, furthermore, characteristically associated with specic
structural contexts. For example, indenite expletives in English occur only in
subject position and only with a restricted subset of predicates; and it expletives
are obligatorily associated with sentential forms and occur only with a restricted
set of verbs. In our analysis, one essential feature shared by all expletives is that
they do not supply annotations; rather, they project requirements: and these
have the eect of ensuring the presence of some later sequence of words which
can be construed as satisfying them. For instance, in Romanian, clitic pronouns
function like agreement devices, projecting requirements with which some sub-
sequent expression must agree (the following examples are from Dobrovie-Sorin
1990):24
(4.36) len nu (i)-am examinat pe multi elevi
yesterday not (them)-past1:Sing examined pe many students
`Yesterday I did not examine many students.' [Hungarian]
(4.37) invita-(l) pe oricare elev
invite-(himACC ) pe any student
Ìnvite any of your students.'
(4.38) no (i)-am scris nimanui
not (himDAT )-past written nobodyDAT
Ì haven't written to anybody.'
Looked at from this perspective, illi becomes less puzzling. It is not so much a
relative subordinator which happens to project agreement properties that have
to be simply stipulated. Rather, it is an expletive relative pronoun that initiates
23 The co-indexing indicates correlation of interpretation between two types of expression,
not any properties of denotation.
24 Note that there is no presumption of deniteness in these Romanian examples, so these
clitic pronouns cannot be taken to be indexical and indicative of some independently identi-
able referent.
the introduction of a LINK transition, and leads to the decoration of the root
node of the LINK ed tree with the requisite formula requirement. This speci-
cation, in virtue of being a requirement, forces the later occurrence of some
expression to provide the appropriate value. In this respect it is like an agree-
ment marker, except that the requirement,25 though lexically projected, does not
have to be satised within the local tree. This property is common amongst
relative complementizers (see also the complementizer marking of Irish { Mc-
Closkey 1988, DuÆeld 1995), a phenomenon forming part of the general pattern
of relative complementizers as encoded anaphoric devices. Indeed, it is one of
the advantages of using a modal tree-description language that such statements
of locally projected requirements to be satised non-locally are straightforwardly
available.
From this account, one might expect that, just as individual languages dis-
play both non-expletive and expletive forms of the same pronominal, subject
to lexical variation, so an individual language might contain both an expletive
and a non-expletive variant of a relativizing particle. Thus Hebrew has both
a relative clause strategy with `gaps' and one with free use of resumptive pro-
nouns, and can be dened as having both the anaphoric-like form of relativizer
analogous to the empty relativizer in English and the expletive26 form of rela-
tivizer analogous to the relativizing complementizer in Arabic. It is notable
that the gap variant is subject to subjacency eects, whereas the resumptive
form is not:
(4.39) ha-?-is se xasavt
p se Dani pagas (?oto) p [Hebrew]
the man that thought2:F em that Dani met (him)
The prex se introducing the relative has no morphological re ex of anaphoric-
ity. To the contrary, it is simply a general subordinating device. Notice the
identical form of the prex se for both relativizing and subordinating particle
and the quite dierent form of the pronoun ?oto. p
There is a further source of variation in Hebrew. Given that the initial

transfer of information from the host node to the new LINK ed structure may
take the form of a required copy of the nominal variable, with no associated
locality restriction, there are two ways this requirement can be met. The rst
is in one step by a pronoun annotating a xed tree node with a copy of the
25 The correlation between expletives and agreement markers is implicit in their near-
equivalent characterizations. Agreement suÆxes project a requirement (or in some cases an
annotation) on a node from the perspective of some other node which is itself currently being
annotated (e.g. agreement on a verb indicating requirements on properties of the accompany-
ing subject, and agreement on a determiner indicating required properties of some associated
nominal). One dierence, on this view, between agreement markers and expletives, on the one
hand, and agreement markers and pronominals, on the other, is the relatively low-level dier-
ence that agreement markers are not morphologically independent forms. Another is that the
requirement projected by agreement marking is essentially local, and does not involve any #
relation.
26 The analysis of Hebrew se as ambiguous has been independently proposed by a number
of other people; see Shlonsky 1993, Suner 1998.
head formula. However, the construction process can also be separated into two
stages: xing the position of the tree node and supplying the Formula value to
that node. This separation will take place just in case the formula projected by
the pronoun annotates an unxed node, so that the annotation on some node
is established, while the tree-identier for that node remains to be xed:
(4.40) ha-?-is se ?ani xosev se ?alav
p p p ?amarta se Sara
p
the man that I think that about him you said that Sara
katva sir
wrote poem
(4.41) ha-?-is
p se ?alav
p ?ani xosev se
p ?amarta se
p Sara
the man that about him I think that you said that Sara
katva sir
wrote poem
(4.42) ha-?-is
p se ?ani xosev se
p ?amarta se
p ?alav
p Sara
the man that I think that you said that about him Sara
katva sir
wrote poem
`the man that I think you said that Sara wrote a poem about him'
As these examples show, a pronominal prepositional phrase can be `fronted' in
Hebrew to any position associated with a clausal boundary occurring between
the head and its point of construal in the emergent structure. This is a process
identied in the literature as a separate process called left-dislocation (see section
4.3.2 and Anagnostopoulou 1997).
Notice how such an update proceeds in steps. In the rst step, an unxed
node is introduced, and the initially imposed requirement is satised (through
*Adjunction and anaphora resolution).27 In the second step, the tree node
identication for that node gets xed (through Merge ). We have the progression
from
fX; ?T y(t); ?hDi(F o(); ?T y(e))g
to
ffX; ?T y(t)g : : : fh" iX; F o(); T y(e); ?9xT n(x)gg
as the rst step, with the construction and annotation of an unxed node, so
that only the identication of the tree node remains as a requirement to be
fullled; and the progression at the second step from
ffX; ?T y(t)g : : : fh" iX; F o(); T y(e); ?9xT n(x)gg
to
ffX; ?T y(t)g : : : fY; F o(); T y(e)gg
27 We assume here for simplicity that the form ?alav is of T y(e), the prepositional content
p
being merely an annotation on that node.

where the tree node is nally xed at Y . No special process needs to be dened
to characterize this sequence of steps { it is merely a combination of processes
that are available independently.
4.3.2 Variation in Locality
As we have seen, the Semitic family of languages displays variation as to whether
the relativizing element is or is not expletive-like. Independently, this element
also displays a lack of any locality restriction on the second occurrence of the
formula annotating the head. Since these two parameters are independent, we
would expect there to be languages in which a relative pronoun projects merely
a formula requirement on the root node of the LINK ed structure (as in Arabic)
but under the restriction that, as in English, the copy of the head formula
occur within the local tree. And indeed, between Romance languages we get
variation relative to a shared locality restriction { for example, Italian and the
closely related language of Romanian (see Dobrovie-Sorin 1990).28 In standard
Italian, resumptive pronouns do not occur in ordinary relative constructions
(see Cinque 1990). A gap is required: the various relative complementizers, as
in English, ensure the presence of the formula from the head in the LINK ed
tree.29 However, in Romanian, whose clitic pronouns have both expletive and
non-expletive functions,30 the relative particle, care, is obligatorily paired with a
subsequent resumptive pronominal, much as in Arabic. However, unlike Arabic,
this Romanian pronominal is subject to strong island restrictions, and may not
occur within a relative clause or a clausal adverbial (the data are from Dobrovie-
Sorin 1990):
(4.43) baiatul pe care l- am vazut [Romanian]
the boy pe which him have1:SING seen
`the boy that I saw'
(4.44) *baiatul pe care am vazut
the boy pe which have1:SING seen
28 There is widespread variation between the dierent dialects of Italian in this respect across
the dierent uses of relative and non-relative pronouns: see Escobar 1997.
29 There are two types of complementizer: the invariant che, the general subordinator, that
does not allow pied-piping (there is also a form cui which occurs with prepositions), and Det
+ quale which is a pronominal form occurring only with pied-piping forms (see Cinque 1995,
Escobar 1997). These two patterns correspond to the two LINK rules postulated, LINK
Introduction and LINK-COPY Introduction. Cinque (1995) reports that in `very colloquial'
Italian, the invariant che is used with free availability of either gap or resumptive pronoun:
see chapter 5 for discussion of how anaphora resolution and Merge may interact in such cases.
30 In our characterization of expletive, the clitic pronouns in (i) are putative expletives:
(i) L-am vazut pe Ion.
him-(we) have seen pe John
`We saw John.'
(4.45) *Omul pe care cunosc femeia care l-a
the manj pe whichj (I) know the womani whichi himj have
întîlnit
met ej
*'the man which I know the woman who met.'
If we dene care as an expletive relativizer projecting a requirement ? # F o()
on the root node of the LINK ed structure for an occurrence of the requisite
formula within the local tree, rather than the weaker form ?hDiF o(), we re-
ect the dierence in distribution of resumptive pronouns in Italian and Roma-
nian: only the latter requires (strong-island-respecting) resumptive pronouns
in relatives. In Spanish we have a mixed system much like Hebrew, in which
resumptive pronouns are in some positions optional with the neutral que form,
and in others (the oblique positions) obligatory.31
In sum, dierent forms of relative clauses vary according as a requirement of
a copy of the head formula in the LINK ed structure is or is not realized immedi-
ately as an annotation on some introduced unxed node. In the absence of any
such realization, the outstanding requirement will necessitate the occurrence of
a pronoun, for only this can secure the presence of the required copy. But in
all variants, the imposition of the restriction that leads to two copies of a single
31 Suner's (1998) analysis of resumptive pronouns in relative clauses is supercially close to
our own, with the distribution of resumptive pronouns due to whether the complementizer
does or does not have strong pronominal features. However, Suner explicitly sets aside re-
sumptive pronouns as being a mere rescue device, and does not include them within a general
characterization of anaphora. Her Minimalist analysis takes left-dislocation structures to be
an imperfection, motivated only by some morphological feature needing to be checked. This
pronominal feature is not an interpretable feature, and is deleted upon checking, playing no
part in construal at the LF interface. Specically, it is not this feature which ensures the
co-indexing with the head, but the predicational `force' of the complementizer. Relative to
these assumptions, resumptive pronominals are the preferred form of analysis, providing a
derivation in which nothing moves. The rst problem for this analysis is that she has to allow
a grammar-internal process of `lexical manipulation' which ensures that the relative pronoun
if not moved surfaces as a regular lexical pronoun. It is by no means obvious how this pro-
cess is to be dened: as she herself observes, the resumptive element may also be realized by
some anaphoric expression other than a pronoun, data which are problematic, given that the
replacement of the relative pronoun at PF by the pronominal resumptive is licensed in virtue
of adding nothing to the interpretation:
(i) The boys who I could see both were shaking with fear, ran away from the bomb.
(ii) There was discussion about infrastructure which I could see he wasn't going to take that
aspect into account.
Third, to re ect the observation that relative wh elements contrast with wh interrogative ele-
ments, Suner proposes that anaphoric operators are restricted so as not to appear as variables
at PF, but are subject to a grammar-internal process of manipulation to ensure their pronom-
inal form. But this stipulated restriction is nothing more than a statement of the problem.
It makes it all the more puzzling that the expressions replacing such relative pronoun forms
are all anaphoric. In the light of this, the assertion that anaphoric operators cannot remain
in situ, because à relative pronoun cannot behave as a bound variable at PF0 is mysterious:
it is referential expressions such as these which are taken not to require movement. More
generally, this account does not explain why wh forms should be quasi-anaphoric expressions,
why they should allow variation, or why they should interact with anaphora resolution to
yield the assigned interpretations.
term in the two structures can be locally dened as part of a LINK introduction
rule.
4.3.3 Topic Structures and Relatives
This analysis of relatives extends naturally to so-called clitic left-dislocation
structures (see Anagnostopoulou et al. 1997), enabling us to bring out the
parallelism displayed by these two types of structure in head-initial languages
in a straightforward way. In clitic left-dislocation structures, there is a pairing
between a left-dislocated constituent and a pronominal in the main structure,
with the dislocated expression characteristically separated from the remainder
of the sentence by a marked intonational break. Re ecting this, we analyse clitic
left-dislocation structures as a pair of linked structures: the initial noun phrase
sequence is taken to project a tree with a root node annotated by some formula
of type e, and that node has the root node of a second tree LINK ed to it,
with the familiar constraint imposed on the development of this new structure
that it contain some node annotated by F o() (see gure 4.8):
(4.46) Magdi, Ali abil-ha [Egyptian Arabic]
Magdi, Ali met-her
`Magdi, Ali met.'
(4.47) irragil, huwwa mabsu:t
the man, he happy
`The man, he is happy.'
(4.48) Gianni, lo vedro domani [Italian]
`Gianni, him I'll see tomorrow.'
(4.49) Ion, l-am întîlnit anul trecut [Romanian]
John, him-(I) met year last
`John, I met him last year.'
(4.50) As for John, I like him. [English]
fX; F o(Magdi); T y(e)g
fhL 1 iX; ?T y(t); ?hDi(F o(Magdi))g
Figure 4.8: The result of processing Magdi in (4.46)

By virtue of there being no expression analogous to relative pronouns in these
clitic left-dislocation structures, languages in which the relativizing element
merely imposes the requirement for a copy of the formula F o() should dis-
play exact parallelism between relatives and clitic left-dislocation structures {
the pronoun has to be construed resumptively in both types of structure. For
instance, in Egyptian Arabic, both topic and relative structures require a re-
sumptive pronoun, except in subject position, and in subject position, where a
pronoun is generally precluded, a strong form of pronoun is acceptable if that
constituent is focused:
(4.51) l-bint illi Ali ?abilha.
p [Egyptian Arabic]
the girl that Ali met her
`the girl who Ali met'
(4.52) irragil illi mabsu:t
the man who happy
`the man who is happy'
(4.53) ??irra:gil illi huwwa mabsu:t
the man who he happy
`the man who is happy'
(4.54) irra:gil illi hatta huwwa mabsu:t
the man who (even) he happy
`the man who even he is happy'
(4.55) Nadja, Ali ?abil-ha
p
Nadia Ali met her

Nadia, Ali met.
(4.56) Ali, mabsu:t
Ali, happy
Àli, he is happy.'
(4.57) ??Ali, huwwa mabsu:t
Ali, he happy
Àli, he is happy.'
(4.58) Ali, hatta huwwa mabsu:t
Ali, even he happy
Àli, even he is happy.'
Hebrew, with its two alternative construals of subordinate relativizing se,
also has left-dislocation structures, parallelling relative clause sequences, with
the dislocated constituent construed as antecedent to a pronoun annotating an
unxed position in the LINK ed structure (see Demirdache 1991):
(4.59) Shalom, ?ani xosev se ?alav
p p p ?amarta se Sara katva
Shalom, I think that about him you said that Sara wrote
sir [Hebrew]
poem
`Shalom, I think that you said that Sara wrote a poem about him.'
(4.60) Shalom, ?alav
p ?ani xosev se
p ?amarta se
p Sara katva
Shalom, about him I think that you said that Sara wrote
sir
poem
(4.61) Shalom, ?ani xosev se
p ?amarta se
p ?alav
p Sara katva
Shalom, I think that you said that about him Sara wrote
sir
poem
On the other hand, in languages in which the relativizing element is a relative
pronoun that presents the copy of the head formula as annotation to some un-
xed node in the LINK ed structure, there should be no such parallelism, and a
resumptive pronoun should be required only in clitic left-dislocation structures.
And so it is that Romanian and Italian dier, with Romanian patterning, like
Arabic, with perfect symmetry between topic and relative structures, but Italian
displaying asymmetry in the two forms of structure. Italian, despite precluding
resumptive pronouns in relative structures, requires a resumptive pronoun in
clitic left-dislocation structures, where it is characteristically associated with
so-called comma intonation, as in Romanian (see Cinque 1990, Dobrovie-Sorin
1990):
(4.62) *l'uomo che l'aprezzavamo [Italian]
the man that we appreciated him
`the man that we appreciated'
(4.63) Gianni, l'apprezzavamo.
Gianni we appreciated him
`Gianni, we appreciated.'
(4.64) baiatul pe care l-am vazut [Romanian]
the boy pe which him-have1:SING seen
(4.66) Ion l-am întîlnit anul trecut.
John him-have1:SING met year last
John, him I met last year.'
Notice, incidentally, that clitic dislocation involves the construal of the clitic
pronominal from a dislocated hence unxed position. In clitic left-dislocation
structures then, in Italian and Romanian alike, the process of interpretation
echoes that associated with the previous Hebrew left-dislocation data { with
a LINK transition followed by the projection of information from a pronoun
onto an unxed node being used to full the modal requirement imposed by the
LINK transition.32 What this display of cross-language variation shows { both
in relative clauses and in left-dislocation structures { is the immediate benet
of the analysis in terms of the building of linked tree structures. For in each
of the analyses { in English, in Arabic, in Italian, and in Romanian { the only
variation that has been manipulated is whether the relativizing element has
pronominal-like properties or not and the dierence in the locality requirement
associated with the node that is to be eventually annotated by the the copy of
the head formula.
4.3.4 Variation in Order { Head-nal Relatives

In the account so far we have concentrated almost exclusively on head-initial
constructions, and head-initial relatives in particular. Such structures have
been the point of departure for the development of the framework, but we shall
see that, nevertheless, the emphasis on underspecication of tree description
allows us to describe languages such as Japanese surprisingly straightforwardly,
providing both a basis for showing what languages such as these have in common
with head-initial ones and how they dier.
As we discussed in chapter 3, the problem posed by Japanese is the multi-
ple ambiguity presented by strings. In a sequence of noun phrases, there is no
certain identication of which belong to the main clause, to some subordinate
clause, or a relative clause. Thus the scanning of sentence-initial Bill ga Mary
o could have both Bill and Mary annotating nodes within the same local tree
structure, or Mary could be in a LINK ed structure, or both could be in a
32 This gives us a basis for explaining why, in Italian, resumptive clitic pronouns in relative
clauses are said to give rise to ungrammaticality, whereas resumptive construal of pronouns
in relative clauses in English, though of reduced acceptability, are not ill-formed:
(i) ?*l'uomo che l'aprezzavamo
`the man that we appreciated him'
(ii) ?the woman that we ill-treated her so badly she reported us to the police
In English, if the pronoun is taken to provide input to Merge, the eect will be simulta-
neously to provide a formula for the node being decorated by scanning the pronoun and an
identication of the hitherto unxed tree node. As a result, the tree description update will
be non-vacuous. In Italian, however, where the clitic pronoun occurs at an unxed node, the
result of Merge would be a vacuous update, since the unxed node would remain unxed. If
this line of reasoning is correct, we would anticipate that the strong form of pronoun occurring
in situ in Italian would, as in English, be of reduced acceptability, used for specic pragmatic
eects such as focus. We take up the interaction of Merge and anaphora resolution in detail
in chapter 5.
LINK ed structure. And so on. There is also the possibility of either one be-
longing to a subordinate clause. It may not be possible to resolve these choices
until the appropriate cluster of verb+nominal or verb+verb sequence becomes
available.
We have assumed that interpretation in Japanese is only very weakly goal-
driven, with case specications narrowing down the space of possible tree posi-
tions for the node they decorate, the driving force of interpretation being indica-
tors of structure such as verbs (which occur nally in any clausal sequence) and
morphological indicators of boundaries between one structural domain and an-
other (such as indicators of relative or other clausal boundaries). Let's consider
the following example containing a restrictive relative clause.
(4.67) Hiroto ga muita ringo o tabeta. [Japanese]

HirotoNOM peeled appleACC ate
`Hiroto ate an apple he peeled.'
or `Sue.. ate an apple Hiroto peeled.'
or `Hiroto ate an apple Sue.. peeled.'
We give a step-by-step construction of (4.67) under the interpretation `Hiroto

ate an apple he peeled'. As in the derivations of Japanese examples in chapter
3, the rst action is that the noun phrase Hiroto creates an unxed node of type
e annotated with F o(Hiroto); T y(e), and address hU iT n(0). The suÆx ga adds
the case requirement ?h"0 iT y(t). Then, having the pointer at the root node,
we invoke from the lexical specication of muita a sequence of actions, which
leads to the construction of a tree whose nodes of type e are annotated with
meta-variables that require identication:
muita
IF f?T y(t)g
THEN make(hDi); go(hDi);
make(h#0 i); go(h#0 i); put(F o(U); T y (e); ?9xT n(x)); go(h"0 i);
make(h#0 i); go(h#0 i); put(F o(V); T y (e); ?9xT n(x)); go("0 );
make(h#1 i); go(h#1 i); put(F o(Muita); T y (e ! (e ! t)); [#]?);
gofirst" (?T y (t))
ELSE ABORT
Having parsed this verb, we have a partial tree of the form:
f?T y(t)g
1 f?T y(t)g
fF o(U),
T y (e); f?T y(e ! t); }g
?9xF o(x)g
fF o(V), fF o(Muita),
T y (e);
?9yF o(y)g T y(e ! (e ! t))g
1 = fF o(Hiroto); T y(e); ?9xT n(x)g
Here we will use the variable projected as the object argument of Muita in
preparation of the intended interpretation, though this is a matter of choice,
a pragmatic action like other anaphoric choices. The problem here is that the
value of the term to be shared between the two nodes is not yet known at the
point of the transition; the variables present have to be identied, through the
rule of Japanese LINK Introduction. We display this as:
f?T y(t)g f?F o(V); }g
1 fT y(t)g
fF o(U),T y(e); f?T y(e ! t)g

?9xF o(x)g
fF o(V), fF o(Muita),
T y (e);
?9yF o(y)g T y(e ! (e ! t))g
1 = fF o(Hiroto); T y(e); ?9xT n(x)g:
ffX; T y(t); : : : }g : : : f(MOD)X; F o(U); T y(e); ?9xF o(x) : : : gg

ffX; T y(t); : : : g : : : f(MOD)X; F o(U); T y(e); : : : gg;
fhLiX; F o(U); ?9xF o(x); T y(e); }g
where X is an address and (MOD) 2 fh"0 1i; h"1 1ig
So, from the pair of a top node of some tree with a Formula value of type t and
some arbitrarily selected local node still requiring a xed value (i.e. annotated
with a meta-variable), the rule licenses the introduction of a LINK ed node
whose Formula value must, by denition, be that of the selected one. As a
result, in the above tree, the value of the object argument of muita and the
head of the LINK ed structure must be the same. The rule having been applied,
the head formula must then be supplied by some following expression of type
e provided by the immediately following natural language string; so, unlike the
parsing of successive noun phrases in the Japanese string, at this point in the
interpretation process, there is a procedural location where the structural role
of the immediately subsequent part of the string is strictly determined. This
rule thus provides the basis for the interpretation of the head-nal structure,
with ringo (= àpple') immediately following the verb muita (= `peel').
The next step in the sequence of actions is the processing of ringo, with its
lexical entry:
ringo
IF f?T y(t)g
THEN make(hDi); go(hDi); put(?T y(e))
DO()
ELSE IF f?T y(e)g
THEN DO()
ELSE ABORT
where DO() = put(F o(U); T y(e)); make(h"0 ih#1i);
go(h"0 #1 i); put(F o(Ringo); T y (e ! cn); [#]?);
go(h"1 i); put(?T y (cn)); make(h"0 i); go(h"0 i); put(?T y (e));
make(h#1 i); go(h#1 i); put(F o(P (; P )); T y (cn ! e));
go(h"1 i)
The sequence of actions dened for ringo has rst the general condition of
requiring T y(t) (which provides the cases we have already seen, where ringo
may project and annotate an unxed constituent of type e); but it should also
be prepared for the structurally specic location immediately following a verb.
In the latter case the noun must project the head of the previous relative clause.
That is, following a verb, the structural role of a noun is as fully determined
as it is in a language like English. If the condition of this word is satised, it
creates a tree structure annotating the current node (requiring type e) with the
nominal variable, but embeds that in a larger structure. The resulting structure
for the interpretation of the object noun phrase in (4.67) (with the additional
restrictor predicate added from the LINK ed structure) has the form displayed
in gure 4.9.
The lexical actions projected by the common noun ringo immediately after
the construction of the LINK transition identify the nominal variable with the
type e required at the type e node resulting from the transition: this creates an
interpretation as a restrictive relative. To the top node of the noun projection
the accusative case marker o then adds the requirement ?h"0iT y(e ! t), after
fF o(; (x; Ringo(x));
T y (e)g
fF oP (; P ); fF o(U,Ringo(U)^

T y (cn ! e)g Muita(U)(Hiroto)),T y (cn)g
fF o(U),T y(e)g fTFy(oe(X (X; Ringo(X )));

! cn)g
Figure 4.9: Incorporating the formula projected from the relative
which the pointer moves back to the root node.33

Finally, in (4.67), the verb tabeta determines a sequence of actions (ignoring
tense):
tabeta
IF f?T y(t)g
THEN make(hDi); put(?9xT n(x); ?T y(t));
make(h#0 i); go(h#0 i); put(F o(X); T y (e); ?9xT n(x)); go(h"0 i);
make(h#1 ); go(h#1 i); put(?T y (e ! t));
make(h#0 i); go(h#0 i); put(F o(Y); T y (e); ?9xT n(x)); go(h"0 i);
make(h#1 i); go(h#1 i); put(F o(T abeta); T y (e ! (e ! t)); [#]?);
gofirst" (?T y (t))
ELSE ABORT
These actions lead to a tree for (4.67) where, again through merging nodes
and anaphoric resolution, all outstanding requirements for F o values and xed
tree positions, are fullled. Figure 4.10 displays, in abbreviated form, the -
nal annotated tree, showing merely the formulae. The structure is exactly the
same structure as is assigned to English relatives. Notice how on this account
the unxed node annotated by the formula F o(Hiroto) entered into the tree de-
scription at the very rst step can freely merge with the embedded subject node,
leaving the meta-variable annotating the matrix subject (the subject of tabeta
{ àte') to be identied by Substitution as F o(Hiroto), leading to the interpre-
tation `Hiroto ate an apple he peeled', even though it is the formula annotating
the embedded subject which is the antecedent of the formula annotating the
matrix subject node.
33 Notice that this case marker requires a completed node (no outstanding requirements), so
the common noun has to be evaluated. This involves incorporation of the information of the
linked structure into the common noun restrictor. The details of these procedural aspects of
parsing in Japanese we leave on one side.
F o(T abeta(; x; Ringo(x)^

Muita(x)(Hiroto))(Hiroto))
F o(Hiroto) F o(T abeta(; x; Ringo(x)^

muita(x)(Hiroto)))
F o(; x; Ringo(x)^ F o(T abeta)

Muita(x)(Hiroto))
F o(P (; P )) Fo(U,Ringo(U))
F o(U) F o(Ringo)
F o(Muita( U)(Hiroto))
F o(Hiroto) Fo(Muita(U))
F o(U) F o(Muita)
Figure 4.10: The head-nal relative { the resulting tree
From the dynamic perspective, the interest of head-nal relatives is two-
fold. First, the tree growth process has to allow upward development from
some node in linking one tree with another, so development is not exclusively
top-down. The second distinguishing property in the sequential projection of
these structures is the need to allow a process of node construction which allows
a node to be introduced which bears the weak tree relation D to the root. This
is not an option that we have allowed hitherto; and it re ects a genuine cross-
language dierence as to the degree to which the interpretive process is goal-
directed and driven by top-down expectations of structure, or, to the contrary,
tolerant of structure being built only as triggered by lexical expressions. The use
of the hDi operator to project sequences of nodes otherwise unrelated re ects
directly the way in which a language such as Japanese presents a at sequence
of NPs while nevertheless projecting a structure for interpretation with familiar
congurational properties. Here we merely note that one possible transition
amongst the disjunction of possibilities at the opening steps within the construal
of a clausal sequence is to allow the building of nodes which are only weakly
related to the node that constitutes the point of departure. It should be noted
that this is a possibility we make available here only from the root, and not
more generally.
Apart from these two distinguishing features, the process of inducing struc-
ture remains locally top-down, the verbs invariably building propositional struc-
ture from a trigger of type t with a set of actions that annotates all the nodes
within that structure. It is then34 subsequent operations of Merge or Substitution
which yield the desired result.
4.3.5 Head-internal Relatives

Even without rounding out all the details of the requisite LINK transitions
in these dierent non-lexical and lexical forms, we can observe that the present
account of head-nal relatives is straightforwardly compatible with an additional
form of relative known to exist in these languages, but not well understood. The
Japanese rule of LINK Introduction was dened so as to allow the introduction
of a head node connected to the previous structure by a LINK relation and
forcing identity between a term of the LINK ed structure and the formula of
type e at the head. Up to now, the formula thus copied has been a meta-
variable. But this is in no way essential. Suppose, to the contrary, that the
structure to35which the head is to be linked contains a full noun phrase structure
of the form:
34 Chinese might seem to pose an obvious problem for this analysis of head-nal relatives,
because the analysis has been constructed upon the assumption that the verb projects a whole
tree of type t with pronominal place-holding devices in the requisite nodes requiring type e.
In other words the sequence of steps is dened for pro-drop languages. Chinese, however, is
not an unrestricted pro-drop language (see Huang 1982). This suggests that in Chinese the
additional nodes that may be projected along with the node for the predicate formula itself
are not assigned meta-variables as in Japanese, but rather have requirements for formulae
of type e as in English, with specic license for the exceptional structures (e.g. question{
answer pairs) in which the subject is reconstructed in context. However, a notable property
of relative clauses in Chinese, distinguishing it from Japanese, is the obligatory presence of
the complementizer de. This complementizer de is a relic of a demonstrative form (Simpson
1999 analyses de as a determiner). Re ecting the spirit of Simpson's analysis, we dene de
as projecting a sequence of actions, rst to project a meta-variable as annotation to one such
node of type e (this is its anaphoric-like property), secondly to induce the LINK transition
onto a head node, imposing a requirement for that same meta-variable on that head node. The
analysis of head-nal relatives in Chinese is thus like Japanese except that the meta-variable
which is to be the common element is projected by de and not by the verb.
35 ronbun = `paper'.
fF o(;U, Ronbun(U)),
T y (e)g
fF o(P (; P )); fF o(U,Ronbun(U)),

T y (cn ! e)g T y (cn)g
fF o(U), T y(e)g fTFy(oe(X (X; Ronbun(X )));

! cn)g
Suppose furthermore that it is the formula annotating the top node of this
sub-tree that is to be identied with the head. Then it is (; (U; Ronbun(U)))
which is copied to the head node; and this will give rise to an interpretation
of the structure as a non-restrictive relative. This we encounter in the so-
called internally-headed relative clause structures known to exist in this type of
language:36
(4.68) John ga ronbun o kaita no ga LI ni notta
JohnNOM paperACC wroteNom NOM LILOC appeared
À paper which John wrote appeared in LI.'
(4.69) John wa Mary ga sanko no ringo o muita no o
JohnT OP MaryNOM threeGEN appleACC peeledNom ACC
tabeta
ate
`John ate all the three apples that Mary peeled.'
These data are highly problematic for analyses in which the wh element is
treated as a complementizer (or an operator in the CP specier position) binding
some gap position in the relative, for in these structures it is the full form of
determiner-nominal which lls what would otherwise be the gap. Kuroda (1999),
for example, argues that some theta requirement must be allowed, exceptionally,
to be saturated by an argument position from within a relative structure in total
disregard of the lack of locality.37
36 Recall that the so-called nominalizer with which these structures are formed is the genitive
marker. Examples from the literature include data from Quechua (see Cole 1987) and Japanese
(see Watanabe 1992), but Korean also displays the same type of structure. These structures
are sometimes called correlatives (see Keenan 1985). `Nom' describes the nominalizer, `NOM'
the nominative case marker.
37 To side-step the diÆculties presented by these structures, it is sometimes argued that these
structures are not relatives at all, but clausal adjuncts (see Murasugi 1996). The existence
of clausal adverbials marked by the particles ga, o and ni as a separate use of these particles
is taken to be uncontroversial. However, the arguments for the necessity of the internally
headed relatives are none-the-less persuasive (see Hoshi 1995, Kuroda 1998), and this is the
analysis we present here. It is notable though that the analysis of relatives as a LINK ed
structure is able to capture the extent to which these clauses are adverbial in so far as the
resulting structure is a pair of independent tree structures. Unlike other analyses of relative
clauses, there need be no radical dierence between the two forms of analysis, for the resulting
structures dier solely with respect to whether a shared term in the two structures is essential.
See Otsuka 1999 for evidence that the dierence between the two types of structure is solely
the result of pragmatic choice.
According to the analysis we suggest (following the general outline of the
Cole 1987 and Watanabe 1992 analyses), unlike the transition dened by the
general Japanese LINK Introduction rule, which is not driven by any relativizing
particle, in these structures, the construction of the relation is induced by the
particle no. The particle constructs a head node LINK ed to the top node of
the tree projected by the verb: in the resulting LINK ed structure, it is this top
node which has no outstanding requirements:
no IF fX; T y(t); ?;; # (F o(x) ^ T y(e) ^ 9xF o(x)) : : : g
THEN make(hL 1 i); go(hL 1 i); put(F o(x); T y(e))
ELSE ABORT
The immediately following case suÆx indicates the required relative position of
the head node within the tree, while returning the pointer to the root. The
construction of the Formula value at the head does not need stipulation that is
exceptional to these structures: the guarantee of the common occurrence of the
term in the two structures is ensured by the lexical actions projected by no.38
This analysis matches the observation of Hoshi (1995) that (4.69) and (4.70)
are not synonymous, with (4.69) but not (4.70) implying that `Mary peeled only
three apples', in other words an E -type form of interpretation:
(4.70) John wa Mary ga muita ringo o sanko tabeta
JohnT OP MaryNOM peeled appleACC three ate
`John ate three apples that Mary peeled.'
According to the analysis of quantication to be given in chapter 7, we would
anticipate this contrast, since the quantication projected from sanko no ringo o
in (4.69) must be compiled as part of the interpretation of the LINK ed structure,
with the skeletal epsilon term shared with the primary structure.39 In (4.70), to
the contrary, it is only a nominal variable that constitutes the shared element,
thereby giving rise to the restrictive interpretation, which is compatible with
Mary having peeled more than three apples.
There is a great deal more that needs to be said about the relationship
between head-nal and internally headed relatives and the way in which ap-
parently empty positions are interpreted in such structures. Nevertheless, the
38 It should be noted that there appear to be pragmatic restrictions on what can and what
cannot be expressed in the form of a head-internal relative, but we leave these on one side,
for further investigation.
39 Notice how examples such as (4.69) provide further demonstration that quantied expres-
sions may give rise to non-restrictive interpretations. See n.12.
anticipated existence within this framework of head-internal relatives provides

welcome initial conrmation of the dynamic perspective adopted.40
4.3.6 The Potential for Lexical Variation
The nal source of cross-language relative clause variation yet to be itemized is
one we have already invoked { namely, whether or not the projection of LINK ed
structures is explicitly encoded in lexical items. The framework allows variation
across languages as to whether a LINK Introduction rule is dened as a freely
available option without lexical triggering, or whether it is initiated by a lexical
item. As we've already seen, we nd a great deal of variation. In English, with
its null complementizers, the availability of a LINK transition has to be dened
as a free option not requiring a lexical trigger. Egyptian Arabic also has both a
LINK Introduction rule (for indenite relatives) and a transition constructed by
a sequence of actions dened as projected from the presence of illi. In Japanese,
there is no relativizing particle in head-nal relatives, though the41 nominalizing
particle no has to be dened to induce a range of LINK relations. In Mandarin
Chinese, relative constructions always require the particle de. And so on.
Despite the diversity, there is one property shared by all morphological en-
codings of the introduction of these linked structures. It is essential that any
such morphological expression be able to project dierent values according as
the information transferred varies, for it is these expressions which provide the
channel through which the information is transferred. And, overwhelmingly,
the most common pattern of relativizing particles across languages (unless, as
in Hebrew, the complementizer used is a simple subordinating device) is the use
of some element which serves elsewhere in the language as a regular anaphoric
40 There are no such head-internal relatives in Chinese, which at rst glance is surprising.
But, in Chinese, the same form of interpretation is achieved by the head-nal structures. In
Chinese, the sequence of words which immediately follows the nal relativizing element de in
head-nal structures is the full noun phrase sequence of quantier-classier-nominal:
(i) ai ta muqin de mei ge ren
loves his mother rel every classifier man
èvery man who loves his mother'
With de projecting both a meta-variable for one argument of the verb and the LINK transition
from some dominating node of type t, it is what follows de that will determine the term
with which the meta-variable will have to be identied; and in this type of structure it is
the term projected from the quantier-classier-nominal sequence which projects the shared
term, not just a nominal meta-variable. As we would expect, this gives rise to non-restrictive
interpretations, again notwithstanding the quantifying determiner.
41 No is also used as a genitive marker, which, as we are about to argue, also induces a LINK
relation.
expression (see Keenan 1985 for a broad typological survey), a fact which ts
our analysis very well.42
It might seem that the framework is over-liberal in allowing so much vari-
ation, with lexically induced and rule-based LINK transitions, both cross-
linguistically and even within a single language. It should be pointed out, how-
ever, that each parameter for variation is an independently motivated property
of the left-to-right form of tree growth { there is the distinction between an
annotation on a node and a requirement; the distinction between whether a
requirement is to be locally or globally satised; and interaction between com-
putational, lexical and pragmatic actions. The type of tree decoration and the
incremental form of goal-directed tree development are also common to all ex-
planations. Nevertheless, within this theoretical space, variation is indeed freely
available both cross-linguistically and within a single language. This provides a
42 Kayne's (1994) analysis shares with the present one a concern to respect considerations
of linear order yet give a unied analysis of relative clauses. According to Kayne's analysis,
non-restrictive, restrictive, head-initial, head-nal and head-internal relatives are all assigned
a base structure in which the Determiner has as complement a CP containing the full wh
expression together with the common noun plus any accompanying adjectives as a DP in the
clause-internal position. This full wh DP is moved overtly to the Spec CP position, and the
nominal associated with the wh determiner is moved over the wh within this internal DP to
Spec DP position to yield:
[D0 the [CP [DP bookj [D0 which tj ]i [C0 [IP I read ti ]]]]]
Non-restrictive relatives involve further movement, at LF, to yield a conguration in which
the IP is moved further up to the Spec DP conguration { providing a basis for its construal
independent of the DP of the common noun to which it is adjoined. Head-nal relatives have a
similar underlying structure, but require overt movement of the entire IP to Spec DP position
(or possibly to some node dominated by the DP node, but dominating the NP node within
which the relative sequence originated). Head-internal relatives are then alike, except that,
with movement construed as copy and delete, it is the copy of the NP which is deleted, leaving
the full NP expression within the relative sequence, contained at the highest level within the
Spec DP position.
Aside from the richness of movement processes, the analysis is not problem-free. It requires
that the nominal remains part of the CP structure, the entire nominal-plus-relative sequence
taken as complement to the determiner. This suggests, wrongly, that noun-plus-relative se-
quences should be pronominalized, like CPs, by the pronoun it :
(i) That book which I left on the table is not as nice as this *it/one.
Pied-piping cases are problematic (with no copy of the noun relative to which the adjoined
relative is an adjunct), particularly non-restrictive ones, where the adjoined wh expression
may contain a non-identical nominal:
(ii) The book whose author I disapprove of has just been published.
(iii) That book, which horrendous piece of writing should never have been accepted for
publication, is making enormous amounts of money.
Furthermore, with only a single occurrence of the nominal, within the CP structure, there
is no basis for characterizing how a noun which is head of some relative can be understood
independently of the relative adjoined to it:
(iv) The hostages who were now arriving had been very badly treated.
(v) The hostages who had been very badly treated were now arriving.
In order to project such an interpretation, one might posit movement at LF, but this would
only characterize non-restrictive forms of construal (not (iv){(v)). There is, furthermore, no
basis for explaining the distinction between the interpretation of head-nal and head-internal
relatives (see Hoshi 1995), or more generally, the correlation between anaphora and relative
clause construal.
new perspective on language variation { that any variation licensed within the
framework may occur in any one language. This turns out to be a bonus, as we
will now see.
4.4 Genitive Constructions as LINK Structures
Nothing precludes the construction of a pair of linked trees with the LINK ed
tree constructed prior to the projection of the head. Such a dierent direc-
tionality might seem a prototypical basis for typological dierences between
languages according as the language is head-initial (e.g. English) or head-nal
(e.g. Japanese). But, as we shall now see, head-nal structures are available in
English as lexically triggered.
In English genitives, a modifying structure of arbitrary complexity precedes
the structure modied, and there is, in consequence, an essential process of
upward tree growth:
(4.71) John's mother's friend died.
Informally, the initial task in parsing such a complex noun phrase sequence,
is the parsing of a full noun phrase, John, as though subject. As the string
continues, it becomes apparent that the formula projected from John is associ-
ated with an increasing level of embedding in the subject node. The sequence
of actions involved in parsing the subject noun phrase is one of constructing a
sequence of nested structures { in other words, remarkably redolent of Japanese
pre-modifying structures, albeit in a language in which the subject and predicate
nodes are constructed by the rules Introduction and Prediction.
In this kind of construction, the Adjunction rule becomes essential, for the
node created to accommodate the subject information is a daughter of the top
node (requiring type t). The genitive construction must apparently now ìnter-
polate' nodes between the top and the putative subject node, and this is an
operation which would not preserve tree structure. So for such structures, we
use Adjunction to create an intermediate node:
fX; ?T y(e); }g
fX; : : : ; ?T y(e)g; fhU iX; ?T y(e); }g
Adjunction, as dened by this rule, allows the creation of an alternative separate
node, which can then be embedded in a larger structure. From this newly
introduced node, the construction of a nested sequence of type e nodes projected
by a genitive NP such as John's friend is then induced by the suÆx 's, which
separates the possessor (John) from the possessee (friend) in the complex NP.
The actions projected by this suÆx are rst to embed the possessor in a tree
with top node of type t and subsequently link the root node of this tree to a head
node in which a term for the possessee is embedded. The eect of the possessive
suÆx 's in the interpretation of John's mother can be seen as falling into two
parts. The rst part, starting from the projection of John, nests that node
Genitive Constructions as LINK Structures 145
within a containing tree involving a binary P OSS relation with the projection
of John and a possessee variable as the two arguments:43
fX; F o(John); T y(e); }g
fF o(P OSS (U)(John)); T y(t); }g
fX; F o(John); T y(e)g fF o(P OSS (U)(John)); T y(t); }g
fF o(Y X (P OSS (Y )(X )))g fF o(U)g

The second part of the action sequence is to introduce a new tree with top node
requiring type e. This tree is structured as an existentially quantied noun
phrase in which the common noun's restrictor formula is still required:
f?T y(e)g
fT y(cn ! e); f?T y(cn); ?h#1 iT y(e ! cn)g

F o(P (; P ))g
fT y(e); F o(V)g f?T y(e ! cn); }g

The node annotated by the variable V of type e is now a head node to a LINK ed
structure and attached to the possessor tree by a LINK relation. The possessee
variable in that LINK ed structure is then identied with the variable projected
by the subsequent common noun (i.e. mother in John's mother ). Thus, the
possessive suÆx projects the following complex lexical action: 9
IF fF o(); T y(e)g >
>
THEN make(h"0 i); go(h"0 i); put(F o(P OSS (U)()); T y(t)); > >
>
>
make(h#1 ih#0 i); go(h#1 #0 i); put(F o(U); T y (e));
=
make(h"0 ih#1 i); go(h"0 #1 i); >

Possessor Tree
>
put(F o(P OSS ); T y (e ! (e ! t)));
>
>
>
>
go(h"1 "1 i);
;
9
make(hL i); go(hL i); put(F o(U); T y (e));
1 1 >
>
make(h"0 i); go(h"0 i); put(?T y (cn); ?h#1 iT y (cn ! e)); >
>
>
>
make(h"0 i); go(h"0 i); put(?T y (e));
=
make(h#1 i); go(h#1 i); put(P [; P ]; T y (cn ! e)); >

Possessee Tree
>
go(h"1 "0 i)
>
>
>
>
ELSE ABORT ;
43 P OSS is a relation whose value is constructed from some contextually provided value.
We leave it here in the form of P OSS for simplicity.
A display of a fully set-out structure for the interpretation of John's mother's

friend as subject is set out in gure 4.11, showing the recursive way in which
one node of type e can be nested inside another.44
Notice that this process of upward growth projected by a genitive noun
phrase must result in a structure with a top node annotated by a type identi-
cal to that decorating the node which initiated the Adjunction rule (i.e. type
e). This restriction does not need specic stipulation, for only if it is satised
can the structure be `clicked together' with the initiating node (by Merge ), and
a tree free from outstanding requirements be constructed.45 The tree growth
associated with the genitive suÆx is essentially that of head-nal relatives, ex-
cept that the structure in question is induced from a single suÆx { a LINK ed
structure is constructed prior to the construction of the associated head.46 The
process of interpretation for a complex genitive construction such as (4.71) is
then one of building successive LINK ed structures, with each one LINK ed to a
node of T y(e) within the structure that is projected immediately subsequently.
A number of welcome consequences follow from this analysis of genitives.
First we anticipate that the internal structure of a genitive determiner (in En-
glish internal to the possessive construction) will always project a structural
domain independent from the structure within which it is nested. There is thus
no locality violation between John and his in the well-known anaphora data,
for the pronoun internal to his is a term in a LINK ed structure and not in the
same local tree as the antecedent John:
(4.72) John admires his mother.
Moreover, we have dened fronted constituents as projecting unxed nodes,
related to the root node by the relation , and this relation does not include
L steps (recall that is the re exive and transitive closure of the relation # ,
which itself is dened as 0 [ 1). Consequently, èxtraction' out of genitive
constructions is debarred, as witness the ungrammaticality of (4.73):47
44 This analysis re ects the extent to which John's is simultaneously an adjunct and a
determiner. In so far as 's projects a tree LINK ed to the nominal variable for some nominal,
it is an adjunct within the containing nominal, but in that it projects a term-operator to bind
the nominal variable, it is a quantifying determiner.
45 Applications of the Adjunction rule yield precisely the auxiliary trees of Tree Adjoining
Grammar, re ecting recursive embedding of an expression of a given type within an expression
of the same type.
46 This correlation between genitive and relative constructions has been proposed by a num-
ber of people { see in particular Baker (1996), who suggests for Mohawk and related languages
that the genitive is an abstract form of relative.
47 Speakers divide sharply between those for whom (4.73) is completely ill-formed, and those
for whom it is borderline acceptable. These latter speakers appear to be redening such left-
peripheral genitives as sister to the whole structure, as in languages where this is an option
encoded in an alternative case realization (Andrew Simpson reports to us that in Hungarian,
the case assigned may optionally be dative, and with such case, èxtraction' is freely available
(see Szabolcsi 1983)). That this is a re-analysis phenomenon in English in the face of left-
peripheral pressure, and not a more general phenomenon, is suggested by the fact that this
form of interpretation is apparently completely unavailable from an object position, even for
those speakers for whom (4.73) is acceptable:
(i) The person who Sue admired's brother is Catholic.
Genitive Constructions as LINK Structures 147
f?T y(t)g
f?T y(e)g f?T y(e ! t); }g
fT y(cn ! e); f?T y(cn);

F o(P (; x; P ))g ?h#1 iT y(e ! cn)g
fT y(e); fF o(F riend);
F o(V)g T y (e ! cn)g
fF o(P OSS ((; x; Mother(x)^

P OSS (John; x));V)); T y (t)g
fF o(; x; Mother(x)^ fP OSS (V)g

P OSS (John; x)); T y (e)g
fT y(cn ! e); fT y(cn);

F o(P [; x; P ])g F o(U, Mother(U))g
fT y(e); F o(U)g fT y(cn);

F o(Mother)g
fF o(P OSS (John;U));

T y (t)g
fF o(John); fP OSS (U)g

T y (e)g
Figure 4.11: Interpreting John's mother's friend ...
(4.73) *The children who Sue said's mother was lazy were often late for
school.
Finally, we also have a basis for characterizing the observed dierence in inter-
pretation between (4.74) and (4.75):
(4.74) John's picture
(4.75) the picture of John
(4.75) can only be construed as the picture which is a portrayal of John; (4.74)
allows any arbitrary relation between John and the picture. On the analysis
suggested here, only (4.74) is taken to lead to the construction of a relation
P OSS between the two terms John and (; x; P OSS (John; x)). The preposi-
tional complement of the common noun picture is, however, taken to project a
second argument to a two-place predicate projected by the common noun, and
must be correspondingly interpreted.
This analysis of the genitive as a form of relative is far from new. Indeed,
it is novel only by virtue of the dynamic perspective within which it is set.48 A
theoretical consequence of postulating this analysis for a construction in English
is that we allow that even structures not generally thought to be available in
a head-initial language may nevertheless be manipulated in lexically specic
environments. We return to this in chapter 8, merely noting here that the
concept of cross-language variation as a parametric switch xed in language
acquisition is not a natural consequence of the dynamic perspective articulated
in this framework (see Chomsky 1981, Manzini and Wexler 1987).
4.5 Summary
In this chapter we have developed an account of relative clauses which abandons
the orthodox variable binding operator account of relatives. The account de-
pends on modelling the way in which information is built up through a left{right
dynamics, re ecting the result achieved by operator variable binding in other
frameworks through an encoded copy device which is an essential intermediate
step in this build-up of interpretation. The major dierence between this and all
48 Further evidence for this form of analysis in other languages is given by Baker (1996),
Kayne (1994). It should be noted that this analysis has no explanation for the occurrence of
anaphors within `picture-noun-phrase' constructions such as John's picture of himself. How-
ever following Pollard and Sag (1994) and Dalrymple (1993), in the face of examples such as
(i){(iii), we assume that the locality restriction constraining anaphor construal is not denable
over tree structure, but is, rather, more indirectly, a trigger for inference of a form yielding
some formula (; ):
(i) John's publicity campaign ensured that there were pictures of himself on every bill-
board.
(ii) John was on the point of leaving the show on the grounds that Mary's portrayal of
himself had been sharper than he had expected.
(iii) John walked in. Pictures of himself were all over the oor.
Summary 149
other approaches is the broadening of the basis for explanation. The emphasis
is no longer merely on construction of a structure re ecting semantic interpre-
tation, but on the sequence of steps needed to establish a route from the basic
requirement for a proposition as interpretation of the natural language string to
the nal result. At every step, emphasis has been placed on what information
is available at that step of the process.
The vindication of this approach lies in its two-fold success. First, it has
made possible cross-linguistic accounts of relative construal, while sustaining
a unitary and structural account of how pronouns are construed. Secondly, it
has led to accounts of head-initial, head-nal and head-internal relative struc-
tures which both integrate the various sub-types of relatives within a unifying
typology, and yet re ect their dierences. In all types of case, the various
sub-types and their specic instantiations in dierent languages have emerged
from general principles of how tree descriptions are monotonically built up. On
the one hand, resumptive pronouns occurring in both relative clause and topic
structures are explained without having to postulate any construction-specic
sub-type of pronominal, merely in virtue of there being two sorts of decoration
on a node { annotations and requirements { the latter inducing subsequent tree
development. On the other hand, the dierence between head-initial and head-
nal relatives is seen to emerge largely as a consequence of the dierence in
ordering; and the head-internal relative constructions, so problematic for other
frameworks, emerge as a natural part of the overall analysis, with the common
term of the two resulting structures licensed to occur in the LINK ed structure
in virtue of the linear order. And a characterization of genitive constructions ts
nicely within the overall family of explanations. Underlying each explanation
has been the concept of tree growth, with descriptions of a tree being built up
monotonically from some starting point through to some nal completely anno-
tated linked tree structure, with no requirements outstanding. The core of the
explanation is thus the dynamics of how structured information is incrementally
developed.
5
Wh Questions: A General
Perspective
5.1 Introduction
We have so far simply taken for granted that the characterization of gap-
containing relative clauses and left-dislocation in terms of an initially unxed
node will extend to the full range of phenomena analysed in other frameworks
by an operator-gap binding mechanism. Accordingly, wh expressions in chap-
ter 4 were dened as projecting a meta-variable WH that, if the wh expression
is initial in a clausal sequence, annotates an unxed node. This analysis of
wh-initial structures ies in the face of the orthodox assumption that wh ex-
pressions are variable binding operators, located in a structure at some Spec
CP position, from which they bind a discrete position (see Chomsky 1981 and
many references thereafter). Even formalisms which posit no gap position as
such, none-the-less characterize wh expressions as variable binding operators in
the sense that a process is dened at the node in the tree they inhabit, which
discharges a feature/assumption annotating a discrete node somewhere lower in
the tree (Steedman 1996, Sag and Fodor 1994, Jacobson 1996, Morrill 1994).1
In this chapter we take up our alternative analysis of wh structures in more
detail by looking at dierent forms of wh questions. First, we shall set out some
semantic problems facing accounts of wh questions in terms of a variable binding
operator. Then we shall show how the framework articulated here not only
captures wh-initial expressions as long-distance dependency phenomena, but
also straightforwardly characterizes wh-in-situ expressions. We shall go on to
demonstrate the fruitfulness of this model by taking up the problematic expletive
1 Analyses within HPSG dier according to whether the SLASH feature proposed con-
stitutes a distinct node in the tree or is projected by a verb, but the assumption that a
left-peripheral wh expression in a clause inhabits a node re ecting its left-peripheral position
is assumed by all (see Pollard and Sag 1994, Sag and Fodor 1994, Sag 1997, Bouma and
Malouf 2000).
150
The Semantic Diversity of wh Questions 151
wh constructions, and show that these expressions can be analysed as projecting
as a requirement what the non-expletive form projects as an annotation, an
analysis which enables us to explain the process of so-called `partial movement'
which is a characteristic feature of expletive structures. So we shall see that the
dynamics of tree growth which yields a typology for dierent forms of relative
clauses across languages also provides a basis in the apparently rather dierent
area of wh question structures.
5.2 The Semantic Diversity of wh Questions
Generally, work on the semantics of wh questions has started from the premise
that assignment of semantic objects to interrogative strings in a compositional
way is a minimum requirement for any account. It is therefore assumed that an
interrogative, like an indicative, should be assigned some form of denotational
content, the standard account being that the content of a question is given by
partitioning the universe of discourse to pick out some set of possible answers
(see Groenendijk and Stokhof 1997 for an overview of attempts to provide a
semantics for questions and a defence of this form of analysis). The slogan is
`To know the meaning of a question is to know what counts as an answer'.
Within this general set of assumptions, there are two major alternative char-
acterizations of wh questions: as an abstraction operator denoting a union of
possible partitions (see Groenendijk and Stokhof 1989, Higginbotham and May
1981, Higginbotham 1996), and as a form of existential quantication denoting
some subset of possible answers (see Karttunen 1977). The problem which these
accounts face is that neither of them is suÆcient to cover all the data, giving
rise to a phenomenon of multiple analyses for a single type of expression, much
as with anaphoric expressions. The question in (5.1), for example, seems to ask
for a listing of all places where poison was put down, but the question posed in
uttering (5.2) does not require any such exhaustive listing:
(5.1) Where did the ratman put down the poison? The sitting-room, the
kitchen and the larder.
(5.2) Where can one buy an Italian newspaper? At the station.
Evidence that this requires an analysis positing dierent types of wh expres-
sions is taken to be provided by the corresponding embedded questions (see
Groenendijk and Stokhof 1997), whose interpretation is presumed to consti-
tute some (exhaustive or non-exhaustive) set of possible answers.2 For, despite
structural equivalence between individual embedded questions, they may nev-
ertheless be interpreted in systematically dierent ways. The truth of (5.3)
depends on John knowing every place where the ratman put down poison: we
would hardly agree that John knew where the ratman put down poison if he
did not know for each room in the house whether or not poison had been put
2 For extensive argumentation against this assumption, see Ginzburg 1992, 1995.
152 Wh Questions: A General Perspective
down there. (5.4), on the other hand, asserts only the weaker claim that John
knows at least one place where an Italian newspaper can be bought:
(5.3) John knows where the ratman put down the poison.
(5.4) John knows where one can buy an Italian newspaper.
Given a standard principle of compositionality and the assumption that the
meaning of expressions is given by their denotational content, such data thus
appear to indicate ambiguity in the wh expressions themselves. Both types of
analysis of wh expressions nevertheless assume that the scope of the wh operator
is indicated by the surface position of the wh expression, c-commanding the
clause to which it is adjoined.
The dierent forms of analysis required for wh questions are not exhausted
by the apparent potential for universal or existential construal of the wh ex-
pression. Evidence has been put forward by Reinhart (1997) that wh-in-situ
constructions require a discrete form of analysis to capture their supposed con-
strual externally to their containing context:
(5.5) Which lady wants to read which book?
(5.6) Who will be annoyed if we invite which philosopher?
As she points out, the in-situ form must not be construed from the position
internal to the conditional internal to (5.6) (on the assumption that the condi-
tional corresponds to material implication) if we are to debar answers to (5.6)
such as (5.7), for the falsity of Donald Duck being a philosopher will have the
unfortunate consequence of satisfying the structure in (5.6) on the assumption
that the wh form is some type of indenite contained within the antecedent of
the conditional:
(5.7) John will be annoyed if we invite Donald Duck.
The choice function analysis which she proposes has the advantage of being able
to characterize wide scope construal of the wh expression vis-a-vis its containing
context while nevertheless being understood as taking narrow scope of the in-
situ wh expression vis-a-vis the other wh expression (see Reinhart 1997, Winter
1997), though such an account of wh-in-situ constructions, distinguishing their
denotational content from wh-initial expressions, has the eect of characterizing
(5.8) in two dierent ways, depending on whether the rst wh expression is
analysed as a regular wh operator or as a wh-in-situ expression:3
(5.8) Which woman read which book?
Despite this addition to the types of wh expression, the list is still not complete.
The assumption that the semantics of a question is given by its possible an-
swers, together with the assumption that a wh expression is a variable binding
3 Reinhart (1997) only considers the diÆculty which wh-in-situ constructions pose for the
Karttunen-style analysis of wh expressions as a form of indenite.
operator whose scope is given by its structural position, gives rise to yet more
ambiguity. It appears that a discrete type of so-called functional questions must
also be dened for wh-initial questions, for in questions such as (5.9) { despite
the structural position of the wh expression c-commanding the quantied ex-
pression { the form of answer demands a construal in which it is the universally
quantied expression which contains everything else within its scope, and not
the questioned element (see Engdahl 1986, Chierchia 1992):
(5.9) Who did everyone recommend? Their best student.
p9f [p = 8x(recommend(x; f (x)))]
`What is the function that pairs people with those they recommend?'
Note that in this case it is the assumption that an initial wh expression marks
its scope relative to the surface string which requires the analysis of wh expres-
sions as a function that picks out an individual relative to some antecedently
articulated quantier. It should be noted that this is not a locally denable
dependency between a pair of quantifying expressions within the same clause.
It would appear, rather, that a wh expression may be construed as having ar-
bitrarily narrower scope than some quantier that it c-commands:
(5.10) What is the Union insisting that almost every British farmer should
get rid of? At least 1,000 cattle.
We know that this phenomenon is not due to the wide scope potential of the
lower quantier in a case like this, as a quantier such as almost every may not
be construed as taking scope beyond the clause in which it is contained { (5.11)
has no such wide scope interpretation of the embedded subject:
(5.11) Most British farmers are complaining that almost every country of the
EU fails to appreciate the problem.
There is yet a further type of question, a so-called choice interpretation, posited
for a pair of disjunctive questions. In these, the answer cannot be construed as
the union of the two answers (as the partition theory of Higginbotham 1996 or
Groenendijk and Stokhof 1989 would lead one to expect), and so the question
re ecting such answers has to be analysed as something other than a regular
disjunction:
(5.12) Where's your father? Or where's your mother?
This problem also arises even within a single sentence, such as:4
(5.13) Who likes Bill or Jill?
Finally, no single one of these suggested denitions of wh expressions provides
a basis for characterizing questions where the set of possible answers is unclear
or unspeciable (open questions):
4 Thanks to Jonathan Ginzburg for pointing this out to us. See Ginzburg and Sag (1999)
for extensive discussion of the problems which disjunction poses for Groenendijk and Stokhof's
account.
(5.14) What are questions?
(5.15) What is Buddhism?
(5.16) What causes a sun's death?
At best, the characterization of a question as the set of possible answers whose
answers are unclear, vague or unspeciable is that the question is correspond-
ingly unclear, vague or unspeciable. But this contradicts common sense: it
does not follow that a question each of whose answers is unclear or vague or
unspeciable is necessarily an unclear, vague or unspeciable question.
The two assumptions on which such analyses of questions are based thus
lead to a multiplicity of types of question. As Groenendijk and Stokhof observe
(1997), the situation is redolent of anaphora, with no single answer to the ques-
tion `What is the meaning of a wh question'. In all cases, the answer is merely
that the meaning of a wh question is one of a number of things, recovery of
which, in context, depends on recovering the type of answer the speaker has
in mind. Not surprisingly, given this unsatisfactory conclusion, as in the lit-
erature on anaphora, there have been various attempts to reduce the observed
heterogeneity, none entirely successful.
5.2.1 Scopal Properties of wh Expressions
There is a further set of puzzles as to what concept of scope should be attributed
to wh expressions. Groenendijk and Stokhof and many others following them
assume that the position of the wh expression in the string determines its scope
as a variable binding operator. However, this poses a number of problems. First,
with the so-called functional questions apparently demonstrating the construal
of the wh form as having narrower scope than other quantifying expressions
in the string, it might seem that there is ambiguity according to whether the
position of the wh expression dictates its scope relative to other expressions.
Second, in addition to either of these possibilities, there is, in some languages,
a discrete form of wh, the expletive wh, which is said to extend the scope of the
wh to that which is wider than would otherwise be projected from its surface
position (see Pafel 1996, Johnson and Lappin 1999, Pollard and Yoo 1998, etc.).
Thus, in German, (5.17) and (5.18) are synonymous. In both cases, so the
explanation goes, the scope of the wh operator is that of the highest clause:
(5.17) Was glaubst du, mit wem Jakob gesprochen hat? German
What think you with whom Jacob spoken has
`With whom do you think Jacob was talking?'
(5.18) Mit wem glaubst du, dass Jakob gesprochen hat?
With whom think you that Jacob spoken has
`With whom do you think Jacob was talking?'
The eect of the expletive was is reported by McDaniel (1989) to indicate in
(5.17) that the lower mit wem is to be construed as taking wide scope with
respect to the higher verb glaubst in the same manner as (5.18). One stan-
dard analysis of these constructions is accordingly that the expletive is a scope
marker, exceptionally extending the scope of the lower wh expression to that of
the clause marked by the expletive (more accurately, the highest of a sequence
of such expletives):
(5.19) Was glaubst du, was Hans meint, mit wem Jakob
What think you what Hans said with whom Jacob
gesprochen hat?
talking has
`With whom do you think Hans said Jacob was talking?'
However, despite this supposed scope-marking property of the expletive form,
when such question forms are combined with other quantifying expressions, the
eect is to strongly encourage readings in which the wh expression is interpreted
as taking narrower scope than the quantier between the expletive and full wh
form (see Pafel 1996 for arguments to this eect):
(5.20) Wo glaubt jeder, dass sie gerne leben wurde
Where thinks everyone that she live like would
`For which place y, for each x, x think she would like to live in y?'
`For each x, for which place y, x think she would like to live in y?'
(5.21) Was glaubt jeder wo sie gerne leben wurde
What thinks everyone where she live like would
`For each person x, for which place y, x think she would like to live in
y?'
`For which place y, for each x, x think she would like to live in y?'
A dierent puzzle is posed by wh-in-situ expressions. As some form of a quan-
tifying expression, these might in principle be expected to be interpreted as
falling within the scope of other quantifying expressions. But, in fact, they are
only sensitive to some other wh expression in the sentence, and not to any other
c-commanding quantier. Thus (5.22) is ambiguous, but only in two ways, de-
pendent on whether the wh-in-situ form is construed relative to the embedded
wh expression or to the higher wh expression. Hence the two possible types of
answer (a) and (b) to (5.22):
(5.22) Which teachers told every student who should concentrate on which
19th century author?
(a) Bill told every student who should concentrate on Wordsworth;
Sue told every student who should concentrate on Henry James.
(b) Bill told every student who should concentrate on which author;
and so did Sue.
We are thus faced with data for which the various analyses do not seem to
add up to a consistent overall analysis. The position of the wh expression has
been taken in the main to indicate its scope over other expressions in the string;
although sometimes (in functional interpretations) it does not. The scope eects
of its position are apparently overridden in two ways { either by some process
construing the wh expression as dependent on some subsequent expression in
the string, or by the presence of a scope-marking expletive apparently extending
the scope of the wh expression so that it has scope over expressions before it in
the string. In this latter case extra problems arise, because that wh expression,
notwithstanding the presence of the expletive, is in some examples construed as
having a scope considerably narrower than that marked by the expletive. And
nally, unlike all the other wh forms, an in-situ wh doesn't seem to interact
with any quantier other than other wh expressions. A single base from which
to explain these various wh forms might seem to be a goal that has to be
abandoned.
These analyses of wh expressions have been, without exception, grounded on
syntactic explanations given in terms of tree congurations decorated by words
of the string (or at least abstract characterizations of those words). Against this
background, the assumption that a wh expression is a variable binding operator
has been taken to be uncontentious, its binding properties dened in terms of
the tree relation between the node it decorates and nodes decorated by other
words. We formulate an alternative analysis not based on these assumptions,
in which a wh expression underdetermines its contribution to the interpretation
of the whole. According to our view, attributing scope to a wh expression is an
attribution of properties of some enriched form { the answer { to the incomplete
structure { the question; and so it does not characterize the properties of that
input structure independently of its available enrichments. What we propose is
that wh expressions project a distinguished meta-variable, WH, which, if initial
in a clausal sequence, is projected as an annotation of an unfixed node in the
logical form under construction. As we shall see, the dynamics of the growing
of tree structures and the progressive xing of tree node positions will, as in the
case of relative clauses, enable us to characterize a supposedly heterogeneous
set of data in a revealing and natural way.
5.3 Wh -initial vs wh -in-situ Structures
The rst observation about the projection of some WH meta-variable, from the
perspective of the present framework, is that the annotation of a node with a
meta-variable is independent of whether that node is or is not xed: in either
case the nal result will be a tree whose root node is annotated with a formula
containing a WH meta-variable. Upon this analysis then, a pair of questions
such as (5.23) and (5.24), whose primary dierence is that one has the wh ex-
pression at a left-dislocated position and the other at an in-situ position, will
Wh-initial vs wh-in-situ Structures 157
give rise to the same formula at the root node of the resulting tree (displayed
schematically in gure 5.1):5
(5.23) Who did Bill see?
(5.24) Bill saw who?
fT y(t);
F o(See(WH)(Bill)); }g
fT y(e); F o(Bill)g fT y(e ! t);

F o(See(WH))g
fT y(e ! (e ! t)); fT y(e); F o(WH)g

F o(See)g
Figure 5.1: The logical form of a wh question

The structural asymmetry between the two types of wh question is re ected in a
dynamic distinction, where in (5.24) the wh expression annotates a xed node,
and in (5.23) a wh expression annotates an initially unxed node whose tree
position is determined at a subsequent step in the parsing process. In English,
as in German, the dierence between the two structures is also, on our analysis,
re ected in the presence or absence of the feature Q: wh-initial questions have
an associated feature Q to identify them as such, whereas wh-in-situ structures
do not. This property, as is familiar, is needed in these languages for a num-
ber of reasons, primary amongst which is the requirement in embedded questions
5 As in other tree displays, we suppress all information except that which is essential to
the point being made. This analysis of wh questions is similar to that of Ginzburg and
Sag (1999), who argue that a wh expression is not a generalized quantier but a specialized
variable { in their terms, a restricted index. However, in construing the concept of index and
its associated uniqueness presupposition semantically, they are forced to postulate ambiguity
as between functional construals of wh questions and what they take otherwise to be wide
scope construals for all wh expressions in questions, and also as between dierent construals
of which, as between what they call n-jective construals and independent construals (both of
which they take to involve no scope dependence). They grant that if they adopted the stance
that such uniqueness presuppositions were only induced as a general pragmatic consequence,
as argued by Engdahl (1986) { a position with which we would concur despite Ginzburg and
Sag's arguments to the contrary { a unitary account would be available. As they themselves
observe, reporting the observation of Engdahl (1986), such uniqueness presuppositions are in
any case not sustained in multiple wh questions.
associated with verbs such as wonder that the wh expression identifying the
complement clause as a question must occur initially:6
(5.25) John wondered who Bill saw.
(5.26) *John wondered Bill saw who.
As we shall see shortly, the feature Q is also critical for determining the distri-
bution of wh forms in the presence of the expletive German wh form.
The meta-variable WH projected by the wh expression can remain unre-
placed in a resulting formula.7 Furthermore, the wh expression itself will not
have its scope xed with respect to other terms in the resulting formula, so a
quantied term contained within a wh question may fail to have its requirement
of an associated scope statement fullled. In this sense wh questions are essen-
tially incomplete { a full semantic evaluation cannot be dened until an answer
is provided. As we shall see in due course, this leads to a dierent account
of scope in wh questions. Before reaching that point, however, we explore the
dierent kinds of analyses which the assumption that wh expressions project a
meta-variable provides.
5.3.1 Wh -in-situ Structures
In Germanic and Romance languages, the so-called wh-in-situ expressions are
standardly reported to be peripheral { normally, though not invariably, asso-
ciated either with an echoic use (characteristically associated with a marked
intonation) or as a second member of multiple wh expressions:
(5.27) John saw who?
(5.28) Who do you think saw what?
However, as pointed out by Ginzburg (1992, 1995), they are associated with a
number of dierent uses:
(5.29) When you get to the oÆce, your rst action will be to do what?
(5.30) Having arrived there, you did what exactly?
In other languages, they form the regular form of question, as in Chinese:
6 We might also posit an additional feature QUE for characterizing all questions, both
yes-no and wh. Each of these features is introduced as a DU formula of the form Cat(Q),
Cat(QUE), etc., picking out a dierent set of structures of type t. It is the feature QUE that
we might use to dene the conditions inducing auxiliary subject inversion in root clauses in
English, a phenomenon which is arguably a re ex of a verb-adjunction process carried over in
a more specialized form from German. In German, all verbs may be introduced at a locally
unxed node, but in English, where this process has atrophied, the option of building such a
node is an action specic to auxiliary verbs (this context for auxiliary occurrence was ignored
in the lexical specication of auxiliary verbs in chapter 3). See sections 3.2.2, and 5.4.2.1.
7 We shall consistently use `WH' to indicate the meta-variable assigned as the F ormula
value, and `wh' to indicate the natural language expression which projects such a meta-
variable.
(5.31) Ta shuo shenme [Chinese]
He say what
`What did he say?'
In most languages, in-situ uses are not restricted to particular kinds of structure.
A wh expression is a noun phrase, and can occur anywhere that noun phrases
can { in relative clauses, sentential subjects, etc. In other words, wh-in-situ
expressions are generally not subject to any of the island restrictions familiar
from long-distance dependency:8
(5.32) The journalist that leaked which document to the press became
famous overnight?
(5.33) That he had missed which movie upset him ?
(5.34) Ni mai-le [[shei xie] de shu] [Chinese]
You buy-asp who write rel book
`Who is the x such that you bought books that x wrote?'
(5.35) Lisi da-le shei shi ni hen bu gaoxing
Lisi hit-asp who make you very not happy
`That Lisi had hit who made you unhappy?'
However in some languages, the occurrence of wh expressions in situ is subject
to more stringent restrictions than in wh -initial structures. For example, in
Iraqi Arabic, wh expressions may normally occur in subordinate clauses only
if the clause is non-nite. If the embedded clause is nite, only the wh-initial
form is possible:
(5.36) Mona raadat [tijbir Su'ad tisa'ad meno] [Iraqi Arabic]
Mona wanted to force Suad to help who
`Who did Mona want to force Suad to help?'
(5.37) meno Mona raadat [tijbir Su'ad tisa'ad]
who Mona wanted to force Suad to help
`Who did Mona want to force Suad to help?'
(5.38) *Mona tsawwarat [Ali ishtara sheno]
Mona thought Ali bought what
(Intended: `What did Mona think that Ali bought?')
(5.39) shenoi tsawwarit Mona [Ali ishtara ti]
What thought Mona Ali bought
`What did Mona think Ali bought?'
8 The one possible exception is why-type adjuncts which are reported not to be able to
occur inside relative clauses (Huang 1982).
In Iraqi Arabic, wh expressions, when in situ, are thus licensed only in environ-
ments in which antecedent anaphor relations are.
Systematically across languages, this pattern of wh -in-situ structures carries
over to multiple wh structures: these display a combination of wh -initial prop-
erties and wh -in-situ properties. In English the initial wh cannot be interpreted
with the supposed gap inside a relative clause (an unxed node cannot nd its
position across a LINK relation), whereas the second wh may occur anywhere
{ for example, inside a relative clause:
(5.40) Who do you think should review which book?
(5.41) *Whoi did the journalist leak the document in which Sue had
criticized ei to which press?
(5.42) *Which documenti did the journalist that leaked ei to the press resign?
(5.43) Who reported the journalist that leaked which document to the press?
In Iraqi Arabic, to the contrary, the second of a pair of wh expressions may occur
in an embedded structure projecting multiple wh-question interpretations only
if the clausal structure in which it is contained is non-nite:
(5.44) sheno ishtara Ali [minshaan yenti li-meno] [Iraqi Arabic]
What bought Ali in order to give to whom
`What did Ali buy to give to whom?'
(5.45) *meno tsawwarit [Ali xaraj weyya meno]?
Who thought Ali left with whom
`Who thought that Ali left with whom?'
This array of properties is problematic for analyses of wh-in-situ forms in which
these undergo the same movement as wh-initial forms but covertly (see Huang
1982 among others), as the constraints on this movement diverge from familiar
subjacency restrictions: they are either more stringent, as in Iraqi Arabic, or
less so, as in English, Chinese etc.9 Multiple wh questions, as Simpson (2000)
9 These and other diÆculties are presented in detail in Simpson (2000), who argues that
features must be checked from their in-situ position, hence in congurations other than a
local Spec Head conguration. However, the checking of features for wh expressions from
an in-situ position poses its own problems, because this raises the question as to why move-
ment should ever take place. Simpson (2000) argues that movement to a Spec CP position of
the wh expression is obligatory in order to `disambiguate' an otherwise `crucially underdeter-
mined' complementizer, but this analysis hovers unsatisfactorily between underspecication
and ambiguity. If the complementizer in question is a `general polarity-oriented head' and is
underspecied with respect to any more specic value, then it will lack any feature particular
to wh forms to license wh-in-situ forms. Ambiguity, on the other hand, is by denition re-
ected in the grammar in the form of discrete items, and on this analysis, each such particular
type of `polarity-oriented' complementizer will project the particular feature needing to be
checked; but then no movement will be necessary. Despite Simpson's informal statements,
conventional assumptions about grammar-parser relations do not permit problems posed by
the need to disambiguate a string as between one of several independent structures to be
re ected in the grammar formalism, for this is not a phenomenon intrinsic to any one of the
structures posited.
demonstrates, provide additional problems for Minimalist assumptions, because
the feature +WH on the wh expression is a strong feature to be checked by
adjacency to a node bearing the feature Q triggering movement from some in-
situ position; but once checked, the whole structure is taken to be licensed for
the wh feature, allowing multiple occurrences of a given wh expression. The
problem with Iraqi Arabic multiple wh structures is that the restrictions on the
wh-in-situ form are carried over to these structures, even though the prohibited
environments for the in-situ form are the very environments in which movement
must take place.
5.3.2 Wh -in-situ from a Dynamic Perspective
In the present analysis, the10in-situ forms are simply the annotation by a wh
expression of a xed node. The lexical characterization of what in English
re ects the two types of structure:
what IF f?T y(e)g
THEN IF fh" i?T y(t); ?9xT n(x)g
THEN put(F o(WH); T y(e); [#]?);
go(h" i);
put(Cat(Q))
ELSE put(F o(WH); T y (e); [#]?)
ELSE ABORT
The condition stipulates that should the node to be annotated be unxed, then
the resulting action is to add the DU formulae F o(WH), T y(e) and annotate
the dominating node with Cat(Q), and that otherwise, it simply adds the DU
formulae F o(WH),T y(e).11 As indicated above, the feature Q is a classicatory
device for picking out wh-initial questions without semantic consequence.12
In the in-situ cases, no such feature Q is added: the only property distin-
guishing wh-in-situ structures from non-wh question structures is the meta-
variable WH itself. It follows immediately that there will be no restrictions
precluding a wh-in-situ expression from occurring in relative clauses or other
environments which do not license long-distance dependencies, as there is, on
this account, no long-distance dependency to establish. Wh -in-situ expressions
also occur in non-restrictive as well as restrictive relative clauses (contra any
analysis which analyses wh-in-situ expressions as subject to covert movement
at LF):
10 We follow Ginzburg (1992) in assuming that in-situ wh forms should be licensed in Eng-
lish as freely available. See Ginzburg 1992 for detailed discussion of in-situ wh questions,
and evidence demonstrating that English wh-in-situ phenomena should not be dismissed as
peripheral, as is standard.
11 As in the case of linguistic names, each wh expression is taken to project a uniquely
distinguished variable of the form WH, so in eect we are assuming an open-ended set of such
meta-variables. We ignore this here.
12 This lexical specication of an annotation to a node requiring type t with a DU formula
Cat(Q) is the lexical analogue within this framework of the wh criterion: see May 1985, Rizzi
1990.
(5.46) Which director, who has directed which lms, did Sophia Loren
marry?
(5.47) Which famous lm star did the Queen, who was visiting which
hospital at the time, insist on seeing?
(5.48) Which famous dog-show in the US was visited by the Queen, who
bought how many corgis there?
Wh -in-situ forms in Iraqi Arabic do not follow this pattern, because they re-
quire a locality restriction analogous to that of re exive anaphors. This can be
characterized in terms similar to the locality restriction formulated for pronouns
(see section 5.4.2.3). W h expressions, e.g. sheno (= `what'), can then be taken
either to annotate an unxed node, or to annotate a xed node, the latter rela-
tive to the locality constraint that the feature Q be local (the characterization,
"0 "1 (Cat(Q)) in the lexical entry below):13
sheno
IF f?T y(e)g
THEN IF fh" i(?T y(t)); ?9xT n(x)g
THEN put(F o(WH); [#]?);
go(h" i);
put(Cat(Q))
ELSE IF f"0"1 (Cat(Q))g
THEN put(F o(WH); T y (e); [#]?)
ELSE ABORT
ELSE ABORT
We have, that is, the analogue of an anaphor for the WH meta-variable { a
wh expression projecting an interpretation only in a certain most-local environ-
ment.14
It should be noted that this characterization depends on the presence of a
feature Q annotating all top nodes of wh question structures; and in making this
assumption, the feature Q is used as a more general classicatory device in Iraqi
Arabic than that for which it was previously used in characterizing languages
such as English and German. In Iraqi Arabic, given the locality restriction
associated with the wh forms themselves, it is essential that the Q feature pick
out the whole class of wh questions. In Germanic languages, to the contrary,
given the particular properties associated with clause-initial wh expressions, it
13 The argument put forward by Ouhalla (1996) that Iraqi Arabic wh -in-situ questions must
have no feature Q on the grounds that if they did, movement would be enforced, as it is in
wh -initial structures, is theory-internal and does not carry over to this framework. In virtue
of the locality eects observed here, Ouhalla argues that wh expressions are anaphors with
the feature +WH as antecedent. The analysis given here preserves the spirit of anaphoricity
which Ouhalla's analysis captures in rather dierent terms.
14 It is arguable that all meta-variables have both locally restricted forms and forms which
are not so locally restricted. See Progovac 1991 for evidence that negative polarity forms
in Serbo-Croatian may range over those which are analogous to anaphors and those which,
conversely, have an anti-locality restriction.
Expletive wh Structures 163
is those fronted wh constituents and not the whole class of wh questions which
the classicatory feature Q picks out.15
5.4 Expletive wh Structures
A further type of wh expression is the so-called expletive form of wh, occurring
in German, Hungarian and Iraqi Arabic and serving as an anticipatory indi-
cation of a wh question with a full wh expression to follow.16 This is where
the relationship between requirements on a node and the subsequent successful
annotation of that node can again be put to use { wh expletives, we propose,
project a requirement for an annotation which only a corresponding full wh
form can supply. It is this relationship between annotations and requirements
for those annotations which provides a basis for explaining the puzzles posed
by expletive wh phenomena and a structure called `partial movement' which in
some languages is characteristic of these expletive wh forms.
5.4.1 Partial Movement
Expletive wh constructions and the process of partial movement go together,
because the latter is part of a movement-based explanation of expletive wh forms
for languages such as German in which the primary form of wh expression occurs
clause-initially. The presence of an expletive wh form, which itself occurs clause-
initially, has the eect of not only inducing a lower clause-initial wh expression,
but also ensuring that that wh expression does not occur in the position it would
have done in the absence of any such expletive expression:
(5.49) Was glaubst du, wen Jakob half? [German]
What think you whom Jacob helped
`Who do you think Jacob helped?'
In (5.49) was is the expletive, wen the `partially moved' full wh form following
a verb which need not itself select for question forms. Expletive wh phenomena
also occur in languages where the wh form characteristically occurs in an in-situ
position. Iraqi Arabic, for example, has an expletive wh form, prexed on the
verb of the higher clause which licenses construal of the wh in a subordinate
nite clause, which is otherwise debarred:
(5.50) sh-'tsawwarit Mona Ali raah weyn [Iraqi Arabic]
Q thought Mona Ali went where
`Where did Mona think that Ali went?'
15 Though we do not explore the consequences of this language-particular use of syntactic
features here, it re ects the fact that in Iraqi Arabic, the in-situ position is the unmarked
form of wh question when this position is licensed, whereas in the Germanic languages, it is
the fronted wh-initial structure which is the unmarked form of wh question.
16 They also occur in Hindi (see Mahajan 1990, Dayal 1994), though we leave Hindi on one
side. In that language the apparently expletive form is anticipatory of any arbitrary type
of question, suggesting a further application of a syntactic feature QUE which picks out the
broader class of questions, to include yes-no questions (see Dayal 1994, 1996).
The encliticized form sh- is a device which explicitly licenses the presence of a
wh expression in the next tensed clause, and does not license wh expressions
across an indenite number of tensed clause boundaries:17
(5.51) *sh-i'tiqdit Mona meno tsawwarit Ali sa'ad meno
Q believed Mona who thought Ali helped who
`Whoi did Mona believe ei thought Ali helped who?'
Unlike Iraqi Arabic, German allows partial movement with an indenite number
of expletive was expressions followed at some lower clausal boundary by the full
wh expression, which must occur initially in the clause:18
(5.52) Was glaubst du, was Hans meint, wen Jakob liebt?
What think you what Hans said whom Jacob loved
`Who do you think Hans said Jacob loved?' [German]
Partial movement is a long-distance dependency eect, since not only the ex-
pletive form but also the full wh expression is free to occur arbitrarily far from
the ultimate gap position:
(5.53) Was glaubst du, wen Hans meint, dass Jakob half?
What think you whom Hans said that Jacob helped
`Who do you think Hans said Jacob helped?'
In some dialects of German, the anticipatory sequence of was expletives must be
unbroken, with each expletive only able to license a wh expression across a single
tensed clause boundary. In others, however, speakers report the acceptability
of a single anticipatory was followed by an intervening dass:
(5.54) Was glaubst du, dass Hans meint, wen Jakob half?
What think you that Hans said whom Jacob helped
What is essential is that there be no occurrence of such a was complementizer
following a full wh form, unless (as in (5.57)), there is some following second
full wh form of which the intermediate was is an anticipation:
(5.55) *Wen glaubst du, dass Hans meint, was Jakob half?
whom think you that Hans said what Jacob helped
(5.56) *Wen glaubst du, was Hans meint, dass Jakob half?
Whom think you what Hans said that Jacob helped?
17 The sh- form is an abbreviated form of sheno meaning `what'.
18 Examples are taken primarily from Simpson 2000 and Hohle 1996.
(5.57) Wer glaubt, was sie meint, wen Hans liebt?
Who believes what she said whom Hans loves
`Who believes whom she said Hans loves?'
In German, other wh expressions also have apparently expletive uses, though
these require successive occurrences of the particular expletive form, without
any intervening dass:
(5.58) Wer glaubst du, wer Uli half?
Who think you who Uli helped
`Who do you think helped Uli?'
(5.59) Wen meint Karl, wen Uli half?
Whom said Karl whom Uli helped
`Who does Karl think Uli helped?'
(5.60) Wie nimmt man, an wie der Prozess ausgeht?
how assumes one how the process ends
`How does one assume that the trial ends?'
Though the expletive form is unambiguously construed as a marker of some
lower wh expression, as re ected in the case assignment to these German ex-
amples, it may on the other hand display case properties which re ect its own
position in the lexical sequence. For example, in Hungarian, which also displays
partial movement, the expletive form is case-marked not by the embedded posi-
tion corresponding to the interpretation to be projected, as in German, but by
the predicate local to its occurrence in the string:
(5.61) Mit mondtal, hogy kinek vett Janos
WhatACC said2:sg:indef:ACC that whoDAT bought JanosNOM
szinhazjegyet? [Hungarian]
theatre-ticketACC
WhatACC did you say for whom John bought a theatre-ticket?
`For whom did you say John bought a theatre-ticket?'
Note in (5.61) the accusative marking on the expletive form mit, and not the
dative, re ecting the resulting position of the WH variable.
5.4.1.1 Minimalist Accounts: Direct versus Indirect Dependency
Faced with this amount of variation, it is perhaps not surprising that heterogene-
ity in the analysis is taken to be required; and current Minimalist explanations
are divided over the underlying uniformity of the data (see Horvath 1997, Dayal
1996, Beck and Berman 1996 for arguments that the phenomena require more
than one form of explanation). But the real puzzle for Minimalist assumptions,
as is demonstrated in detail by Simpson (2000), is why partial movement should
exist at all. Given that the expletive ensures (e.g. through movement to a Spec
CP position) that the wh feature is checked under adjacency, there should be
no need of movement of the full wh form, and with no need of movement, no
movement should be licensed. It is thus far from clear how the obligatory nature
of partial movement in a language such as German can be licensed, let alone
forced. To make matters worse, such partial movement should be to a node
which is marked for the feature Q, but partial movement constructions char-
acteristically exist as complements to predicates which select as complements
propositions which are not questions, hence with the feature Q.19
The various Minimalist analyses of these data can be seen as attempts to
overcome this problem { none of them fully satisfactory. There are two princi-
ple variant forms. One entails analysing the anticipatory wh expression as some
form of expletive, creating a single chain from the expletive form to the wh form
(and thence to its gap). This direct dependency approach is so called because
it directly re ects the interpretation, which corresponds to long-distance depen-
dency construals as in English (McDaniel 1989, Chomsky 1995, Brody 1994,
Beck and Berman 1996). The other, so-called indirect dependency approach
analyses the expletive forms as question words in their own right, each clause
corresponding to an independent question with local binding of some posited
empty position for the expletive wh form (Dayal 1996, Horvath 1997). The
direct dependency approach is clearly inappropriate for Hungarian, since case
assignment in these Hungarian structures must not re ect the position of the
gap. However, the indirect dependency approach cannot be seen as a simple
apposition of two entirely independent questions, since a quantier within the
initial clause may bind a pronominal in the second (Beck and Berman 1996,
Dayal 1996, Horvath 1997):
(5.62) Was glaubt jederi , wohin eri geht? [German]
what thinks everyone where he went
`Where does everyone think he went?'
(5.63) Mit nem hisz senki, hogy milyen torteneteket
WhatACC not believe no-one that what stories
terjeszt rola a felesege [Hungarian]
spreads about him the wife-hisNOM
`What stories doesn't anyone believe that his wife spreads about him?'
Moreover, there is the problem of what should be taken as the trigger of the
two movement processes that have to be invoked { on the one hand, partial
movement to a Spec CP position whose CP is immediately dominated by a
predicate which precludes questions, and, on the other, a subsequent pied-piping
at LF, which is needed to ensure the interpretation of the whole as a single
question.20
19 This problem applies to all analyses in which the feature Q is taken to be a feature
indicating the semantic type of the clause, as in Brandner 1996.
20 Pied-piping at LF, which is highly problematic under Minimalist assumptions, is also
needed in some direct dependency analyses (see Fanselow 1997).
Even leaving aside how to dene appropriate triggers, there is the problem
that indirect dependency approaches fail to explain why no expletive forms occur
as second or subsequent members of a multiple wh construction:21
(5.64) *Wer hat was gedacht, wen wir anrufen sollen? [German]
Who has what thought whom we call up should
`Who thinks what who should we call up?'
There is also the problem that the relation between the expletive wh form and
its supposed site does not parallel wh movement. As Horvath points out, in
Hungarian a wh determiner may occur externally to its DP, indicating movement
across a DP boundary, an option which is not available to the expletive wh
forms:22
(5.65) Kinek lattad a feleseget? [Hungarian]
WhoDAT saw2sg:def:DO the wife3sg:poss:ACC
`Whose wife did you see?'
(5.66) *Mit lattal kinek a feleseget?
WhatACC saw2sg:indef:DO whoDAT the wife3sg:poss:ACC
`What did you see whose wife?'
The expletive also cannot be followed by the full wh expression embedded in
a non-nite clause, despite the fact that movement across such a boundary is
unproblematic:
(5.67) Kivel akarsz beszelni?
Who-with want2sg:indef:DO talkINF
`With whom do you want to talk?'
(5.68) *Mit akarsz kivel beszelni?
WhatACC want2sg:indef:DO who with talkINF
What do you want with whom to talk
`With whom do you want to talk?'
This restriction is quite unlike wh movement, and requires special explanation.
We should notice in passing that it echoes the Iraqi Arabic data in which the
function of the expletive is to signal the presence of a wh expression in a locally
subordinate tensed clause. And all of these indirect dependency approaches are
inappropriate for the German non-was forms of expletive (5.58){(5.60), which
have to be set aside as involving a separate pronominal form of wh expression
21 Simpson (2000) suggests a case explanation, but this leads to the consequent need in
explaining Hungarian sentences analogous to the German (5.54) (in which there is an inter-
vening non-expletive complementizer) to postulate a PF-driven movement of an expletive form
in Hungarian of which there may be no phonetic realization.
22 Hungarian marks deniteness and indeniteness constraints on the object as a suÆx on
the verb, here indicated as e.g. `def.DO'.
despite the parallelism observed between these and the more regular was form
of expletive (see McDaniel 1989, Reis 1996).23
It is in the light of these and other problems that some authors grant the
possibility of setting up quite dierent types of analysis for expletive forms, with
German said to display a direct dependency type of structure, Hungarian an in-
direct dependency form of structure (see Beck and Berman 1996, Horvath 1997,
Dayal 1996). The current overall position in Minimalism debates about partial
movement thus presents a lack of consensus as to how to resolve the problems
these structures pose, with no very strong justication for the movement of the
intermediate form and no apparent means of characterizing the various dierent
case realizations allowed across dierent languages in a uniform way.
5.4.1.2 HPSG { A Feature-based Account of wh Variation
The exibility of feature specications provided by the HPSG formalism allows
the wh phenomena to be characterized relatively unproblematically (Johnson
and Lappin 1999), though the formalism's lack of intrinsic dynamism makes it
not well suited to explain why the data should pattern as they do. The diversity
in the data of long-distance dependency, wh -in-situ and expletive wh forms is
simply carried over to the analysis itself. Three separate binding operations
are dened for the dierent structures: dislocated wh expressions, in-situ wh
expressions, and expletive wh expressions. Regular wh structures are character-
ized by feature percolation of SLASH features to be discharged by the abstract
Q complementizer in the presence of the initial wh expression; wh -in-situ struc-
tures are characterized as projecting a distinct feature QUE, also discharged by
the abstract Q complementizer. The Q complementizer itself involves a three-
way disjunctive specication to embrace rigidly wh -in-situ languages, mixed
languages that display both wh -in-situ and `moved' forms, and languages which
allow all wh forms to cluster at the front of the clause. The expletive wh is anal-
ysed as itself a complementizer that discharges a newly dened feature LQUE, a
feature that is introduced by a second abstract complementizer in the presence
of the full wh expression and subsequently passed up the tree. Each form of
feature percolation is subject to discrete forms of island, with additional locality
23 A pronominal copy cannot provide any basis for predicting the sharp fall-o in accept-
ability if the number of occurrences exceeds two, an asymmetry with was which is directly
contra-indicated by a pronominal account of these additional non-was expletive forms.
(i) ?*Wer glaubst du, wer Hans meint, wer Uli half?
`Who do you believe Hans said helped Uli?'
(ii) Was glaubst du, was Hans meint, was Jakob sagt, wen Uli gesehen hat?
`Who do you think Hans said Jacob said Uli has seen?'
(i) is markedly worse than (ii), notwithstanding a form of wh, which, on a pronominal copy
analysis, should make the sequence as easy to process as, say, (iii):
(iii) Jeder student weiss dass er seine Hausaufgaben zu seinem Lehrer schicken hatte sollen
wenn er sein Ergebnis rechtzeitig erhalten wollte.
Èach student knows that he should have sent his assignments to his teacher if he
wanted to know his results in time.'
features specic to each construction to induce whatever restrictions are neces-
sary (island restrictedness in the rst, not so restricted in the second, possibly
locally restricted to a single clausal boundary in the third). The particular as-
sociation of type of restriction to each feature is coincidental. A SLASH feature
insensitive to islands and a wh -in-situ expression sensitive to such islands could
equally be dened. Where the expletive form itself displays case, as in Hungar-
ian, a discrete analysis would be required in order to re ect the case feature,
a prediction not made by the LQUE analysis. So the account succeeds, but it
does little more than mirror the divergence of the data, with no explanation of
the data in terms of independently motivated general principles.24
5.4.2 Partial Movement as a Re ex of a Requirement
In contrast to these exclusively conguration-based approaches, the dynamics of
the present framework might have been devised with the expletive wh phenom-
ena in mind; for the shift from an imposed requirement to a resulting annotation
{ in combination with the concept of nodes in a tree being describable as unxed
{ is, we suggest, what is needed to capture the underlying uniformity in these
data.
We start from the simplest case { a regular wh expression annotating directly
a xed position with Fo(WH). If we were to dene a specic lexical item to
ensure the subsequent presence in the string of such an expression, and in this
sense anticipate it, this item would have to project a requirement:
? # (F o(WH ))
To ensure that such an expression would annotate a node properly dominated
by the node under construction, the requirement would have to take the form:
?h#i # (F o(WH ))
This gives us the means of characterizing an expletive for a wh-in-situ form, as
we shall see for Iraqi Arabic.
Supposing, however, we take the set of annotations projected by a clause-
initial wh expression. In wh-initial questions in English and in German such
expressions simultaneously annotate two nodes: the node requiring type t which
it annotates with the feature Q, and an unxed node, which it annotates with
Fo (WH). Given that any DU formula may be imposed on a node as a require-
ment, the eect of this set of actions can also be imposed as a requirement.
We will construct this requirement in stages. Seen from the perspective of some
node requiring ?T y(t), an unfixed node annotated with F o() can be described
as h#i(F o()) (note the use of the internal operator, as opposed to the external
operator used in the requirement for the in-situ expletive). However, wh-initial
expressions in questions also add the annotation Q to the node from which they
24 So far as we are aware, there is no categorial grammar account of partial movement
phenomena.
dangle as unxed. At this node, the eect of this pair of actions can be described
as:
Cat(Q) ^ h# i(F o(WH ))
As in the simpler wh-in-situ case, we have to ensure that this requirement can
be fullled only by wh expressions annotating a discrete, dominated node.
Thus, we need a requirement of the form:
?h#1i # (Cat(Q) ^ h#i(F o(WH )))
This gives us the requirement needed to characterize an expletive which is antici-
patory of a clause-initial wh expression, as is needed for German and Hungarian.
This requirement is satised only in the presence of two nodes (apart from
the one it decorates): the unxed node and a node on the branch leading to
that unxed node. In principle, either of the two nodes to be introduced may
be subject to further restrictions { by providing a more precise description of
their relative position in the tree (case marking) { or in terms of their relative
distances. This gives us directly the span of variation we nd in expletive wh
forms: German optionally case-marks the unxed node, indicating its relative
position in the resulting tree, with variation over how local the intermediate
node needs to be to the node decorated by the requirement. Hungarian case-
marks the intermediate node.
5.4.2.1 German wh Expletives
German, like English, denes its full wh expression, if initial, as not only
annotating an unxed node with Fo(WH), but also annotating the dominating
node requiring T y(t) with the additional formula Cat(Q). Accordingly,
the expletive was form is dened to impose the composite requirement
Cat(Q) ^ h# i(F o(W h) ^ T y(e)) on some dominated node. However, in or-
der to dene the update provided by the expletive form, we need to bear in
mind appropriate word order restrictions, as it is the lexicon which has to de-
termine obligatory aspects of word order. So before turning to this task, we
provide a sketch of German word order and the general assumptions we shall
make in addressing the problems raised.
5.4.2.1.1 German word order German presents notorious problems of
word order. Primary amongst these is the verb-second phenomenon: in root
clauses German allows almost any constituent to be initial, but following this
constituent, the verb must occur on its own, leaving the subject and any other
noun phrases to follow (in the so-called Mittelfeld), in which they occur in the
canonical order subject/indirect object/direct object:25
25 There are minor exceptions to this verb-second restriction: e.g. topic constructions, for
which an analysis of the rst and second constituents as the root node of a LINK ed structure
and an unxed node respectively seems promising:
(5.69) Den Marie liebt der Jakob. [German]
MarieACC loves JacobNOM
`Jacob loves Marie.'
(5.70) Der Jakob liebt den Marie.
JacobNOM loves MarieACC
`Jacob loves Marie.'
To add to the complications, if such a verb form is compound, as in the perfect,
then it is the auxiliary which occurs in the second position in root clauses, and
in such a case the verb must occur at the end, following the sequence of noun
phrases in the Mittelfeld:
(5.71) Heute hat der Jakob den Marie gesehen.
Today has JacobNOM MarieACC seen
`Today Jacob has seen Marie.'
(5.72) Den Marie hat der Jakob heute gesehen.
MarieACC has JacobNOM today seen
`Marie, Jacob has seen today.'
(5.73) *Der Jakob hat gesehen den Marie heute.
JacobNOM has seen MarieACC today
`Jacob has seen Marie today.'
(5.74) *Heute, der Jakob hat gesehen den Marie.
Today JacobNOM has seen MarieACC
`Today, Jacob has seen Marie.'
This nal position in a clausal sequence is the required position for verbs in
subordinate clauses, unless the verb takes a complement clause, in which case
in both root and subordinate clauses the complement clause must follow the
verb:26
(5.75) weil der Jakob den Marie liebt
because JacobNOM MarieACC loves
`because Jacob loves Marie'
(5.76) Hans sagt, dass er glaubt, dass der Jakob den Marie liebt.
Hans said that he thinks that JacobNOM MarieACC loves
`Hans said that he thinks that Jacob loves Marie.'
(i) Den Jakob, den liebt die Marie.
JacobDAT , thatDAT loves MarieNOM
`Jacob, Marie loves.'
Dialects of German vary as to whether case is obligatory with names. Here we give examples
with case marking on all names, though case marking of names, in particular when nominative,
is often avoided.
26 Again with minor exceptions, in which a subordinate clause may be taken to mimic the
root clause ordering.
In wh questions, the initial position in root clauses is always taken by the wh
expression (which is, as in English, its canonical position), and in subordinate
clauses, the wh expression must, as in English, occur rst in the sequence of
noun phrases:
(5.77) Wen liebt der Jakob?

WhoACC loves JacobNOM
`Who does Jacob love?'
(5.78) Heute fragt der Hans, wen der Jakob liebt.
Today asked HansNOM whoACC JacobNOM loves
`Today Hans asked who Jacob loves.'
(5.79) *Heute fragt der Hans der Jakob wen liebt.
Today asked HansNOM JacobNOM whom loved
`Today Hans asked Jacob whom loved.'
As a nal twist, in the expletive wh questions, the expletive wh form, despite

its apparently empty semantic contribution, nevertheless counts as an initial
constituent, and must be immediately followed by the verb (or the auxiliary if
there is one) in root clauses; so any subsequent wh expression, which obligatorily
occurs with such an expletive, cannot occur in that position, and moreover does
not occur immediately following it but at the left periphery of some embedded
clause:
(5.80) *Was wer glaubt, dass der Jakob den Marie liebt?
What who thinks that JacobNOM MarieACC loves
`Who thinks that Jacob loves Marie?'
In trying to make sense of these data, we shall assume as a general property

of German that the rst noun phrase in any root clause in German annotates
an unxed node. This node is induced by a step of *Adjunction, as is famil-
iar.27 We shall then assume that the presence of the verb second in any root
clausal sequence is ensured by a language-specic rule of Verb Adjunction which
is dened to license an unxed node most-locally, i.e. within the limits of a single
27 What can occur in such a left-dislocated position is somewhat more liberal than in English,
but we leave this on one side.
clause, which is actually a schema for a number of rules, one for each dierent
type of verb:28
Verb-Adjunction (X)
fT n(0); ?T y(t); }g; fh"i(T n(0)); F o(); T y(Y ); ?9xT n(x)g

fT n(0); ?T y(t)g; fh"i(T n(0)); F o(); T y(Y ); ?9xT n(x)g,
fh"1 i(T n(0)); ?T y(X ! (e ! t)); }g
Recall, that there are two type specications involved in the lexical entry for
a word such as a verb. There is the trigger condition (ÌF') for the sequence
of actions to take place, which characteristically is a type specication; and
there is the type specication which is provided as one of the annotations to
be added. In chapter 3 a sample lexical entry for a verb was given for English,
which introduced a sequence of actions relative to the condition requiring type
e ! t, with these actions creating a node with a type specic to the predicate
provided by the verb, type e ! (e ! t). In English and Arabic the two types
involved in the lexical entry for a verb dier. However, in German, they will
be identical. German verbs may also vary according to whether they impose a
requirement for a sister node of type t, or whether they require a sister node
annotated by T y(e). For illustration, we dene glaubt as follows:29
glaubt
IF f?T y(t ! (e ! t))g
THEN IF f" >g
THEN put(F o(Glaub); T y (t ! (e ! t)); go(h"1 i);
put(?h#0 iT y (t))
ELSE put(F o(Glaub); T y (t ! (e ! t)));
go(h"1 i); put(?h#0 iT y (e))
ELSE ABORT
Notice that, after glaubt has been processed, in root clauses the creation of the
28 This characterization might appear to leave on one side all the discontinuous verb forms
with auxiliary plus participle in so far as it makes no special reference to them. One way to
address this problem, within this current analysis, is to analyse auxiliaries as expletive verbs
in the sense that they decorate a node introduced by the general rule of Verb Adjunction with
a requirement for a verb of a certain type, but then leave it without providing the required
annotations, so that an unxed node requiring the type of a verb is left to be subject to a
later step of Merge when the verb duly annotates a xed node that has been introduced by
successive application of Introduction and Prediction. The denition of Verb Adjunction (X) is
by no means problem-free, presuming as it does a generalization across verbs of dierent logical
type; but for present purposes this will suÆce, and in all that follows we shall ignore problems
particular to such discontinuous tense projections: see Marten 1999 for a characterization of
all verbs as systematically underspecied as to type, characterized as e ! (e ! t), with
concomitant introduction into the tree description of the required formula at an unxed node,
allowing a more elegant denition of the rule of Verb Adjunction.
29 We are abstracting away here from any properties of agreement or tense. Tense, as
indicated in chapters 2{3, would be imposed as an action annotating the rst node dominating
the current node requiring type t. Agreement would be specied through dening syntactic
features which would be projected by one node as a requirement upon the node with which
it is paired as an annotation.
subject node is required and in non-root clauses the creation of the verb com-
plement.
Adopting these assumptions about German provisionally, we can now give
the full lexical characterization of the expletive was:
wasexpl
IF f?T y(t)g
THEN IF f# 9xF o(x)g
THEN ABORT
ELSE IF f" >g
THEN put(Cat(Q),
?h#1i # (Cat(Q) ^ 9xT en(x) ^ h#iF o(WH))
make(h#0 i); go(h#0 i)
ELSE put(Cat(Q),
?h#1i #1(Cat(Q) 1^ 9xT en(x) ^ h#iF o(WH)))
make(h# i); go(h# i); put(?T y (t ! (e ! t)))
ELSE ABORT
From this complex we can now extract what are properties of the expletive
which are general to an expletive for a wh-initial expression, and what are par-
ticular to German. The rst general property is the condition that the node
to be decorated should as yet dominate no nodes annotated by formulae { i.e.
an expletive precedes all words within a clausal sequence. The second general
property is the annotation of the current node with a syntactic feature Q. The
third is the30complex requirement for a wh-initial expression, itself projecting the
feature Q. All other complexities are specic to the problems of German word
order. The rst sequence of actions is dened for the subordinate environment
(`something is above the current node'), which includes the construction of a
subject node. The alternative sequence of actions is dened for the root clause
environment, and includes a requirement that ensures that a nite verb imme-
diately follows. Here we cannot rely on the general process of Verb Adjunction
to guarantee the verb-second position, for the expletive does not annotate an
unxed node, and this is what Verb Adjunction requires. In consequence, as
part of the expletive, we have to dene actions that ensure that in a root clause
a verb (or auxiliary { see n.6) will follow. These two alternatives are displayed
in gure 5.2.31
Focusing now solely on the complex requirement denitive of the expletive
itself, notice that there is no type or case specication for the required wh
expression, thus licensing the presence of a WH formula of whatever type or
case. Once an unxed node has been introduced in the tree, a wh expression
can both annotate that node with Fo(WH) and the dominating node with the
30 The specication of 9xT en(x) as part of the complex requirement is included to ensure
the tense marking on the subordinate clause. Whether this restriction is in fact necessary
turns on an analysis of non-nite constructions. If we analyse these as projecting a predicate
of type e ! t directly, with no immediately dominating node requiring T y(t), this specication
is unnecessary.
31 In this tree display, the annotations given are in the form of tree descriptions.
I The expletive initial in a subordinate clause:
fT n(a); ?T y(t)g
fT n(b); ?T y(t); }g
fT n(a); ?T y(t)g
fT n(b); ?T y(t); Cat(Q); ?h#1 i # (Cat(Q) ^ 9xT en(x) ^ h# iF o(WH))g
fh"0 iT n(b); ?T y(e); }g

II The expletive initial in a main clause:
fT n(a); ?T y(t); Cat(Q); ?h#1 i # (Cat(Q) ^ 9xT en(x) ^ h# iF o(WH))g
fh" iT n(a); ?T y(t ! (e ! t)); }g
Figure 5.2: The eect of wasexpl in German
feature Q, and from then on, the construction mirrors the one for wh-initial
structures. The nal result is that two strings diering only with respect to
the presence of an expletive plus an intermediate clause-initial full form versus
full form in initial position will yield the same tree structures. Both (5.81) and
(5.82),
(5.81) Wer glaubst du, dass Uli half?
Who think you that Uli helped
(5.82) Was glaubst du, wer Uli half?
What think you who Uli helped
will yield the same annotation of the root node:
Cat(Q); F o(Glaub(Half(Uli)(WH))(du)); T y(t)
Either one of the two nodes required by the expletive might be additionally
restricted. For example, to the requirement of an unxed node annotated with
Fo(WH), one might add a case specication constraining its eventual position
in the tree. This yields the additional type of German expletive, the lexical
variants of wer, wen, etc. (repeated here from section 5.4.1):32
(5.58) Wer glaubst du, wer Uli half?

who think you who Uli helped
(5.59) Wen meint Karl, wen Uli half?
who said Karl whom Uli helped
`Who does Karl say Uli helped?'
(5.60) Wie nimmt man, an wie der Prozess ausgeht?
how assumes one how the trial ends
`How does one assume that the trial ends?'
These cases, licensed in some dialects of German, also illustrate the second basis
for variation outlined earlier, that of constraining the tree relation between the
node decorated by the expletive and the one annotated by Fo(WH). In these
forms, the relation between the node decorated by the expletive requirement
and the dominated node with the feature Q is restricted by the fact that there
may be only one intervening type t node. This more restricted relation, together
with some case specication such as subject, can be expressed as:
?h#1i #1#0 (Cat(Q) ^ 9xT en(x) ^ h#i(F o(WH) ^ "0 (T y(t))))
The case-marked expletives constrain their tree positions. We give here the
lexical characterization of expletive wer:33
32 In movement analyses of expletive wh phenomena in German, these are referred to in
the literature as an entirely separate `copy-wh strategy', on that analysis the broad range of
properties (see Hohle 1996, Reis 1996) shared by the two processes being coincidental.
33 Here we are systematically ignoring person considerations, though their specication
might be incorporated by a restricted indexing system. See the lexical account of the denite
article and the encoding of its restrictor properties suggested in chapter 7.
werexpl
IF f?T y(t)g
THEN ABORT
ELSE IF f" >g
THEN put(Cat(Q),
?h#1i #1#0 (Cat(Q) ^ 9xT en(x)^
h# i(F o(WH) ^ h"0 i(T y(t))));
make(h#0 i); go(h#0 i)
ELSE put(Cat(Q),
?h#1i #1#0 (Cat(Q) ^ 9xT en(x)^
h# i(F o(WH) ^ h"0 i(T y(t)))));
make(h#1 i); go(h#1 i); put(?T y (t ! (e ! t)))
ELSE ABORT
Notice how this form, like the expletive forms of wen and wem, display system-
atic variation from the lexically neutral was on two counts: case and locality,
and also in fact person marking, which we have ignored here.
By way of comparison we give in gure 5.3 schematic output descriptions
for (5.81), (5.82) and (5.58), each of which gives rise to the same formula. Note
that the rst characterization involves two nodes, but the two following only
one; the third imposes a case requirement at some unxed node annotated by
Fo(WH), and a locality requirement on the intermediate node.
All these German wh expletives share overall properties. On the one hand,
the Q feature which they impose as a requirement on a dominated node will not
clash with the semantic property of being the object of a propositional attitude
(which is required of complements to predicates such as glaub ), since Q is indica-
tive of the structural property of a formula containing the meta-variable WH, not
of any general property of questionhood. The signicance of this is that, on the
one hand, there is no con ict with whatever semantic restrictions are imposed
by intervening verbs such as glaub, which may not select for questions.34 On the
other hand, there is no problem in having a tree structure of which several nodes
are annotated with the feature Q, say re ecting the formulae Liebt(WH)(Jakob)
at one node, the formula Meint(Liebt(WH)(Jakob))(Hans) at some higher node,
and the formula Glaubt(Meint(Liebt(WH)(Jakob))(Hans))(du) at some node
which dominates both, all in re ection of a sequence of expletives followed by a
single wh expression, each of which annotates a node with the feature Q:
(5.52) Was glaubst du, was Hans meint, wen Jakob liebt?
34 Unlike wonder in English, glaub imposes no requirement other than ?T y(t) on its induced
complement argument node, though nothing precludes its acquiring a feature such as Q.
fCat(Q); F o(Glaubt(Meint(Liebt(WH)(Jakob))(Hans))(du)g
...fCat(Q); F o(Meint(Liebt(WH)(Jakob))(Hans))g
...fCat(Q); F o(Liebt(WH)(Jakob))g
Result of processing wer-initial non-expletive in (5.81)

fT n(a); Cat(Q); ?T y(t)g
j
fh"iT n(a); F o(WH); T y(e); ?h"0i(T y(t)); }g
Result of processing was-expletive in (5.82)
fT n(a); Cat(Q); ?T y(t); ?h#1i # (Cat(Q) ^ 9xT en(x) ^
h# i(F o(WH))); }g
Result of processing wer-expletive in (5.58):
fT n(a); Cat(Q); ?T y(t); ?h#1i #1#0 (Cat(Q) ^ 9xT en(x) ^
h# i(F o(WH); h"0 i(T y(t)))); }g
Figure 5.3: Updates from processing wh
While, in ignoring all the complexities of agreement and tense, this sketch
of wh expletives in German is woefully incomplete, it none-the-less gives an
indication of how the present formalism can be used to capture a broad array of
German partial movement data in a unifying way. It captures the central data
(5.52){(5.53) in which an initial was precedes a later full wh expression, itself
possibly to be followed by a complementizer dass intervening between full form
and the `gap':
(5.53) Was glaubst du, wen Hans meint, dass Jakob half?
It also requires that a subject-construed noun phrase follow the expletive in
subordinate clauses:35
35 This restriction will preclude instances in which the verb occurs immediately after the
expletive in a subordinate clause, mimicking the root-clause restriction:
(i) Was glaubst du, was meint Hans heute, wen Johannes gesehen hat?
`Who do you think Hans is saying today that Johannes saw?'
These we would have to analyse as involving apposition of two question structures { in eect
the indirect dependency form of analysis (see Gartner 1997 for discussion of non-root verb
inversion eects, particularly in relative clauses).
(5.83) Was glaubst du, was Hans meint, wen Johannes liebt?
what think you what Hans said whom Johannes loves
`Who do you think Hans is saying that Johannes loves?'
However, what is precluded is the projection of an unxed node requiring type
e immediately following the expletive, because the expletive as dened must be
immediately followed by a verbal form in all root clauses:
(5.84) *Was wen hat Jakob gesehen?
what whom has Jacob seen
`Who has Jacob seen?'
Finally, in all occurrences of the expletive as dened, the subsequent apparently
partially moved wh expression cannot precede a non-nite verb, in virtue of the
requirement ?9xT en(x).
The denition provided correctly characterizes a further range of data.36
First, it precludes (5.85), in virtue of specifying that some dominated node
must be annotated by a full wh expression:37
(5.85) *Was glaubst er was?
what thinks he what
`What does he believe?'
Since, as a clause-typing feature, a Q feature is consistent only with the type
annotation T y(t), it will never hold of some node internal to a single clause.
Moreover, the second occurrence of was here could not be a wh-in-situ form,
since, as in English, this would not project the Q feature required by the exple-
tive. The analysis also, for the same reason, precludes a clause-internal occur-
rence of the expletive was, a fact extremely puzzling on an indirect dependency
analysis:
(5.86) *Wer meint was, wen wir gewahlt haben?
Who thinks what whom we elected have?
`Who thinks we have elected who?'
Unlike the indirect dependency analysis, the account here explains the lack of
any construal of was in (5.86), either as an in-situ form or as an expletive. The
was, if it were an expletive, could only annotate some induced complement node
requiring type t as Q, but in so doing it would preclude the presence of wen in
(5.86), for the same reason as (5.84). And if the was were a full form satisfying
the second argument of meint with that verb lexically dened to project a
36 We do not provide any characterization of the negative island data discussed in Beck and
Berman 1996, Fanselow and Mahajan 1996, since no account of negation is as yet provided in
this framework.
37 These data are on the face of it troublesome for Simpson (2000), for the wh expletive
projects the feature which licenses the presence of the full form of wh expression, and move-
ment is solely a means of disambiguating the trigger (see Simpson 2000 for discussion and also
section 5.3.1 n.9 above.
predicate of type e ! (e ! t), then the following complement clause sequence
is not licensed at all.
We also correctly characterize (5.56) as ill-formed on the basis that the
requirement projected by the expletive form fails to be satised:
(5.56) *Wen glaubst du, was Hans meint, dass Jakob half?
Whom believe you what Hans said that Jacob helped
Here the characterization of the requirement specic to the expletive as having
to be satised at a daughter node is essential. If the requirement added to the
tree description by the expletive were of the form
? # (Cat(Q) ^ 9xT en(x) ^ h#i(WH ));
then the Q specication lexically projected as part of this requirement could
in (5.56) be satised at the node meeting the condition for the actions asso-
ciated with the expletive, and the unxed node that was annotated through
processing wen, which as yet remains unxed, could satisfy the second part of
the requirement requiring an unxed node.
W h expletives may also occur heading a subordinate clause (without having
to construe that intermediate clause as an independent question):
(5.87) Wer fragt sich, was sie meint, wen Hans liebt?
who asks herself what she thinks whom Hans loves
`Who wonders who she thinks Hans loves?'
We would anticipate that this structure, with the expletive inducing some lower
wh formula as a requirement, is not construed as a multiple wh question { which
is correct (see Fanselow and Mahajan 1996). Nothing, also, determines that the
requirement of there being a full wh expression be satised locally, so though
no replication of such a requirement is precluded, nothing dictates it either, and
(5.54) is, correctly, characterized as well-formed:
(5.54) Was glaubst du, dass Hans meint, wen Jakob half?
What think you that Hans said whom Jacob helped
This is contrary to any indirect dependency analysis in which all sequences of
clauses must be questions in their own right.38 In accordance with those anal-
yses, (5.54) is often reported in the literature to be ungrammatical at least for
some speakers (Beck and Berman 1996); yet our informants almost invariably
judged it to be acceptable. This variation between speakers is notably straight-
forward to express in our system, given the exibility provided by requirements
38 Simpson (2000) also has to characterize these forms as involving PF deletion of an inter-
mediate occurrence of was, though in virtue not of clausal typing, but of an imposed locality
constraint. This is not so much was `deletion' as `substitution' of was by dass.
on further development of a node, which may or may not dictate locally re-
stricted satisfaction of that requirement. It is simply that in the variant of
German in which (5.54) is unacceptable, a more stringent locality specication
imposes the requirement of the annotation Q on the daughter of the predicate
node, like the more restrictive non-was forms of expletive (see examples (5.58){
(5.60)). According to this model, the analysis re ects the data as two variants
of the language which have the same vocabulary and the same logical forms,
but nevertheless license distinct string sets.
A further advantage, which is particular to the present account, is that
nothing precludes an additional in-situ wh expression occurring between the ex-
pletive form and some lower clause-initial wh expression, allowing a construal as
a multiple question (see Hohle 1996, Fanselow and Mahajan 1996, Reis 1996):
(5.88) Was meint wer, wen wir gewahlt haben?
What thinks who whom we elected have
`Who thinks we have elected whom?'
(5.89) Es is mir vollig egal, was wer meint, wen sie
it is to me completely equal what who thinks whom she
liebt.
loves
Ì do not care who thinks she loves who.'
In all these cases, the expletive serves to provide an annotation for the top node
of the feature Q. It also has the eect of guaranteeing the presence subsequently
in a grammatical string of a wh-initial form projecting the composite annota-
tion of a Q feature plus that WH variable. Nothing, however, precludes an
intervening in-situ wh form; and with both root and complement node having
a Q feature, nothing precludes this form from being the rst of a sequence of
WH variables, even though it does not satisfy the requirement of the expletive.
So, unlike39other analyses, here such examples can be characterized as multiple
questions.
So, to sum up, characterizing the expletive forms through the concept of
requirement as anticipatory of a full wh form, with both non-case-marked and
case-marked forms, provides an elegant and relatively successful characterization
of a broad array of data.
5.4.2.2 Hungarian wh Expletives
As pointed out initially, with the more complex expletive requirement neces-
sitating the introduction of two nodes, the node which is initially unxed and
39 The tension, leading to reduced acceptability for some speakers (observed by Pafel 1996),
is that the wh expression which enables the expletive's requirement of a subsequent full wh
form to be satised follows the projection of the in-situ wer in (5.88), so any expectation by
some human parser that the interpretation of the question will involve the partially moved
form as projecting the rst open variable would not be fullled. Judgements are for some
reason clearer with the embedded question form, as in (5.89).
an intervening node which is associated with the Cat(Q) annotation, either of
these two nodes may be subject to further restrictions. Hungarian provides a
language in which the morphological case on an expletive wh form projects a
requirement on this intermediate node, the complement of the predicate verb
of the clause containing the expletive. In most instances the case selected in
Hungarian is the accusative-marked form mit, but individual predicates may
impose idiosyncratic restrictions, as in (5.91):40
(5.90) Mit mondtal ok mit tudnak melyikk
WhatACC saidindef:2sg they whatACC knowindef:3pl which
ut szereted ek ? [Hungarian]
boyACC likedef:2sg
`Which boy did you say they know you like?'
(5.91) Mire szamitasz, hogy mit fognak mondani a
WhatAL count2sg that whatACC the kids will
gyerekek?
say
`What do you think the kids will say?'
We dene the expletive mit as follows:
mit (expletive)
IF f?T y(t)g
THEN ABORT
ELSE put(Cat(Q),
?h#1 i #1#0 (9xT en(x) ^ h"0iT y(e ! t) ^
# (Cat(Q) ^ h# iF o(WH ))))
make(h#1 i); go(h#1 i); put(?T y (t ! (e ! t)))
ELSE ABORT
That Hungarian should license an expletive of this more complex form is ex-
pected, for all Hungarian wh expressions characteristically occur initially in a
clausal sequence41 { hence the similarity to German in displaying partial move-
ment phenomena. In these expletive structures, the verb is marked with a
Def ({) agreement, distinguishing such cases from normal subordination mark-
ing on the verb, which is indicated by a Def (+).42 The characterization of
mitexpl , like the analogous denition for expletive was, will ensure that it
40 This preliminary account of Hungarian expletives is really no more than an indication of a

story yet to be told, particularly in view of the complexities of Hungarian word order. In what
follows we shall side-step all such issues, merely indicating the way in which the Hungarian
expletive diers from its German analogue.
41 In-situ forms are reported to invariably involve echo interpretations.
42 We do not formalize this here.
co-occurs with a subsequent full wh expression at some level of embedding within
a case-marked complement structure as in (5.90).43
Such expletive structures display only partial parallelism with movement
constraints (contra direct dependency analyses). Given that F o(WH) occurs
dominated by the feature Q, albeit initially at an unxed node, that property {
and indeed the requirement on the top node which is projected by the expletive
{ will only be satised if the intermediate node and the initially unxed node
nd positions in44the local tree, and not in some linked tree structure (i.e. in a
relative clause):
(5.92) *Mit hallottal a hirt hogy kivel
WhatACC heard2sg:indef:DO the newsACC that with whom
talalkozott Mari?
met3sg MaryNOM ?
`What did you hear the news with whom Mary met?'
Yet the requirement of a local complement with a tense annotation correctly
precludes any wh form internal to the same clause as the expletive, despite
the fact that, in these languages, the resolution of an initially unxed node
annotated with F o(WH) is licensed in such structures:
(5.93) Kinek lattad a feleseget?
WhoDAT saw2sg:def:DO the wife3sg:poss:ACC
`Whose wife did you see?'
(5.94) *Mit lattal kinek a feleseget?
WhatACC saw2sg:indef:DO whoDAT the wife3sg:poss:ACC
`What did you see whose wife?'
The proposed characterization of the Hungarian expletives is thus consonant
with an array of data, much of which is problematic, in particular for direct
dependency analyses.
There are, in addition, a range of other case-varying expletives; and, as in
the German case, all these require individual lexical denition re ecting dierent
case requirements, and, as in German, allowing speaker variation as to expletive
forms. The signicance of the proposed analysis of wh expletives is that in this
framework both German and Hungarian can be seen as minor variants of the
same basic pattern, that expletives are forms which are anticipatory of regular
full wh forms.
43 (5.90) is the data taken by Horvath as evidence that the expletive form had moved and was
not an expletive, here characterizable without invoking a non-expletive form of analysis. These
are also the data problematic for Simpson's assumption that the intermediate full clause-initial
wh form is licensed only in virtue of a local relationship to some higher expletive.
44 The one case which appears to be problematic is the apparent license of wh expressions
separated from an expletive wh by a clausal adjunct boundary. In such a case, we would
dene the expletive miert (= English why ) form as inducing a pair of conjuncts, annotating
the second with a feature Q.
5.4.2.3 Iraqi Arabic wh Expletives
Finally, we have the expletive form which is anticipatory of an in-situ form, and
here, recall, the WH variable is not required to occur at an unxed node, and the
expletive is characterized by means of the simpler requirement ? # F o(WH ).
Iraqi Arabic, however, requires a complication to this minimal specication, for
the full wh forms are licensed only in the presence of some most-local Q feature.
Accordingly, any wh expletive, if anticipatory of the in-situ form, will have itself
to project the feature Q as an annotation, so that subsequent presence of the
full wh form can be suitably licensed:
(5.95) sh-'tsawwarit Mona Ali raah weyn? [Iraqi Arabic]
Q thought Mona Ali went where
`Where did Mona think that Ali went?'
(5.96) *sh-i'tiqdit Mona meno tsawwarit Ali sa'ad meno?
Q believed Mona who thought Ali helped who
`Whoi did Mona believe ei thought Ali helped who?'
With this in mind, we dene the complex of morphological prex plus verb
as projecting the feature Q at the top node and at the node projected as the
complement of the verb, as well as constructing and annotating the node for
the predicate itself. Because this expletive form is a verbal prex, we dene
the projection lexically, each propositional attitude verb having an alternative
sh-marked form:
sh-'tsawwarit
IF f?T y(t)g
THEN ABORT
ELSE put(Cat(Q));
make(h#1 i); go(h#1 i);
put(F o(T sawwarit); T y (t ! (e ! t)); go(h"1 i);
make(h#0 i);go(#0 ); put(Cat(Q); ? #1 #0 (F o(WH)));
gofirst(?T y (t)); make(h#0 i); go(h#0 i);
put(F o(U); T y (e); ?9x(F o(x)))
ELSE ABORT
As in the German case, we have to disentangle what is pertinent to the expletive
itself and what is a re ection of Arabic word order. Recall from chapter 4 that
Arabic is a free subject-pro-drop language, with the verb taking a node requiring
T y(t) as the condition for its sequence of actions. The eect of the prexed verb
is then to ensure that all nodes for the arguments to the predicate are induced,
as well as a Q feature on both top and complement nodes.45 There are two
45 As it stands, this characterization will preclude the occurrence of an explicit subject noun
phrase following the verb in expletive structures, as in (5.95). We leave this here, as one of a
number of word-order problems.
critical properties of this expletive. First, there is the projection of the feature
Q at the top node of the complement structure as an annotation, rather than
a requirement, since this is necessary to satisfy the locality restriction on the
full form of the wh expression if the lower clause is tensed. Secondly, there is
the requirement itself, which ensures an occurrence of F o(WH) at some locally
dominated node. Notice that, unlike German, the requirement for a WH meta-
variable can be met directly by some action annotating a xed node, hence the
modal requirement using the local variant of the external # form of operator.
Here is where the distinct classication provided by the Q feature for Arabic is
doing some work, because it is not specic to wh-initial questions as in German
(or English and Hungarian).
The wh expression may occur initially in the subordinate clause where it
too would project the label Q, but this position in the sequence is not required.
Hence partial movement in Iraqi Arabic is, unlike in German, optional:
(5.97) sh-tsawwarit Mona Ali gabal meno?
Q thought Mona Ali met who
`Who did Mona think Ali met?'
(5.98) sh-tsawwarit Mona meno Ali gabal?
Q thought Mona where Ai went
`Who did Mona think Ali went?'
Finally, given the reported optionality of the expletive form when the comple-
ment clause is non-nite (Wahba 1991),
(5.99) (sh)-raadat Mona Ali ygaabal meno?
Q wanted Mona Ali meetINF whom
`Who did Mona want Ali to meet?'
we specify the decoration on the complement node as simply the annotation
Cat(Q), leaving open whether or not the node should also be decorated with
?T y(t) and ?9xT en(x) as in German and Hungarian.46
There is much more that can be said about these structures in the dierent
languages, and we freely grant that these sketches are no more than promis-
sory notes for fragments of the languages yet to be given, and much remains
poorly understood. In particular, we have made use of syntactic features which
are, arguably, nothing more than a shorthand for talking about the dierent
structural classes that need to be identied in the language. None-the-less we
take these characterizations to indicate that the various types of wh expletives
can be analysed in ways that directly re ect the informal observation that they
serve to anticipate the full form of wh expression within the string.
46 This optionality is problematic for both Simpson's (2000) Minimalist account, and John-
son and Lappin's (1999) HPSG account.
5.5 Wh Expressions and Scope Eects
In these outlined analyses of wh question structures, what we have relied on
are two basic properties of tree development. There is the development from
an unxed node to a xed node, and the development from requirements on
a node (or set of nodes) to annotations on a node or (set of nodes). It is
these two properties which have formed the basis for articulating the space for
possible variation. The success in providing a natural typology of observed cross-
language variation suggests that the intrinsic underspecication to all analyses
vis a vis semantic evaluation is the clue to addressing the semantic problems
raised at the outset.
One striking dierence between this approach to wh questions and all others
is that it completely lacks the concept of the wh expression corresponding to an
operator binding some discrete position/variable.47 The wh expression projects
a meta-variable without denotational properties. It is this assumption which
forms the common element in the dierent analyses of wh-initial/wh-in-situ/wh-
expletive structures, while allowing diversity in the various expletive and non-
expletive forms. The extent to which the construal of wh expressions interacts
with quantifying expressions as part of the interpretation of the question will
thus have to follow from something other than intrinsic properties of the wh
expression itself.48
This negative conclusion about the correlation between wh expressions and
scopal properties of quantifying expressions follows from the commitment to
characterizing the question as a question, rather than in terms of possible an-
swers. In this connection, there are two rather dierent concepts, both of which
47 The sense in which wh questions fail to display a semantic scope phenomenon was rst,
to our knowledge, argued by Reis and Rosengren (1992), but they retain a concept of what
they call interrogative scope, distinguishing it from the semantic concept of dependence of one
quantifying expression upon another. See also Pafel 1996, which argues that the expletive was
is not a scope marker.
48 Hirschbuhler (1981) argued for a wh -scope ambiguity following Baker (1970), on the
grounds of unacceptability of VP ellipsis with multiple wh questions as antecedent:
(i) *Which professor knows which boy subscribed to which magazine and which girl did?
(ii) *Which boy subscribed to which magazine and which girl did?
and the unambiguity of
(iii) Which professor knows which boy subscribed to which magazine and which girl does?
While, with no formal account of the various opacity eects induced by question-embedding
predicates, we have no account of (iii) to oer, we nevertheless have a natural explanation for
(i) and (ii) as a starting point for such an account: namely, that VP ellipsis is an anaphoric
process from an input with a meta-variable of type e ! t requiring substitution by some ap-
propriate term of the same type, hence associated with a requirement ?(9xF o(x); T y(e ! t)),
a requirement which precludes meta-variables as values. The acceptability of (iii) thus has to
turn on the fact that the antecedent which provides input to the reconstruction of the ellipti-
cal form is interpreted as `knows the answer to the question Which teacher interviewed which
student?', however that is to be formalized. It should be noted in this connection that bare
argument ellipsis, equally a form of predicate ellipsis, appears to allow the very interpretation
of the full form which is precluded in (i):
(iv) Which examiner knows which linguistics student answered which question. And which
anthropology student?
Wh Expressions and Scope Eects 187
have been taken to fall under some broad concept of scope, which need to be
distinguished. The rst is the familiar one, which concerns the interpretation
of one (quantied) term relative to another. Relative to this concept, had we
adopted the perspective of analysing a question in terms of the answers provided
in some larger discourse sequence of which the wh question is a part { as a means
of characterizing the semantics of the question itself { the scope potential of the
WH meta-variable would have looked very dierent. In such a larger discourse
setting, in which the variable is provided with some value, these substituends
would interact with other quantifying terms in the projected structure, for the
substituend would be interpreted relative to the total structure provided by
the question{answer pair together, hence to other quantifying elements in that
structure (this type of example includes the functional questions distinguished
by Ginzburg (1992), following Engdahl (1986)):
(5.100) Who did everyone write to? Their mother.
It is the con ation of question and answer in providing a semantics of the ques-
tion tout court which gives rise to the stated interaction of scope between wh
expressions and quantifying expressions, not the construal of the wh expression
itself within the question.49
There is a rather dierent concept to which the term scope has also been
applied, and this is the relation between a WH variable and the formula within
which it occurs, a problem which is specic to the question form itself.50 What
a wh-initial expression projects into the tree description, in addition to some
meta-variable, is the annotation of the node in a tree structure by a set of
features { Q (the classier in questions of wh-initial questions) and QUE (the
classicatory device for questions as a whole: see notes 6 and 16 of this chapter).
This node-marking property is shared both by wh-initial forms and by exple-
tive wh forms, one of whose primary purposes is indeed to project such features.
This relationship has been used to articulate a concept of scope for a wh ex-
pression which is dierent from the concept of quantier dependency, being the
relation between a WH meta-variable and the propositional formula suitably
annotated with Q/QUE features within which it is contained. It is this distinc-
tion between the concept of quantier scope and the concept of propositional
formulae containing a meta-variable which explains the apparently con icting
scopal properties of wh expressions, set out as problematic in section 1. For
example, the con ict between the supposed wide scope properties of the wh
expression as indicated by the position of the expletive in some sequence of
words, and yet the availability of supposedly narrow scope interpretations for
49 It is argued in May 1995 and accepted by many others since that while (i) has an in-
terpretation in which who has narrow scope with respect to the quantifying expression every
overseas dignatory, (ii) does not:
(i) Who did every overseas dignatory meet when they arrived o the plane?
(ii) Who met every overseas dignatory when they arrived o the plane?
But in the light of the possibility of answers to (ii) taking the form of (iii), we do not take
this observation to be robust (see Ginzburg 1992 for further examples):
(iii) Sean met Clinton; Alison met Mandela; Sue met Blair.
50 This is Reis and Rosengren's (1992) concept of ìnterrogative scope'.
both so-called functional questions and expletive forms of question as in (5.10)
and (5.21) repeated here (observed in Pafel 1996), can be explained in terms of
this distinction:
(5.10) What is the Union insisting that almost every British farmer should get
rid of? At least 1,000 cattle.
(5.21) Was glaubt jeder wo sie gerne leben wurde? [German]
What thinks everyone who she live like would
`For each person x, for which place y, x think x would like to live in y?'
On the one hand, both these questions have an initial wh expression (in (5.21)
an expletive) which serves to annotate the root node with the feature Q. The
formula annotating this node will invariably contain a WH meta-variable, and
in this sense the wh expression might be said to be taking that formula as its in-
terrogative scope. On the other hand, the logical form constructed by this string
will contain both a WH meta-variable and a quantied term, and any choice of
answer to the question will lead to an interpretation in which the substituend
either does or does not take narrow scope with respect to other quantier(s) in
the string { hence the narrow scope interpretations for both (5.10) and (5.21),
despite the alleged scope-marking properties of the wh expression at the initial
position in the string.
We can also provide a basis for explaining the apparent ambiguity of wh
questions noted in section 1. If, as is traditional, wh questions are construed
in terms of the denotational content of their possible answers, ambiguity is
inevitable, for all scope choices have to be xed as a part of interpreting the
answer provided. Relative to this assumption, a question for which some answer
appears to dictate a particular scope relation (e.g. the functional interpretation)
will have to re ect that selected relation itself. But at the level of characterizing
the interpretation of the question itself, according to the present account, no
ambiguity has to be proposed, for the characterization of the wh expression is
neutral between any selection of scope.
Of the supposed variation in the interrogative scope of the wh expression
in embedded questions such as (5.101) according to whether they are construed
relative to the embedded question or to the higher question (see Baker 1970),
our analysis has nothing to say:
(5.101) Who remembered who reviewed which nineteenth century author?
In this connection, it should be noted that this phenomenon is sensitive to lexical
variation. Embedded questions functioning as complement to a verb such as ask
fail to give rise to any such ambiguity:
(5.102) Who asked who reviewed which nineteenth century author?
This dierence corresponds to the distinct ways in which these embedded ques-
tions are construed. As complement of verbs such as remember, the embedded
question has to be construed as providing the answer, with existential implica-
tion for the wh expressions. As complement of verbs such as ask, however, the
Wh Expressions and Scope Eects 189
embedded clause has to be construed as a question, and there are no existential
implications. As pointed out in detail by Ginzburg (1992) and Ginzburg and
Sag (1999), there is considerable heterogeneity in the interpretation of embed-
ded questions (Ginzburg and Sag isolate four classes of predicate), despite the
coherence of embedded wh questions from a structural perspective. This sug-
gests that all such variation in interpretation is best analysed as a consequence
of lexically dened idiosyncrasies. Given that we have had nothing to say in
this book about lexical inferences, we leave this on one side. Nevertheless, we
tentatively assume that this heterogeneity is suÆcient motivation for presum-
ing a lexically based analysis of this phenomenon, and retaining the assumption
that wh expressions project WH meta-variables without scope properties.
On the strength of the syntactic generalizations that form the bulk of this
chapter, then, we advocate an analysis in which the meta-variable WH which is
projected by a wh expression in questions, is a mere substitution site awaiting
a substituend. The dierent types of wh expressions, on this view, are a simple
consequence of variations in the dynamics of projection.
6
Crossover Phenomena
In exploring the dierent types of relative clauses in chapter 4, we had a rst
indication that anaphora resolution can play a central role in a structural con-
struction process, while retaining its essentially unrestricted, hence pragmatic,
nature. In this chapter, with the characterization of wh questions now also es-
tablished, we turn to so-called crossover phenomena, which have been treated
as constituting a heterogeneous set needing restrictions which are independent
both of anaphora construal and of long-distance dependency. Even a subset of
these phenomena has been described by Postal (1993) as a mystery, split into
an increasing number of discrete sub-phenomena. Together they constitute a
phenomenon in which anaphora and wh construal evidently interact; and the
challenge will be to provide a unifying analysis while nevertheless respecting the
diversity of possible interpretations.
6.1 Crossover { The Problem
In the 1970s, as a movement account of wh expressions was under development,
it was noted by Postal (1972) that a pronominal cannot be identied with a wh
expression (in (6.1)-(6.3) the relative pronouns) if it occurs between the position
in which that wh expression occurs and the position at which the wh expression
is to make its contribution (in the movement metaphor, the gap):1
(6.1) *Johni, whoi Sue thinks hei worries ei is sick, is on half pay.
(6.2) *Every man whoi Sue thinks hei worries ei is sick, is on half pay.
This phenomenon extends equally to wh questions:
(6.3) *Whoi does Sue think hei worries ei is sick?
1 In this chapter and elsewhere, we indicate this position by co-indexing of ei with the
appropriate expression only when either reporting on movement analyses or when it is essential
for clarity of the exegesis.
190
Crossover { The Problem 191
The rst problem posed by these data is that they seem to show that wh ex-
pressions cannot be analysed as regular quantier-variable binding expressions,
because, if they were, they would be expected to bind any or all variables within
their c-command domain. The solution, uncontroversially, is taken to be a dis-
crete denition of a syntactic operator. The second problem the crossover data
pose is that a subset of the data appears to be sensitive not merely to the rela-
tion between the wh element, the pronominal and the position of the so-called
gap, but to the larger context in which that structure is contained.2 Despite the
apparently identical congurational relationship between wh, pronominal and
gap in relatives and questions, the crossover eect appears not to be constant
across the two structures. In relatives, the restriction debarring an intervening
pronominal appears to hold with much reduced strength, if at all, if either the
wh or the pronominal expression is nested within a determiner:3
(6.4) The studenti whoi Sue thinks hisi mother worries ei is sick is on half
pay.
(6.5) The studenti whosei mother Sue knows hei worries ei is sick is on his
way to the hospital again.
(6.6) The book whichi I had to tell itsi author ei wasn't selling well was
taken o the market.
(6.7) Everyone whoi hisi mother had helped ei did well in the exam.
(6.8) The nurse whosei patientsj shei admitted ej were unhappy, was
transferred to another ward.
(6.9) Any student whosei abstractj the committee failed to tell themi ej
had been accepted, had assumed itj had been rejected.
This asymmetry between the pairing of a pronominal and a wh expression both
to be construed as arguments and the pairing of pronominal and wh when at
least one of them is not an argument does not apparently extend to questions,
where the unacceptability of an intervening pronominal appears to be sustained
for all sequences of wh ... pronoun ... gap:4
(6.10) ?*Whoi does Sue think hisi mother worries ei is sick?
2 The crossover restriction is said also to underlie the unacceptability of (i), a constraint
holding of movement at LF, but following Williams (1994), we take this to be a linearity
restriction, and address only the interaction of pronominal anaphora and wh expressions:
(i) *Hisi mother annoyed every boyi .
3 It is reported in the literature that in relatives, it is only the non-restrictive form in which
weak crossover eects are suppressed, but in view of the fact that we have not met a single
speaker who does not accept all of (6.4){(6.9), we shall take the primary data to be those in
which weak crossover and extended strong crossover eects are suppressed whenever either
the wh expression or the pronominal is a sub-part of a determiner.
4 We shall see in due course that the data are not this clearcut. For some speakers, questions
with which are also acceptable under the indicated reading:
(i) ?Which childi is hisi mother looking for?
192 Crossover Phenomena
(6.11) ?*Whosei mother does Sue know hei worries ei is sick?

Considered to be a restriction on movement of the wh element that it cannot
cross a co-indexed pronoun, this phenomenon was originally taken (following
Postal 1972, Chomsky 1981) to divide into two discrete phenomena, strong and
weak crossover, a distinction which is still widely assumed. Strong crossover
is said to be due to the universal restriction which prohibits a pronoun from
being interpreted as dependent on a c-commanding operator if it both precedes
and c-commands the trace bound by that operator (a principle C eect) { a
restriction constant across questions and relatives:
(6.12) *Whoi did hei think that Bill liked ei?
Weak crossover, on the other hand, is a separately stated restriction precluding
a pronoun from being bound by some c-commanding operator in the presence
of a following trace (a principle C-based explanation being inapplicable because
the pronominal does not c-command the gap):
(6.13) ?*Whoi did hisi mother ignore ei?
It is weak crossover that presents the greatest problems, over and above the
bifurcation of the data into weak and strong crossover, for, as demonstrated
by Lasnik and Stowell (1991), these structural characterizations do not provide
any basis for capturing the fact that weak crossover eects are suppressed in a
number of environments, one of which is relative clauses:
(6.14) John, whoi hisi mother had regularly ignored ei, fell ill during the
exam period.
(6.15) Whichi of the team did the judge put away ei without hisi mother
being able to see ei ?
(6.16) Sami is all too easy for hisi mother to ignore ei.
According to Lasnik and Stowell, this context sensitivity of weak crossover ef-
fects should be analysed away by invoking a further type of empty category { an
empty pronominal epithet, which occurs in a structural conguration described
as weakest crossover. Weak crossover eects, then, are said to arise in a con-
guration containing a sequence of an operator, a pronominal and a trace, if
the operator which is the putative binder of both pronoun and trace is a `true
quantier phrase, supposedly as in (6.12){(6.13) (notice that this is the reverse
of what one would expect if the wh operator were a regular variable binding
operator). And data such as (6.14){(6.16) are said by Lasnik and Stowell to
constitute the discrete phenomenon of weakest crossover, the empty category
not being subject to principle C, but merely to principle B.
If this invocation of a further type of empty category succeeded in resolving
all remaining problems, the implicit weakening of the theory to which it gives
rise might be acceptable. However, a whole series of problems remain, as pointed
out in detail by Postal (1993). Amongst these is the problem that neither of the
characterizations of weak or weakest crossover extend to constructions in which
it is the wh element which is in a determiner position, for in these the wh element
itself does not c-command the pronoun, and the pronoun does c-command the
gap. Yet this conguration { which, given the c-command relation between
pronoun and gap, classies them along with strong crossover phenomena { is
subject to exactly the same context-sensitive suppression of crossover eects,
licensing the sequence of wh pronominal gap in relative clauses, though not in
questions:
(6.17) ?*Whosei motherj did you tell himi ej had refused the operation?
(6.18) ?*Whosei exam resultsj was hei certain ej would be better than
anyone else's?
(6.19) John, whosei motherj I told himi ej had refused the operation, was
very upset.
(6.20) John, whosei exam resultsj hei had been certain ej would be better
than anyone else's, failed dismally.
The proposed account of weakest crossover therefore provides no basis for ex-
plaining the parallelism between these data, dubbed by Postal secondary strong
crossover, and weak crossover data; and the parallelism remains entirely unex-
plained.
An additional problem which Postal notes is the considerable variation in
weak crossover eects across languages. In dierent forms of Arabic, for exam-
ple, there are no crossover eects at all in relative clauses, in virtue of the use
of resumptive pronouns. And a language such as Hebrew, which allows relative
clauses to be projected both without and with resumptive pronouns, displays
no crossover eects in the event of a pairing between wh operator and resump-
tive pronoun, but crossover eects as marked as strong crossover eects if the
wh operator is paired with a gap (Hebrew), even in what in English are weak
crossover environments:
(6.21) ha-?isi se ?im-oi
p p ohevet ?otoi
p [Hebrew]
The man that his mother loves him
`the man who his mother loves'
(6.22) *ha-?isi se ?im-oi
p p ohevet ei
the man that mother his loves ei
Sar (1996) has argued that the weak crossover phenomenon can be resolved by
positing two alternative analyses: a derivational analysis involving movement,
and a base-generated analysis involving resumptive pronouns, with the phenom-
ena dubbed weakest crossover being resumptive and not involving movement at
all (a non-movement analysis of resumptive pronoun wh-pronominal pairings
has been taken to be uncontroversial since Ross 1967, where the characteristic
lack of any island eects with wh resumptive pairings was rst observed). How-
ever, this promise of a solution by simply providing two modes of analysis has
not provided the easy, albeit weak, solution for the crossover mystery that Sar
promised.5 To get the analysis to work, Sar has to posit a principle whereby a
syntactically derived gap (= a derivational variable) is converted to a resumptive
pronoun whenever it is not bound to a true quantier, and the relative pronoun
in non-restrictive relatives is transformed from being an operator at LF to a
referential expression at LF0 (this presumably despite current assumptions that
LF is the level at which interface properties with the semantics are dened).
As Aoun and Choueiri (1997) and Aoun and Benmamoun (1998) have pointed
out, the pairing of a dislocated expression with a resumptive pronoun is in any
case not clear indication of a non-movement analysis, for there is interaction in
all forms of Arabic between left-dislocation eects associated with a gap and
left-dislocation eects with a pronoun if both occur within the same string. If a
dislocated NP and its paired resumptive pronoun are contained within the con-
guration intervening between a wh expression and its gap, then the dislocated
NP and its paired resumptive pronoun become subject to the very same restric-
tions as those of the wh expression and its gap, viz. sensitivity to islands. Thus,
despite the acceptability of (6.23) and (6.24), (6.25) and (6.26) are reported to
be quite unacceptable:
(6.23) Drna
y l-masra iyye yalli t?arrafna ?ala l-muXriz
y p p yalli
saw1st:P lur the play that met1st:P lur on the director that
?aXraz-a
p [Lebanese Arabic]
directed3:Sing:Masc it
`We saw the play that we met the director that directed it.'
(6.24) t?arrafna ?ala l-muXriz yalli Laila seet l-masrahiyye
p p
met1st:P lur on the director that Laila saw the play

yalli (huwwa) ?aXraz-a
p
that (he) directed3:Sing:Masc: it

`We met the director that Laila saw the play that he directed.'
(6.25) *su Naadya xabbaro S- Sabe
yalli seefa
what Nadia told3:sing the boy that saw3:sing:masc: her
`What did they tell the boy who saw Nadia?'
(6.26) *ala miin Naadya arrafo r-re?eel yalli
p
to whom Nadia introduced3:sing the man that

zaara
visited3:sing:masc her
`To whom did they introduce the man who visited Nadia?'
5 Within a paradigm in which there has been extensive debate over whether explanations
should be given in terms of movement or in terms of properties of a single representation
(see Brody 1994), the conclusion that both are required has to be seen, even if correct, as
disappointing.
Aoun and Benmamoun accordingly assume that some resumptive pronouns are
generated through a movement analysis. But if this is the correct conclusion
to draw, the crossover mystery is only deepened, for one might then anticipate
that all such movement-derived cases would display weak crossover eects. But
there are no weak crossover eects in Arabic relative clauses at all.
Though the phenomenon of crossover was rst analysed within movement ac-
counts, the assumption of a division at least between strong and weak crossover
is very generally assumed { though not, however, in the categorial analysis
of Dowty (1993). On this account, the restriction arises from the interaction
between two operator binding mechanisms, one resulting in a discharge of an
initially posited assumption (roughly approximate to the gap) and one resulting
in anaphora construal (the latter having no re ex in any corresponding step of
type deduction). In the order wh ... pronoun ... gap, the pronominal binding
abstraction mechanism may bind the gap assumption, and its own binding then
renders the binding by the wh operator vacuous { hence the ungrammaticality
of the wh { pronoun { gap sequence. For this account, it is the persistence of
strong crossover eects whatever the context which is problematic:
(6.27) *Johni, who Sue said hei knew Bill dislikedi, refused to stand for
President.
To explain examples such as the non-restrictive relative in (6.14), on the in-
dicated interpretation, one might posit an indexical construal of the pronoun
referring to the same individual as denoted by the preceding head NP (see
Hepple 1990 which explicitly excludes indexical anaphora construal and any at-
tendant analogue of principle B eects from his account of pronominal binding).
But on this explanation, there is no reason to preclude a construal along these
lines for (6.27). Moreover, the explanation as it stands applies equally to weak
and strong crossover environments, so there is no basis for explaining why for
many English speakers such weak crossover eects fail to emerge; nor are there
grounds for explaining why in languages such as the various forms of Arabic all
crossover eects disappear in relative clauses.
More recently, Bresnan (1998) has argued that weak crossover eects display
dierent kinds of restrictions in dierent languages, some being sensitive to lin-
ear order considerations, others to semantic/thematic structure, and the best
that can be hoped for in characterizing the phenomena is a typological charac-
terization of how these various constraints are dierentially weighted in dierent
languages. As she points out, this raises the problem that very closely related
varieties of the same language (even within members of a single family) might
fall into discrete typological classes. Thus the interaction between anaphora and
wh construal is recognized to be heterogeneous { at best problematic, at worst
a mystery (see Postal 1993). The challenge is to provide an integrative form of
analysis, yet with suÆcient exibility to predict the observed variation.
6.2 Crossover { The Dynamic Account

By way of taking stock, we should observe that the distinctions between strong,
weak, extended strong, or weakest crossover are all of them theory-internal di-
visions, and none of the central data are ungrammatical strings. It is merely
a question of what interpretations are precluded and how. The problem, in
theory-neutral terms, lies in the complexities that arise in the interaction be-
tween anaphora construal and the association of a wh expression with a discrete
position in the string from which its interpretation is projected (the gap). The
distinguishing feature of the present framework is that it makes available under-
specied descriptions of some conguration to be assigned in initial stages of its
projection, and denes the dynamics whereby such specications are progres-
sively strengthened. And we shall now see that it is through the step-by-step
process of update that we can explain why crossover eects are uniformly ob-
served in some structures and in others partially suppressed. We take rst the
construal of relatives, where the cross-language variation seems most striking,
starting as before with English.
6.2.1 Crossover in Relatives
The dynamics of the interpretation of English relatives is most straightforwardly
displayed in non-restrictive relatives, so we take these as the point of departure.
For these, the LINK Introduction rule attaches to the root node of the LINKed
structure an unxed node with a requirement for a copy of the formula annotat-
ing the head, and the projection of the initial wh then satises that requirement,
leaving this node with only the need to provide a xed tree position. The rst
question is then why a subsequent pronoun intervening between the projection
of the unxed node and the point at which its tree position is established can-
not be interpreted in like manner, with a copy of the head formula providing its
value. Once, however, the unxed node has found a xed position, a pronoun
can freely be interpreted as identical to the head:
(6.28) *Johni, whoi hei knows Mary likes ei, is insisting on ignoring her.
(6.29) *Johni, whoi Sue thinks hei knows Mary likes ei, is insisting on
ignoring her.
(6.30) Johni , whoi ei knows Mary likes himi , is insisting on ignoring her.
(6.28) and (6.29), though as strings as well-formed as (6.30){(6.34), preclude
any interpretation of he as identical to John. The second question then is
why if either the pronominal or the wh element is nested within a possessive
construction such identication becomes possible:
(6.31) Johni , who hisi aunt knows Mary likes ei , is insisting on ignoring her.
(6.32) Johni , who ei knows Mary likes hisi aunt, is insisting on ignoring her.
Crossover { The Dynamic Account 197
fT n(a); ?T y(t)g

John
fhL 1 ih"0 iT n(a); ?T y(t); }g
fh" ihL 1 ih"0 iT n(a); F o(John); T y(e); ?9 xT n(x)g

who
Figure 6.1: The processing of John, who
(6.33) Johni, whosei auntj hei knows Mary likes ej , is insisting on ignoring
her.
(6.34) Johni, whosei auntj ej knows Mary likes himi , is insisting on ignoring
her.
(6.35) Johni, whosei auntj hisi sister knows Mary likes ej , is insisting on
ignoring her.
The reason for this emerges from considering the successive steps of interpreta-
tion.
We take, rst, the interpretation of (6.28), paying close attention to the
construction steps. In (6.28), the relative clause is initiated by the imposition
of a requirement that within the resulting representation there must be some
node annotated with the formula F o(John) (through LINK Introduction ). The
immediately subsequent processing of the relative pronoun annotates the un-
xed node bearing this requirement, so that all that is left is to establish a xed
tree position. We have, that is, the partial tree description which we might
display as in gure 6.1. The next word to be interpreted in the sequence is he,
which, in virtue of its case specication, is established as annotating a subject
node (constructed through a step of Introduction and Prediction ). The meta-
variable associated with this pronominal requires a substituend, and a candidate
antecedent formula is F o(John) annotating the node of the previously estab-
lished incomplete tree described as fh"0iT n(a); F o(John); T y(e)g. However,
this annotation cannot be chosen. Interaction between an unxed node anno-
tated by John and the locality constraint on Substitution associated with the
pronominal ensures that there is no possible means of assigning this interpreta-

tion: the unxed node can become a co-argument with the node annotated by
the pronominal meta-variable, but the locality restriction on pronoun construal
(section 3.3) then excludes the annotation at that unxed node taking the same
value as that assigned to the meta-variable (see gure 6.2 for a diagrammatic
display of the locality domain within which a pronoun's meta-variable may not
establish a value).
T n(a); ?T y (t)
h" iT n(a); h"0 iT n(a); h"1 iT n(a);

?T y(e) ?T y(e) ?T y(e ! t)
h"0 iT n(a); h"0 ih"1 iT n(a);
?T y(e) ?T y(e ! t)
h"0 ih"1 ih"1 iT n(a); h"1 ih"1 ih"1 iT n(a);
?T y(e) ?T y(e ! (e ! (e ! t)))
Figure 6.2: The domain of locality
The result is that any application of Substitution replacing the meta-variable U
with John in (6.28) would be in violation of the dened locality restriction.
There is only one possible way of side-stepping this locality restriction while
nevertheless assigning the pronoun the same interpretation as the head, and that
is to merge the unxed node with the node annotated by the pronoun. Indeed,
nothing precludes an annotated but unxed node merging with an appropriately
located node decorated with F o(U) as a means of providing the substituend for
that meta-variable, for update from F o(U) to F o() is a monotonic update.
However, any identication of the unxed node and the xed node annotated
by F o(U) ensures that the following string cannot contain what is familiarly
called a gap. For should there be any node subsequently constructed with a
requirement ?T y(e) which is not met by input from the string, there will no
longer be the unxed node to merge with. So this requirement would remain
outstanding, and the string could not lead to a completed logical form. This
situation arises in two types of circumstance: the gap occurs nally and the tree
is left with outstanding requirements; or the gap occurs in the middle of the
string, the condition of the incoming word is not met, and the parsing process
aborts.6 There will therefore be no interpretation of a relative clause containing
a pronominal preceding the gap in which the pronominal is identied with the
head; and we derive the fact that (6.28) is precluded.7 An application of Merge,
6 This analysis is an extension of an earlier analysis of Kempson and Gabbay (1998) purely
in terms of locality, an account which provided no basis for explaining resumptive use of
pronouns.
7 Notice of course that this is a question of exact denotation, since nothing whatever pre-
cludes the pronoun from being identied as some other individual named John. The repre-
which simultaneously provides a xed tree position for the unxed node and a
value to the U variable for a pronoun, will yield an acceptable structure as long
as there is no gap in the following sequence of words:
(6.36) Johni, whoi Sue thinks hei realizes hei has done wrong, is looking
suitably contrite.
Because the unxed node is addressed by h"iX , the preclusion of an interpre-
tation in which the same value is assigned to the head formula, the pronoun
and an unxed node applies even if the pronoun preceding the gap is contained
within a subordinate clause. So (6.29) will be ill-formed on the construal of he
as `John', just as (6.28) is.
If the pronoun is interpreted as projecting some formula other than that
annnotating the head, then identication of the unxed node and the node
decorated by the pronoun would yield an inconsistent description. So under
such an interpretation of the pronoun, the unxed node must be merged with
some other node. So we capture successfully the fact that (6.28) and (6.29) are
well-formed strings, but disallow co-construal of the pronoun and the head of
the relative.
When we turn to cases of so-called weak crossover and extended secondary
strong crossover, the question of locality of an unxed node with respect to the
node to be decorated through processing a pronoun will not arise, because in
these cases one or other (or both) of the wh and pronominal are contained within
a possessive construction. Genitives, recall, are dened as projecting a structure
LINKed to the variable projected by the following nominal, with the structure
of that LINKed tree, one of its arguments and the subsequent quantication
over the argument variable of the following head, all being lexically induced by
the genitive particle. In accordance with this analysis of genitives, a U meta-
variable internal to his lies on a path from the root of the LINKed structure
which includes an L step, so it can be identied by pronominal substitution of
the formula assigned to the head. There is no interaction with Merge, because
resolution of the unxed node cannot take place across an L step. This has
the consequence that the unxed node cannot be merged with that decorated
by the pronominal in these structures { hence the lack of weak crossover eects
in relative clauses. The interpretation of the sequence John who his as in (6.31)
is displayed in gure 6.3, where the lack of any two occurrences of F o(John) in
the same tree is evident.
The extended strong crossover cases such as (6.33) are equally unproblem-
atic, and for essentially the same reasons. It is the wh expression in these
structures which is in a LINKed structure, and the sub-tree projected by the
possessive expression is LINKed to the unxed node which initiates the inter-
pretation of the relative sequence. A subsequent pronoun he in that relative
structure may in consequence be construed as taking the value F o(John) with-
out violating any locality requirements, for there is no occurrence of F o(John)
sentation here is simplied, but nothing turns on this approximation in the specication of
proper names.
fT n(a); ?T y(t)g
fh"0 iT n(a); fh"1 iT n(a); ?T y(e ! t)g

F o(John); T y (e)g
John
fhL 1 ih"0 iT n(a);

?T y(t)g
fh" ihL 1 ih"0 iT n(a);

F o(John); T y (e)g f?T y(e)g f?T y(e ! t)g
who
fT y(cn ! e); fT y(cn ! e);

F o(P (; x; P ))g F o(P (; x; P ))g
fT y(e); F o(U)g f?T y(e ! cn); }g
fhL 1 iT n(m); F o(P OSS (U)(John));

T y (t)g
fF o(John); T y(e)g
fF o(P OSS (U); T y(e ! t)g
his
Figure 6.3: The result of having parsed John, who his

within the tree containing the unxed node; and identication of this node and
the unxed one can never yield an interpretation matching the specication
of the pronoun. Put simply, identifying he as `John' in (6.33) does not solve
the problem of how the sub-tree projected from whose aunt should contribute
to the building up of the overall tree structure: the identication of the po-
sition for this unxed node hence remains as a task yet to be satised in the
construction process (see gure 6.4). Hence, as in the weak crossover cases,
there are no crossover eects.8 Interpretation of restrictive relative clauses, we
have suggested, proceeds exactly as in the non-restrictive case, except that the
head of the LINKed structure is the nominal variable projected by the common
noun. This being the case, there is a copying of a nominal variable U onto
(some daughter of) an unxed node within the LINKed tree. It is sometimes
suggested in the literature that some speakers dierentiate between restrictive
and non-restrictive relatives, with suppression of crossover eects in associa-
tion with possessive constructions only in non-restrictive relatives. However, we
are unable to nd any speaker who consistently distinguishes the two types of
structure, so we characterize them both in exactly the same terms.9
The consequence for English is, on the face of it, surprising. According to
the analysis, merging of an unxed node annotated by F o(WH) and a xed node
annotated by F o(U) may be a means of anaphora resolution, F o(WH) replacing
the F o(U) and thereby providing the antecedent for the pronominal. This
leads to the prediction that structures in which resumptive pronouns replace
the empty position will be wellformed:
(6.37) ??The student who Sue says he should have discontinued his studies
has nevertheless turned up to class.
(6.38) ??The student who Sue says she wished she had failed him has applied
to the Ph.D program.
(6.39) ??The man who I argue with him is at home.
(6.40) ??The man who he agrees he's been overworking is coming on holiday
with us.
In all of these, nothing precludes Merge unifying the unxed node and the node
decorated by a meta-variable, leading to replacement of the U variable by the
8 Some speakers report asymmetry between weak crossover eects and extended strong
crossover eects, with the latter retaining the strong crossover eect both in questions and
in relatives. For these speakers, we would anticipate a correlation with extraction possibili-
ties, such speakers also allowing extraction out of this determiner-initial position (recall the
reported observation in section 4.3.4 that some speakers tolerate (i):
(i) The children who Sue said's mother was lazy were late for school.
9 The suppression of weak crossover eects is not restricted to relative clauses (see Lasnik
and Stowell 1991). However, the other environments in which such eects are suppressed are
substantially similar. For example, easy to please constructions are explicable exactly as other
crossover eects, and behave accordingly:
(i) *Johni is easy for himi to ignore.
(ii) Johni is easy for hisi mother to ignore.
fT n(a); ?T y(t)g
fh"0 iT n(a);
F o(John); T y (e)g fh"1 iT n(a); ?T y(e ! t)g
John
fhL 1 ih"0 iT n(a);

?T y(t)g
fh"0 ihL 1 i
h"0 iT n(a); fh"1 ihL 1 i
fh" ihL 1 ih"0 iT n(a); ?T y(e)g F o(John); T y (e)g h"0 iT n(a);
?T y(e ! t); }g
he
fT y(cn ! e); ?fT y(cn);

F o(P (; x; P ))g ?h#1 iT y(e ! cn)g
fT y(e ! cn);
fT n(m); T y(e); F o(Aunt)g
F o(U)g
aunt
fhL 1 iT n(m);
F o(P OSS (John;U));
T y (t)g
fF o(John);
T y (e)g fF o(P oss(U)(John));
T y (e)g
whose
Figure 6.4: Interpretation of John, whose aunt he

nominal variable and updating of the tree node identier for that node. Indeed,
in cases such as (6.37){(6.40), the structure assigned to the string will not be
well-formed unless the pronoun is so construed and the two nodes are unied,
for otherwise it would still contain an unxed node. Yet, despite frequent use of
such pronouns in naturally occurring data, judgements on these sentences are
mixed, as the occurrences of `??' indicate, with judgements of unacceptability
of (6.39) and (6.40) generally stronger than judgements on (6.37) and (6.38).
We leave further explanation of these resumptively used pronouns on one side
for the moment, merely noting that under certain circumstances such forms
become fully acceptable without any change of structure, suggesting that it is
pragmatic considerations that dictate why some pronouns cannot be construed
resumptively:
(6.41) The assistant who I made sure that at least he got a rise was duly
grateful.
(6.42) The professor who I made sure that at least I didn't get him as a tutor,
cornered me at the party to ask if I knew why he'd never tutored me.
(6.43) My son, who he, even, agrees he's been overworking, is coming on
holiday with us.
(6.44) My son, who I argue even with him, has refused to come home, on the
grounds that it's too stressful.
6.2.1.1 Arabic/Hebrew Relatives

The account of crossover in relatives so far has turned on the assumption that the
relative pronoun annotates an unxed node and the identication of the pronoun
with the head may provide a xed position for that unxed node. This account
suggests that crossover eects in languages in which the relative pronoun fails
to provide an unxed node should be dierent. In particular, we would expect
that in a language such as Arabic, there would be no analogue to these strong
crossover eects, because the relativizer does not project an unxed node. So
there is no gap, and, consequently, there are no crossover eects. The only
way of ensuring the presence in the LINKed tree of the formula to be copied is
through the use of a pronoun. This is indeed the situation { there is a complete
lack of crossover eects in Arabic relatives. Nothing special needs to be said
about this case. In addition, because there is no unxed node, there is nothing
corresponding to the pied-piping eects displayed by Romance languages, where
a complex formula annotates an unxed node. Furthermore, since the process of
anaphora resolution is freely available (modulo its locality restriction), nothing
precludes a pronominal from establishing its antecedent across a LINK relation
L (from Demirdache 1991):
(6.45) qara?tu ll-maqaalat llaÆii saafara

p
Saabu
S- llaÆii
read-I the article that travelled the young man that
kataba-ha. [Classical Arabic]
wrote it
Ì read the article that the young man who wrote it travelled.'
6.2.1.2 Hebrew `Moved' Resumptive Pronouns

Perhaps surprisingly, this analysis is compatible with the `moved resumptive'
data displayed by Hebrew, which supercially yields the very crossover order
the system debars (see Demirdache 1991). Hebrew is a mixed system. Like
Arabic, it licenses resumptive pronouns, but, like English, it licenses gaps in
object and subject positions in relative clauses, gaps and resumptive pronouns
in general being in free variation in these positions, except the topmost subject
(see Shlonsky 1993):
(6.46) ha-?is se- ra?iti (?oto)
p p p [Hebrew]
The man that I saw (him)
`the man I saw'
(6.47) ha-?is se- xasavt
p se- Dani pagas (?oto) p
the man that youF thought that Dani met (him)

This goes together with a licence for preposed subjects and objects in general
to be associated with gap or pronominal resolution:
(6.48) Dani xasavt se- ra?iti (?oto)
p p
Dani, youF thought that I saw (him)

Às for Dani, you thought I saw him.'
The analysis we give re ects this mixture. As in Arabic, Hebrew imposes no
restriction at all on where in a LINKed tree the information transferred from
the head node should occur. This is because the subordinating particle se
may, like the Arabic complementizer yalli(or illi),10 simply serve to induce the
LINK transition, imposing a modal requirement for an occurrence of a copy
of the formula annotating the head node anywhere in the LINKed structure.
Because there is no antecedent occurrence of this formula within the induced
LINKed structure, locality does not restrict the occurrence of the pronominal.
However, like English, Hebrew also allows se to serve an anaphoric function
which, following upon a step of *Adjunction, leads to the construction of a copy
of the head formula as an annotation on an unxed node. This second function
of se forces this initially unxed node to be identied within the local tree (i.e.
not crossing a LINK relation).
10 Note the form llaÆii of Classical Arabic in (6.45).
The free variation between pronouns and gaps in the construal of rela-
tive clauses allows us to explain why a resumptive pronoun can apparently
be `moved' to a higher clause yielding the typically precluded crossover order
relativizer : : : pronouns : : : gap, a phenomenon which we saw in a dierent
context in section 4.3.1:
(6.49) ha-?is se ?ani xosev se ?alav
p p ?amarta se Sara
p p
the man that I think that about him you said that Sara
katva sir
wrote poem
`the man that I think that about him you said that Sara wrote a poem'
(6.50) ha-?is
p se ?alav
p ?ani xosev se
p ?amarta se
p Sara
the man that about him I think that you said that Sara
katva sir
wrote poem
`the man that about him I think that you said that Sara wrote a poem'
(6.51) ha-?is
p se ?ani xosev se
p ?amarta se
p ?alav
p Sara
the man that I think that you said that about him Sara
katva sir
wrote poem
`the man that I think that you said that about him Sara wrote a poem'
The explanation turns on the fact that there are two aspects of underspeci-
cation in the projection of LINKed structures. On the one hand, there is the
requirement for the construction of a copy of the formula from the head node.
On the other hand, there is the requirement for some tree node to host this re-
quired annotation. Though these requirements may be met by the projection of
a xed node duly annotated directly (by suitable construal of a pronoun taken
to annotate a xed tree node), it is also possible for these two requirements to be
met separately. In particular, the construction of a copy of the required formula
can be met by first building an unxed node and decorating it. The second
step would then be nding a xed position for this node. This two-stage sat-
isfaction of a modal requirement initially imposed by LINK Introduction takes
place when the pronoun is taken to annotate a position which is itself unxed.11
We have the transition from
fhL 1 iT n(a); ?T y(t); ?hDi(F o(); T y(e))g
11 As Demirdache (1991) points out, reporting Borer (1984), topic constructions are possible,
albeit marked, within a relative clause in Hebrew (?et is a free standing morpheme which
p
marks the denite direct object of a verb):

(i) ?eyle ha- ?ansim se ?et David salaxti ?eleyhem
p p p p
these the-people that ACC David sent-I to-them

`These are the people that David, I sent to them.'
to
fhL 1 iT n(a); ?T y(t); ?hDi(F o(); T y(e))g; fh"ihL 1 iT n(a); ?F o(); T y(e)g
as the rst step,12 and the transition to
fhL 1 iT n(a); ?T y(t); ?hDi(F o(); T y(e))g; f" hL 1 iT n(a); ?F o(); T y(e)g
as the second step. Note that the external operator " replaces the internal
operator h"i, describing the replacement of an unxed node by some xed
node, hence a monotonic form of update. We thus derive the dierence between
the English case (6.52) and the Hebrew (6.49){(6.51):
(6.52) *Johni, who Sue said hei knew Bill disliked ei, refused to stand for
president.
In (6.52) the interpretation for the pronoun is projected as part of the tree
structure at a xed node of type e, precluding any subsequent gap. In (6.49){
(6.51) the pronoun projects a formula onto a node, but the position of that
node in the description is not xed. This satises the modal requirement, but a
further step is required to identify a xed position for this node within the tree.
6.2.1.3 Apparent Crossover Eects in Hebrew
A second property of Hebrew is the marked weak crossover eects observed in
restrictive relatives, reportedly as strong as strong crossover eects in English
(see Demirdache 1991):
(6.53) *ha-?is sei ?im-oi
p p ?ohevet ei
p
the man that mother him loves e

In Hebrew, the genitive can be expressed in two forms: either by a regular
noun head with a modifying PP; or by the so-called construct state structure,
in which the head noun, which occurs in a phonologically de-stressed form, is
immediately followed by a full NP (and not a PP):
(6.54) ha-bayit sel ha-?is p
`the house of the man'

(6.55) beyt ha-?is
p
house the man

`the house of the man'
12 As in chapter 4, we are assuming for simplicity that the form ?alav is of T y(e), with the
p
prepositional content being an annotation on that node.

We dene a specic sequence of actions for these de-stressed nouns that re-
ects the bi-nominal structure of the construct state directly.13 Any such form
takes as a trigger ?T y(e), and induces a skeletal structure of type e with nodes
for determiner, nominal, nominal variable and restrictor. Within this struc-
ture, the nominal variable and restrictor get annotated in the same way as
other (full) common noun forms. What is specic to the projection of this
structure is the construction of a complex determiner, which builds from the
determiner node a pair of daughters, one annotated with a P OSS operator,
xP [; (U; P OSS (U)(x) ^ P (U))], the other decorated with a requirement of
the form ?T y(e). It is the node requiring ?T y(e) which is where the pointer
resides, thus expecting some noun phrase to immediately follow.14 In (6.53), for
example, this node of type e within the determiner tree is annotated by scan-
ning the pronoun and construing it as identical with the formula annotating the
head: it is this node which may provide the copy of the head formula required
by the LINK Introduction. Given that in this language LINK introductions are
projected from se either as a transition which imposes a requirement for a copy
of the head on the root of the LINKed structure (as in Arabic), or as a transition
which involves building and annotating an unxed node, the unacceptability of
(6.53) must follow from both procedures. With respect to the rst of these,
the interpretation of o as identied with the head will indeed satisfy the re-
quirement imposed by LINK Introduction, but the string, lacking any object
following ?ohevet, will not induce a complete logical form. On the analysis in
p
which se annotates an unxed node, the locality restriction constraining the
interpretation of the pronominal will preclude the appropriate substitution, so
the only possible way to assign the requisite interpretation to the pronominal is
by Merge. But with the unxed node having merged with the node annotated
by the pronominal, the projection of the tree will thereafter remain incomplete
in the presence of a subsequent gap. Either way, that is, the presence of the
suÆx o within the construct state structure will have the eect of establishing
a xed position in the tree for the required copy { hence the reported eects
that the ungrammaticality is of the same order as strong crossover in English
(and not a weaker weak crossover violation). Relative to our model, (6.56) is
well-formed, as the analysis would lead us to expect:
13 The form of construct state nouns is discrete from their independent forms, is phono-
logically incomplete, with reduced stress and consequent shifts in both vowel and consonant
forms, and cannot occur independently.
14 As this analysis re ects, such construct state forms lack a morphological determiner for
the head noun, can be recursively complex, and allow a determiner to intervene between the
head noun and its modifying noun. They also allow a clitic-doubling form of construction
with a following PP which apparently duplicates the suÆxed clitic:
(i) beit-o sel ha-more
house his of the teacher
`the teacher's house'
This structure we could dene as involving a LINK relation, the structure projected from the
PP with sel LINKed to the nominal variable of the preceding noun{clitic sequence happening
to project the very same P OSS relation, and therefore imposing identity of interpretation on
the pronominal clitic and the following full noun phrase.
(6.56) ha-?isi
p se ?im-oi
p ?ohevet ?otoi
p p
the man that his mother loves him

This analysis of the Arabic and Hebrew data is that we have explained the
resumptive pronoun distribution and crossover eects in the same terms as the
earlier English data, with only minor modication according as the relative
pronoun is or is not assumed to provide a copy of the formula of the head.
The explanation has not been in terms of any concept of movement; rather,
it has been in terms of how anaphora construal interacts with the process of
constructing a logical form to give rise to long-distance dependency eects.
6.2.1.4 A Dutch Puzzle
Not all crossover data are as straightforward to characterize within the proposed
framework. Crossover phenomena in Dutch relatives, for example, appear to
present problems for the analysis. Dutch poses the problem that it has a relative
pronoun die which displays strong crossover eects in relative clauses while
disallowing all resumptive use of pronouns, suggesting that the explanation of
strong crossover phenomena cannot turn on the application of Merge to yield a
resumptive construal:15
(6.57) *de mani die hiji verwacht dat ziek is [Dutch]
the man that he expects that sick is
`the mani that hei expects ei is sick'
(6.58) de mani die verwacht dat hiji ziek is
the man that expects that he sick is
`the mani that ei expects that hei is sick'
(6.59) *de mani die hiji verwacht dat hiji ziek is
the man that he expects that he sick is
`the mani that hei expects that hei is sick'
In our model, the relative die, like other relative pronouns in Germanic lan-
guages, annotates an unxed node, but this node must be resolved by merging
it with a node which has no decorations other than ?T y(e). To ensure this, we
can assume that in Dutch the variant of Merge made available is the more strin-
gent form dened in chapter 3 as Gap Merge (see chapter 3, section 3.2.1.3).
However, Dutch also uses a pied-piping form of construction with van wie which
allows resumptive construal of pronouns, thus suggesting that perhaps the more
general form of Merge is required after all:
15 We thank Ad Neelemann for this observation and the discussion to which it led.
(6.60) de man van wie Sue verwacht dat Tom (hem)
the man of whom Sue expects that Tom (him)
opzettelijk negeert
deliberately ignored
`the man whom Sue expects Tom deliberately ignored'
(6.61) de man van wie Sue verwacht dat hij ziek is
the man of whom Sue expects that he is sick
`the man whom Sue expects is sick'
However, the proposed analysis of crossover is undisturbed by such examples.
We assume that in Dutch all Merge operations are of the more restricted variant,
Gap Merge, ensuring that any left-dislocated constituent cannot be identied
with a lexical pronominal. In strong crossover environments, any pronoun in-
tervening between the projection of the unxed node and the supposed point
at which Gap Merge should apply is precluded by locality restrictions from
identifying with the annotation of the head. With the relative pronoun die,
this is the entire story. With the van wie forms, on the other hand, a discrete
LINK Introduction can be projected by van wie, imposing, as in Arabic, a re-
quirement on the top node of the structure linked to the nominal variable for
a requirement of a copy, which the pronoun provides. In the latter case, the
pronoun is given a value by Substitution. Thus, though there may be individual
expressions that allow resumptive pronouns, Dutch has no general process of
pronominal construal through a Merge operation. The general form of expla-
nation for crossover phenomena is, however, unaected: it is merely simpler,
for the locality restriction associated with Substitution suÆces to preclude a
pronoun between the antecedent16 and the point in the interpretation process at
which Gap Merge might apply.
6.2.1.5 Constraining the Liberality of the Computational System?
A major question for the proposed analysis is its liberality. By allowing Merge
to be a means of assigning an interpretation to a pronoun, all head-initial rela-
tive structures will in principle be expected to license the occurrence of lexical
pronouns interpreted resumptively in positions where a gap may occur. Yet
we have already seen that this cannot be the whole explanation. In particular,
there appears to be a very general avoidance of pronouns being resumptively
construed too close to the start of the clausal sequence:17
(6.62) ?*The man who he likes my mother has come to visit us again.
(6.63) ?*The man who my mother dotes on him has come to visit us again.
16 We have not analysed English in terms of Gap Merge because, as we shall see immediately,
the acceptability of using pronouns resumptively in English is determined by an array of
system-external factors.
17 The `?',`??' or `?*' in the examples that follow are merely crude indicators of relative
judgements of (un)acceptability.
Whatever restriction this is against resumptive construal of pronouns, it is

weaker if the pronoun occurs in a clause subordinate to the relative clause;
and it is suspended altogether in so-called weak island environments:
(6.64) ??The man who Sue says he dotes on my mother has come to visit us
again.
(6.65) ??The man who Sue says my mother dotes on him, has come to visit
us again.
(6.66) ?The man who Sue says she's certain he is courting my mother has
come to visit us again.
(6.67) ?The man who Sue says she's certain my mother has slept with him
has come to visit us again.
(6.68) That colleague of mine who if he leers at me once more, I promise I'll
scream abuse at him, has been assigned the same oÆce as me.
(6.69) I have to interview the student who we're trying to decide whether he
should be admitted to the program.
The question then arises as to whether any form of restriction should be dened
within the computational/lexical system, analogous to what has been charac-
terized as an Avoid Pronoun principle (see Montalbetti 1984, Chomsky 1981).
The rst observation is that any such restriction would not have the robustness
of, for example, the restriction debarring pairing of an unxed node projected
by a left-dislocated expression and a node within a structure projected from a
relative clause. In these cases, even if pronouns are used in place of a gap, the
strings remain equally unacceptable:
(6.70) *I'm going to the meet the man who Sue has published every article
she has written about him.
(6.71) *I distrust the man who our neighbours bought a car which he had
renovated.
Secondly, the restriction does not seem to be structurally denable, since even
in structures in which the pronoun occurs within the top clause of a relative,
the presence of either contrastive stress or the presence of some accompanying
adverbial phrase, which requires a stressed pronoun, substantially alleviates the
judgement of unacceptability:
(6.72) ??The friend of mine who he admits he needs a holiday has come to
London with us.
(6.73) ?The idiot who he had to be the one to admit that we did very little
teaching was roundly condemned.
(6.74) ?The friend of mine who only he has no idea how ill he is, is rmly
rejecting the operation.
(6.75) ?The friend of mine who he, even, admits he needs a holiday has come
to London with us.
(6.76) ?That friend of yours who, despite all our eorts to dissuade him, he is
determined to join the expedition, is arriving in Kathmandu tomorrow.
(6.77) ?The man who Sue says my mother especially dotes on him is
presenting her petition.
Finally, the restriction is not specic to any particular rule such as Merge,
as we would expect if it restricted a particular process. This is so, because
the same pattern of data is displayed in Arabic, even though the processes
whereby LINKed structures are constructed are, as we saw in chapter 4, quite
distinct (with LINK Introduction inducing only a requirement on the top node
of the LINKed structure and the normal lack of pronominal in subject position
re ecting the projection of the subject node by the verb):
(6.78) l-m?allme ?aaSaSit
p p l-walad yalli Laila
the teacher punished3:Sing:F em the boy that Laila
bta?if
p miin (huwwa) Darab. [Lebanese Arabic]
know3:Sing:F em who (he) hit3:Sing:Masc
`The teacher punished the boy that Laila knows whom he hit.'
(6.79) ??irra:gil illi huwwa mabsu:t [Egyptian Arabic]
`the man who he is happy'
(6.80) irra:gil illi hatta huwwa mabsu:t
the man who even he happy
`the man who even he is happy'
(6.81) ?irra:gil illi huwwa mabsu:t
`the man who he is happy'
That is, just as in English, resumptively construed pronouns can occur in sub-
ordinate clauses with judgements of borderline acceptability, and even in the
top clause of a LINKed structure, resumptively construed pronouns are licensed
if such a form is stressed.
Data such as these present us with two choices. On the one hand, we can
try to dene a restriction which would ensure that strings in which a pronoun
resumptively construed occurs in a position where it is unnecessary for the re-
sulting interpretation do not give a completed logical form. Such a restriction
would be relatively straightforward to express for a language such as English,
dening a restriction on the actions to be associated with resolving an unxed
node, for example through Gap Merge.18 However, it is hard to see how the
required restriction in Arabic could be expressed. In Arabic relatives, there is
no projection of any unxed tree node, and the oending lexical pronoun is in
competition with a meta-variable projected by the verb, both of which must
make provision for a value to be assigned from an external antecedent (their in-
dexical construal). The move to debar these data as ill-formed in either English
or Arabic would, furthermore, create the concomitant diÆculty of explaining
why such strings may be, and often are, unproblematically assigned an inter-
pretation. The alternative is to dene a system which is liberal with respect to
the data in allowing as well-formed some strings which speakers may express
varying degrees of hesitation over accepting. We opt for this route as our general
strategy. In a system with a monotonic concept of tree growth, with no possibil-
ity of retraction, it is harder to explain how some string interpretation pairings
judged as acceptable are precluded than it is to explain how some unaccept-
able pairing of string and interpretation might nevertheless be included. In the
face of data where acceptability varies according to non-structural factors, we
assume that the constraints determining acceptability are a consequence of rel-
evance or other processing considerations, retaining the characterization given
in chapter 3, that it is the triple of computational, lexical and pragmatic ac-
tions which determines available pairings of strings of the language and assigned
logical forms.
We do not oer here a detailed pragmatic account of focus and contrastive
stress with which to substantiate this methodological decision.19 If a relevance-
theoretic perspective of this sort is adopted, with its commitment to explaining
interpretation in terms of balancing inferential eects against cognitive cost,20
the exceptional resumptive use of pronouns might be seen as arising from the
speaker setting aside the consideration of cost to the hearer of parsing an unnec-
essary pronoun, in the light of wanting to convey particular inferential eects
which can only be ensured through the use of a (contrastively stressed) pro-
18 In order that the application of Gap Merge in English should not be vacuous, the more
general form of Merge would have to be precluded from applying in the same context. This
would leave the admittedly borderline acceptability of (6.62){(6.69) and (6.72){(6.77) quite
unexplained unless a discrete lexical specication was posited for the stressed form of each
pronoun, such forms being analysed as expletives that fail to provide any meta-variable to sat-
isfy the requirement ?T y(e) (thereby enabling Gap Merge to apply), but nevertheless project
a focus feature relative to which focus eects could be dened. Not only would such a char-
acterization be at best partial, excluding the unstressed resumptive use of pronouns, but it
would also be predicted to be a fully acceptable form of wh question, at least if the widely held
assumptions that the left-dislocated position is a focus position, and that fronted wh expres-
sions in wh questions are associated with focus, are retained. For discussion, see Rochemont
and Culicover 1990, Chomsky 1971, Erteschek-Shir 1997. Yet it is these structures for which
resumptive construal of pronouns is problematic.
19 See Breheny 1996 for a relevance-theoretic characterization of focus giving a representa-
tionalist alternative to Rooth's model-theoretic projection of accompanying context (Rooth
1992, 1996), in terms of sets of premises made available for inference.
20 Recall the characterization of the presumption of optimal relevance as an assumption that
the intended array of inferential eects is achievable with minimal cognitive eort necessary
to retrieve that interpretation (Sperber and Wilson 1995).
noun.21 Without any such inferential eects, such use of pronouns is avoided,
because unnecessary. Adopting the methodological preferences of a more, rather
than less, liberal system, with nothing more than a promissory note of some
forthcoming pragmatic explanation,22 we retain a formal system which licenses
resumptive use of pronouns in relative clauses and, accordingly, an account of
crossover phenomena which has this construal of pronouns as a consequence.
6.2.2 Crossover Phenomena in Questions
We have so far left entirely on one side the apparent asymmetry between
crossover phenomena in relatives and in questions. It might seem that the
decision to allow pronouns in English in all positions in which a successful ap-
plication of Merge can take place is too hasty, given that pronouns cannot gen-
erally be used resumptively in wh questions (6.82) and (6.83), and many people
provide judgements of weak crossover eects that his cannot be identied with
who in (6.84) and (6.85):23
(6.82) ??Whoi did Sue insist that we invite himi to the party?
(6.83) *Whoi did Mary like himi ?
(6.84) ??Whoi did hisi book irritate?
(6.85) ??Whoi did hisi mother wish wasn't sick?
This reported unacceptability of weak crossover construal of pronouns in ques-
tions, and the accompanying unacceptability of pronouns resumptively con-
strued, is, as in relatives, not explained by the account to hand. On the basis of
the analysis so far, we would expect that (6.84) and (6.85) allow an interpreta-
tion in which the pronoun is identied by substitution of WH for U in F o(U)
projected by his, since the possessive in English projects a LINKed structure,
and this substitution would not undercut the possibility of some subsequent
application of Merge. Furthermore, we would anticipate, just as in relatives,
that a pronoun might under certain circumstances serve to identify the position
of the initially unxed node in a wh question, thereby yielding a resumptive
construal. Yet it is very generally agreed that wh questions are obligatorily
associated with a gap, an association which is observed even in those languages
21 The question of estimating cognitive cost is not unproblematic, since there is a tension as
to what it is that contributes most to cognitive cost { the parsing of a pronoun, uncertainty
as to where Merge should take place, or delay in establishing some application of Merge.
22 This characterization of the underlying pressures upon choice of form of expression as
outside the remit of the computational system is in contrast to the Minimalist program,
where analogous restrictions in the form of Minimal Move are dened to be central to the
computational system (see Johnson and Lappin 1997, 1999, for an attack on this attempt
to assign system-internal status to any such restriction). In this respect, it is also unlike the
recent moves within LFG to adopt an optimality-theoretic perspective to characterize dierent
`strengths' in pronoun types (see Bresnan 1999).
23 As before, the `??' and '*' are indicators of acceptability judgements, rather than indica-
tors of ungrammaticality.
where the resumptive pronoun strategy is the only strategy available in rela-
tives.24 We could take this as evidence that wh questions are dened along
lines suggested for Dutch relativizer die as simply being gap-seeking. There is
reason, however, to think that the apparent display of asymmetry between rela-
tive clauses and questions with respect to the licensing of a resumptive construal
of pronouns is not a lexically dened restriction, but is rather due to the inter-
action of computational and pragmatic considerations jointly delimiting the set
of sentence-interpretation pairs licensed by the computational/lexical actions.
In eect, we shall be arguing that the phenomenon dubbed `weak' crossover
is a phenomenon where the preclusion of certain interpretations of a pronoun
is a consequence of pragmatic considerations alone, and is not an exclusively
structural phenomenon.
We take rst the strong crossover data, only to set them aside as exactly
parallel to their counterpart in relatives.
(6.86) *Whoi did Sue believe hei realized ei had given the wrong answer?
Any attempt to provide the value WH for the meta-variable U projected by the
pronoun by Substitution falls foul of the imposed locality constraint, and any
application of Merge to the node annotated by F o(U) and the unxed node
leads irrevocably to an incomplete structure. We leave these data without more
ado.
With the weak crossover data, we face the question of the degree to which a
wh expression can be used to provide an antecedent for a following pronominal
in wh questions. In this connection, the rst point to observe is that whatever
the supposed asymmetry between (6.87) and (6.88),
(6.87) ?Whoi did hisi mother defend?
(6.88) The man whoi hisi mother defended won his case.
the tendency to avoid using a fronted wh expression as antecedent for a subse-
quent pronominal cannot be put down to any lack of content for a WH meta-
variable. A wh expression in such a position can be successfully used to provide
an antecedent for construal of a pronoun within some following inserted paren-
thetical structure { and notice that once having been so used, the interpretation
of the pronominal within the following weak crossover environment becomes
fully acceptable:
(6.89) Whoi , if hei could be persuaded to stay, would hisi supervisor agree to
support?
24 Some languages which have a distinction analogous to the what/which distinction of
English are reported to dierentiate between the what/who types of form and the which
type of form, with the former requiring a gap, the latter a pronoun resumptively construed:
see Dobrovie-Sorin 1990 for discussion of Romanian. It is notable in this connection that
the Romanian wh interrogatives that require a clitic pronoun also require a particle pe, a
particle which appears to project a LINK relation between the left-dislocated expression and
the remainder, being obligatory also in relative clause sequences where a clitic pronoun is also
required.
(6.90) Whoi , no matter how long hei cries, will even hisi mother ignore?
(6.91) Whati , given that iti had been discussed at the previous AGM, did
itsi proponents decide not to put on this year's agenda?
(6.92) Which horsei, given that iti was badly injured in the recent National,
did itsi owner reluctantly decide not to race again ?
(6.93) Whosei dissertationj , if theyi complete itj in time, will you
recommend for the prize?
According to the representational account of pronominal construal advocated
in the present framework, this is as it should be. Any accessible representation
of the right logical type should be suÆcient as a substituend for the U meta-
variable projected by a pronoun; and the semantic weakness of a WH meta-
variable is not grounds for its unsuitability as antecedent for a pronoun.
As these examples also show, the availability or non-availability of a left-
peripheral wh expression as antecedent for a pronoun within a following subject
position does not seem to be structural either, but is rather due to a medley
of pragmatic factors. First, wh questions in which the pronoun takes the wh
expression as antecedent are more acceptable if the wh expression conveys the
implication that the construct being questioned is one of a set of individuals
already established in the discourse context, as in questions with which:25
(6.94) Which painting did its owner complain was underpriced?
Secondly, the explicit addition of modifying material to a wh expression substan-
tially increases its viability as an antecedent for some subsequent pronominal {
compare (6.96) with (6.95):
(6.95) What type of car would its owner never part with?
(6.96) ??What would its owner never part with?
Such increased specication of the restrictive predicate for some WH meta-
variable does not need to be made explicit in the wh form itself. Interpreted
against an appropriate context, a question containing an unmodied wh ex-
pression may substantially improve in acceptability. Thus, (6.96) is consider-
ably more acceptable construed against a context in which what is as in (6.97)
construed relative to a contextually restricted set of entities:
(6.97) A: Some cars seem to give endless trouble, but there are a few jewels
that people get irrationally fond of, once they own them.
B: What would its owner never part with?
A: An XJS, to begin with. Then, the old long-wheel-base
Landrovers.
25 Questions with which are commonly said to presuppose that there is an entity of appro-
priate type (see Lappin and Reinhart 1988).
And this phenomenon cannot merely be put down to some quasi-referential

property of a generic form of construal of the wh expression, for it also applies
to wh expressions which are not taken to be questioning properties of some type
of object:
(6.98) Whati personal attack on Blair did itsi perpetrator get away with?
As (6.98) shows, examples in which the sole plausible antecedent for the pro-
noun is the WH meta-variable projected by the wh expression are notably more
acceptable than ones in which the salience of the pronominal-antecedent rela-
tionship is less marked:
(6.99) What committeei did itsi chairman eventually admit had made a
succession of serious errors?
(6.100) ?What girli did heri mother annoy?
Given that there seems to be neither a semantic nor a structural basis for this
hesitation over using a wh expression as an antecedent for a subsequent pronom-
inal, we might again turn to a pragmatic theory such as Relevance Theory
to explain how and why these various factors enhance the accessibility of the
pronominal antecedent relation. There are a number of possibilities to explore,
any one of which appears to contribute to the accessibility of the antecedent of
a pronominal. One hypothesis is that selection of antecedent for the pronomi-
nal is driven by the availability of putative inferential eects relative to which
the choice of value can be conrmed { in other words, driven by the pressure
to establish the relevance of an interpretation. Such an account is suggested
by the apparent increase in acceptability with greater predicative content of
the antecedent, providing an enriched context for assessing the relevance of the
choice.26 The increase in acceptability of examples when either there is no
other antecedent than the pronoun within the following subject or there are
specic eects which cannot be achieved otherwise further suggests that the
relevance-theoretic concept of relevance (with its combination of accessibility
considerations weighted against suitable inferential eects) is a fruitful direc-
tion in which to search for an answer. But without a formal understanding of
the process of context construction, which we have not provided here, we cannot
do more than again note the potential of relevance-theoretic assumptions.
It should be noted that this analysis allows the presence of pronouns resump-
tively construed in wh questions: nothing will prevent Merge applying to the
pair of a xed tree node annotated with a formula F o(U) and the unxed node
annotated with F o(WH). Yet resumptive pronouns in questions are judged to
be much less acceptable than resumptive pronouns in relative clauses:
(6.101) ?*Who does Sue think he's a selsh idiot?
26 The checking for inferential eects is presumed to involve retrieving additional assump-
tions from stores associated with the values of the F ormula predicate within the propositional
structure under construction, the association of a store with a F ormula expression constitut-
ing the content of that expression (see Sperber and Wilson 1995, Carston 1996, Breheny 1999
for this form of characterization of a concept).
(6.102) ?*Who has the Committee decided they are going to dismiss him?
However, as in relatives, the acceptability of a pronoun co-indexed with a clause-
initial wh expression is enhanced if the use of the pronoun is essential to achieve
the specic intended eect,27 and accordingly we treat these as well-formed:28
(6.103) ?Who did Sue say only that he should not be left o the short list?
(6.104) ?Who did Sue insist at least he must be interviewed?
(6.105) ?Who did Sue say that if necessary he could be left o the short list?
(6.106) ?Whose upgrade seminar did even he admit it had been an abject
failure?
(6.107) ?What did Sue tell her mother it shouldn't be cooked for longer than
an hour?
(6.108) ?What type of car would you suggest that the makers should have
taken it o the market years ago?
(6.109) ?Which results did Sue say she was checking whether they could be
conrmed?
It might be objected that, in positing an analysis in which a pronoun can nd
an antecedent either through Substitution or through Merge, we have done little
more than mimic the very same ambiguity between a base-generated pronominal
and some morphological realization of trace29 (or in HPSG terms, the analysis
of a pronoun as being replaced by a SLASH feature). Supercially this might
seem true, but the dierence lies in the fact that, in the present account, the
input provided by a pronominal is unitary { a meta-variable { and the processes
of Merge and Substitution simply provide dierent update processes whereby
underspecied aspects of tree descriptions can be further specied.
In presuming that Merge can provide a value for a pronominal, we are pre-
suming no more than the assumption that a formula at an unxed node in a
tree description can provide a means of assigning an interpretation to a pro-
noun (in this sense being its antecedent).The unifying basis for the explanation
27 Particles such as only, even, etc. cannot be used to modify the wh form in the left-
peripheral position, a restriction which is mysterious.
28 It has been suggested to us by Ronnie Cann (personal communication) that in all cases
where pronouns appear to be taking the wh expression as antecedent, the construal of the
pronominal is in fact constructed from some representation of an entity in some sense inde-
pendently identied. However, though many such uses involve what have been reported as
D-linked interpretations (Pesetsky 1987), this is not an essential property of such construals.
Furthermore, in cases such as (6.103){(6.109), where Merge must be allowed to take place,
no such contextually provided interpretation of the pronominal is possible, as the application
of Merge would then have the eect, incorrectly, of replacing the WH meta-variable by this
contextually determined value.
29 The distinction between pronouns as base-generated and as re exes of movement goes
back to Ross (1967).
of anaphora lies in the form of input a pronominal provides to the interpretation

process, and in the pragmatic assignment of an antecedent.30
6.2.2.1 Pragmatic Explanations and Language Variation
The previous discussion might seem to miss what many might take to be the
central problem posed by weak crossover phenomena and resumptive pronoun
construal, which is the considerable cross-language variation, a variation which
seems to demand a system-internal explanation if we grant that pragmatically
driven accounts, being entirely general, are unsuited to capture any such vari-
ation.31 However, such a swift rejection of a pragmatically based explanation
is not warranted. To the contrary, the present framework provides a rich basis
for anticipating variation in resumptive pronouns, despite leaving the account
of when such interpretations are licensed to pragmatics. The variation can be
seen as stemming from dierent lexical forms in the individual languages and
the dierent range of tree structures for which resumptive pronouns can provide
a suitable update.
The rst basis for cross-linguistic variation in resumptive construal of pro-
nouns is that pronouns may take one of two forms, either a full or a clitic form.
One distinguishing feature of clitic pronouns, in all languages that display them,
is that they can never be phonologically stressed, with the consequence that,
unlike the corresponding full form, no clitic pronoun can provide a particular
contrastive eect.
The second basis for variation lies in the two possible analyses of left-
peripheral noun phrase sequences that we have proposed. Left-dislocated con-
stituents may either be taken to annotate the root node of a tree (of type e) to
which some following tree is LINK ed, or they may be taken to annotate some
unxed node within a single tree. Not all languages freely allow the rst of these
constructions { English, for example, does not, requiring the signal as for for
30 In assuming that Merge can provide a value for a meta-variable, and so license resump-
tive uses of pronouns in wh questions, we have presumed that pronouns lack any restriction
that their annotation must decorate a terminal node of a tree, unlike other lexical items. An
alternative analysis would be to retain this restriction, so that the formula selected, however
complex, must itself be a decoration on a terminal node without dominating any further sub-
structure. This restriction would distinguish the licensing of resumptively construed pronouns
in relative clauses and questions, since the LINK ed structure projected by a relative sequence
involves the construction of a copy of a formula from the head node, hence a single formula
satisfying the terminal node requirement. In wh questions, however, the unxed node may be
of arbitrary complexity, so application of Merge which led to update of a pronominal meta-
variable by a branching structure would be debarred by the terminal node restriction. In
order to predict the unacceptability of all resumptive pronominal construals in questions, this
analysis would require re-analysis of who as projecting a complex structure with determiner
and common noun nodes (much as Japanese nominals). However, this analysis would predict
the unacceptability of (6.103){(6.109), and in the light of their relative acceptability, we leave
this option on one side as an alternative to be explored.
31 Witness to this conviction is provided by the persistent attempts to characterize the
various restrictions on weak crossover eects in structural terms, even when this leads to
the need to typologically distinguish speakers of an individual family (brother and sister) as
suggested in Bresnan 1998.
any such construal:
(6.110) As for Samantha, she annoys me.
Many others do: for example, Romance and Semitic languages, as we've already
seen, with the clitic left-dislocation structures (section 4.3.3.2). An initial noun
phrase construed as a LINK ed structure will induce the obligatory presence of
some anaphoric expression within the following clausal sequence to ensure the
required copy, so any language which has clitic pronouns will be able to use
them for this purpose. There is no question of such resumptive pronouns only
being motivated by the need to obtain contrastive eects, as in this type of
structure resumptive construal of some anaphoric expression is the sole means
of creating the correct logical form.32
A third basis for variation is that pronouns themselves may be licensed to
project an annotation only for an unfixed node { this is arguably displayed in
Romance languages, in which clitic pronouns obligatorily display a local scram-
bling phenomenon, which we might analyse as projecting an annotation to a
most-locally unxed node (see Radford 1977):33
(6.111) Paolo glielo vide far cadere. [Italian]
Paolo to-him it saw made drop
`Paolo saw him drop it.'
This form of analysis provides a potential explanation for cross-linguistic vari-
ation in the resumptive use of clitic pronouns. The coincidence of a pronoun
projecting an annotation on an unxed node and having a reduced phonological
form (hence disallowing any contrastive eects) will ensure that the only possible
role for a clitic pronoun in such a language is to full an outstanding require-
ment for some particular Formula value. On this basis we can anticipate the
dierence between Italian and Romanian relative clauses, that the latter allow
{ indeed require { the presence of a pronoun construed resumptively, whereas
the former do not (see 4.3.1). The LINK transition in Italian, recall, we anal-
ysed as leading to the projection of an unxed node which the relative pronoun
annotates, whereas in Romanian, the LINK transition is dened as imposing
only a modal requirement on the root node of the LINKed structure:
(6.112) *l'uomo che l'aprezzavamo [Italian]
the man that we appreciated him
(6.113) baiatul pe care l-am vazut [Romanian]
the boy pe which him have1:SING seen
32 As noted in Anagnostopoulou et al. 1997, full forms of noun phrase can be used in such
structures to achieve the required result of a common formula in the pair of LINKed structures.
33 Such an account requires complementary distribution of clitic pronouns and full forms
of noun phrase, and this can be ensured by dening the required locally unxed node to be
projected only by lexical action as dened by the clitic pronouns. Notice that this allows
individual clitic pronouns to vary. In this connection, see Monachesi 1998.

The dierence between annotating a xed node and annotating one that is
initially unxed provides a means of explaining the optional use of resumptive
pronouns in some languages. If the clitic pronoun is in a position in the string
from which it will annotate a xed tree node position, then it may serve to
provide a xed tree node position for some hitherto unxed but annotated node;
but in such a position in the string, its presence will not be obligatory. The
Semitic languages allow the presence of a clitic resumptively construed with a
left-dislocated expression; and in standard Arabic, this includes wh question
forms which are the equivalent of who (data from Demirdache 1991, 1997):
(6.115) mai ra?ayta-(hui )?
p [Classical Arabic]
Who saw you (him)
`Who did you see?'
On the assumption that in (6.115) the wh form annotates a node which is
initially unxed within an initiated tree, we would anticipate, correctly, that
such clitic forms cannot be used with this form of construal when they occur in
a relative clause:
(6.116) *mani ras?ayta l-fataata llatii Æarabat-hui
p
who saw you the girl that hit him

`Who did you see the girl that hit him?'
Then there is the contrary possibility that in some languages (in some struc-
tures), resumptive use of pronouns may be debarred altogether (as in Dutch),
with Merge taking only the more restricted Gap Merge form, and any resump-
tive pronoun occurring only as fullling the modal requirement imposed by some
LINK Introduction.
Finally, we would anticipate that, all such lexical dierences aside, the prag-
matic pressures which determine the limited extent to which a WH formula
annotating an unxed node is used as an antecedent for a pronoun would in-
deed be common across languages. We do not wish to suggest that this list
exhausts the possible sources of variation. It is merely intended as an indication
that the interaction between the options made available within a relatively lib-
eral tree growth mechanism and system-external constraints provides a fruitful
point of departure for a cross-linguistic investigation of the interaction between
the resumptive construal of pronouns and long-distance dependency eects.34
Given that our general aim throughout this book is merely to develop a frame-
work for natural language interpretation and demonstrate its application to a
34 In articulating a form of interaction between pragmatic and system-internal factors, this
account stands as an alternative to proposals advocating a considerable increase of complexity
within the grammar formalism itself. In this connection see Bresnan 1999 and the advocacy of
optimality-theoretic syntax, which necessitates the encoding of system-internal default rules.
Summary 221
range of syntactic puzzles, we leave the discussion of weak crossover phenomena,
without probing further the nature of the interaction between pragmatic con-
straints and particular lexical specications. Notice meanwhile that the concept
of well-formedness relative to which weak crossover data are explained, is one
in which pragmatic actions and constraints, though dierent in kind from the
formal operations of the computational/lexical system, nevertheless act, like the
computational/lexical update mechanisms, as a lter on logical forms. In so do-
ing, pragmatic actions and constraints conform to the general pattern that the
parsing process involves monotonically establishing annotated tree structures.
6.3 Summary
So what has this characterization of crossover phenomena in head-initial struc-
tures demonstrated overall?35 In chapters 2{3, we set up an account of the
interpretation of left-dislocated expressions in terms of a process of updating an
initially underspecied tree node characterization for the formula projected by
the dislocated expression. Anaphora was analysed in terms of a place-holding
expression, to be replaced by some independently provided formula. We charac-
terized relative clause construal as a process of inducing LINKed tree structures,
with an obligatory process of transfer of information from one tree to another.
And we gave a characterization of wh expressions in relatives as ensuring the
presence of the required copy on some unxed node in the LINKed structure,
and in questions as projecting a place-holder. From these assumptions, the re-
lated stories of crossover and resumptive pronoun construal have emerged, with
an analysis of a broad array of supercially heterogeneous data, for which we
have not posited any additional construction-specic stipulation.
The signicance of the fact that the value assigned to the pronominal meta-
variable can be provided by a formula at some unxed node through an appli-
cation of Merge should not go unnoticed. This interaction between pronominal
construal and long-distance dependency is only expressible if both of these pro-
cesses are dened as resolving structural underspecication within the develop-
ment of annotated tree structures. Characterizing the assignment of content to
a pronoun in model-theoretic or other denotational terms, to the contrary, does
not provide a basis from which to articulate this interaction. It is accordingly
only with a representationalist stance that we can expect to characterize all uses
of pronouns as variants of the same phenomenon. Thus we arrive at the stance
we took as the initial motivation for developing this framework in concluding
chapter 1.
35 In head-nal structures, with the absence of any anticipatory feeding relation from the
head into the projection of unxed node within some LINKed structure, the constraints on
pronominal construal within such relatives cannot be subject to the same form of explanation.
In the main, the only type of crossover observed in head-nal languages is that displayed by
His mother loves everyone, which we have indicated should not be analysed in crossover terms,
but rather as a linearity phenomenon. However, see Kempson and Gabbay 1998 for discussion
of crossover phenomena in Chinese, which displays a mixed system, in part like Japanese, in
part like English.
One caveat is in order by way of a coda. The apparent reliance on the

linearity of processing in the explanation of crossover phenomena has been a
consequence of the parsing dynamics. It is not a property which has had to
be stipulated. This allows for the reduction of sensitivity to linear order if the
value for an anaphoric expression is determined solely relative to properties of
the resultant logical form rather than on the process of establishing that form.
Arguably this is true of anaphors such as herself (see Reinhart and Reuland
1993, Hepple 1990, and many others). In this connection, Bresnan (1998) has
claimed that in Bantu languages at least some crossover phenomena have to be
explained in terms of a hierarchy of semantic/thematic relations. Early on, we
noted and set aside lexical anaphors such as herself as items whose constraints
on interpretation had to be dened over the constructed formulae (actually
in some cases more indirectly over required patterns of inference from some
resultant formula { see chapter 4, note 40). So a full account of the interac-
tion between wh expressions and anaphora of this type remains to be given.
Nevertheless, we take the signicance of crossover phenomena and pronominal
construal in English, Arabic, Hebrew and Dutch to lie in the demonstration
that problematic cross-linguistic variation can be explained in terms of the pro-
gressive accumulation of tree structure descriptions on a left{right basis.
7
Quantication Preliminaries

7.1 Introduction
Throughout the previous chapters, we have argued that the expressions under
analysis are not quantied expressions. First, wh relativizing elements were
argued not to be quantifying operators but to be quasi-anaphoric expressions.
Pronominal expressions were also argued to supply substitution sites and not
quanticational expressions in their own right (contra e.g. Neale 1990, Breheny
1999). Then even wh question words were argued to project meta-variables,
with no quantifying force of their own, even in clause-initial position (contra al-
most everybody, though see Ginzburg and Sag 1999).1 In making these claims
we have presumed that noun phrases project annotations of type e, an as-
sumption which might seem to y in the face of the numerous formal semantic
arguments that a homogeneous characterization of natural language semantics
must assume that noun phrases are all assigned the generalized quantier type
(e ! t) ! t, stemming back to Montague (1974a, 1974b). This idiosyncrasy
now needs to be addressed explicitly, bearing in mind that natural language ex-
pressions systematically underspecify denotational content, so that diversity of
denotational content is not a contra-indicator of the most appropriate analysis,
as long as there is a unitary basis for this diversity. We need to see what kind of
asymmetry there is between the actions projected by quantifying determiners
and the resulting logical structures, and how the lexical inputs to be dened for
the various determiners interact with the construction of annotated trees. More-
over, we need to show how such an account of quantication interacts with the
substitutional account of anaphora to yield bound variable and E-type eects,
in order to provide a unied and structural account.
By way of preliminary, we might note that the plan to construct an account in
terms of expressions of type e, rather than some higher type, is based on the intu-
ition that a formal system for compiling natural language interpretation should
1 W h-in-situ expressions have been analysed as a variable, but they are then taken to be
distinct from wh-initial expressions { see Pesetsky 1987, Aoun and Li 1993, Berman 1991.
223
224 Quantication Preliminaries
be grounded in some counterpart to the logical concept of arbitrary name. The

motivation behind this adopted desideratum is a parallelism between anaphora
and scope eects for a certain subset of quantifying determiners. These are the
indenite determiners, all expressions interpreted as an existential quantier,
in English the set comprising the singular indenite a, the determiners some,
many, a few, many a(n) X, the numerals and the quantiers constructible from
at least X, at most X, for numbers X. The parallelism between anaphora and
scope of indenites lies in the fact that, in both, an antecedent has to be found.
In the case of anaphora this is the referent; in the case of indenites this an-
tecedent is some term to depend on. In both cases, the process of construal of
the antecedent is free and subject to pragmatic constraints. If this parallelism
between indenites and anaphoric expressions is to be captured, then there are
two consequences for the characterization of indenites. It will require, on the
one hand, that we take the logical terms projected by indenite noun phrases
to be names in a sense to be dened, and on the other hand, that the choice of
term to depend on is given by some pragmatic action of substitution. This is
indeed what we shall propose in what follows.
Our rst step will be to show that the free choice of interpretation particular
to indenite noun phrases cannot, as is sometimes argued, be reduced to a lex-2
ical ambiguity between a referential (or specic) and a quanticational sense.
We also show that the relation between such a quantifying expression and other
expressions in its scope is not subject to restrictions familiar from long-distance
dependency, as might be expected from any analysis of quantier construal in
terms of covert movement or other mechanism correlating it with long-distance
dependency, but is rather to be handled analogously to anaphora resolution as
involving an essential element of choice. Then we shall set out our alternative
analysis, which explicitly re ects the incremental process of establishing a log-
ical form, taking into account the fact that scope relations between terms can
generally not be determined until the full propositional structure is compiled.
We will show how it is possible within the system to dierentiate between dif-
ferent types of quantier; and we shall give a brief demonstration of how the
model extends naturally to expressing E-type anaphora eects.
7.2 Scope Eects and Indenites
The point of departure is to recall that until quite recently, it has been stan-
dardly assumed that the ability of a quantifying expression to take scope over
a domain considerably larger than is apparent from the surface position of the
quantier, should be reconstructed in the same terms as long-distance depen-
dency { that is, in terms of a binding of a variable by a binding operator arbi-
trarily far removed { essentially in the manner of predicate logic quantication.
In the case of the GB paradigm, such quantifying expressions were presumed
to have moved covertly, as part of mapping onto a level of LF by a process of
2 This was originally argued for by Fodor and Sag 1982, but has been copied by many others
since, e.g. Diesing 1992, Szabolcsi 1997.
Scope Eects and Indenites 225
Quantier Raising (QR).3 In the case of HPSG, the characterization is, at least
in some analyses, presumed to involve feature percolation, a feature which is
discharged at some arbitrarily higher position in the conguration, by a process
of abstraction, matched by a semantic process of Quantifying in (see Pollard
and Yoo 1998). The process of Quantifying In familiar from Montague (1974b)
presumes that all noun phrases are generalized quantiers of type (e ! t) ! t,
and are subject to a semantically dened process of storage with a process
of abstraction at arbitrary points, binding an individual variable to create a
predicate to which the generalized quantier can apply. Categorial grammar
formalisms dene the proof-theoretic analogue of this semantic device, with the
construction of a premise at the extraction site, which is abstracted over at some
arbitrary point in the proof sequence by conditionalization to yield a predicate
type appropriate for combining with the generalized-quantier type (see Mor-
rill 1994). The primary exception to this general style of analysis for natural
language quantication has been Discourse Representation Theory, at least in
its original conception, with its characterization of universally and existentially
quantied expressions in terms of discourse referents; but even these have been
supplemented in more recent variants with generalized quantier characteriza-
tions in which the name-like property of the quantifying expression is no longer
apparent (see Kamp and Reyle 1993). In all of these frameworks, the charac-
terization of quantier scope construal is essentially general, leaving no room
for lexical variation, or for under-specication of interpretation. This is even
true of the discourse referent analysis of indenite expressions in DRT, for, ac-
cording to this analysis, the individual properties of the determiner are due to
the embedding relations of the DRS as a whole, and not to properties of the
determiner itself.
However, there is evidence that quantifying expressions in natural language
fall into two distinct classes according to their freedom with respect to scope
construal (we shall see in due course that there is considerably more varia-
tion than this). There are quantifying expressions whose scope relations are
determined by their position in the natural language string, and do not war-
rant unrestricted processes of quantier storage/LF-movement. These are the
non-indenite quantiers, such as every, most, few, almost no, whose scope
relations generally follow the sequence of words:
(7.1) Almost no student has read every book by Chomsky.
(7.2) No more than one examiner marked most scripts.
(7.3) Few students attended most concerts.
(7.4) Every student there is certain that almost nobody has read a book by
Chomsky.
(7.5) John wrote an article which few friends of mine had heard of.
3 For a detailed discussion of the problems arising for analyses which incorporate a process
of QR at LF, see Reinhart forthcoming.
Thus, in (7.1), there is no reversed scope interpretation for the two quantifying
expressions, though in (7.2) and (7.3) such interpretations are marginally pos-
sible. Wide scope interpretations are not possible for either almost nobody in
(7.4) or for few friends of mine in (7.5).
On the other hand, there are the indenites, which may all be construed
as taking arbitrarily wide scope with respect to the environment within which
they are contained, their order relative to other expressions in the string, in
consequence, providing little clue as to how they are to be interpreted:
(7.6) Two of my colleagues wrote an article which a friend of mine
complained about.
(7.7) John is planning to have several parties in which we all celebrate the
work of two Dutch colleagues.
(7.8) John went to see a lm and in it many friends of mine were portrayed
as idiots.
(7.9) John was visiting his mother while two painters I know were stripping
his house.
It has been suggested that the very wide scope eects associated with indenite
determiners are evidence of ambiguity between a sense in which the indenite
acts as a referring expression, its so-called specic use, uniquely picking out
some entity which the speaker has in mind (Fodor and Sag 1982 and many
authors following them) and a regular existentially quantied term, and that
it is the uses of indenite noun phrases as names which give rise to the ap-
parently unconstrained process of quantier construal. However, it has been
demonstrated by a number of people (see Farkas 1981, Cormack and Kempson
1991, Abusch 1994, Reinhart 1997) that intermediate interpretations are avail-
able which cannot be explained by any such supposed discrete referring term. In
these intermediate cases, the indenite can be construed as having wide scope
with respect to the relative clause, but narrow scope with respect to some higher
quantifying expression elsewhere in the clause:4
(7.10) Every professor rewarded every student who read a book he had
recommended.
(7.11) Each student has to come up with three arguments that show that
some condition proposed by Chomsky is wrong.
The assumption, furthermore, that the process of quantier construal can be
analysed as a process of long-distance dependency, in the same terms as move-
ment, feature passing or assumption discharge, faces the additional puzzle that
the indenite determiners do not display any of the constraints associated with
4 Cormack and Kempson (1991) argue that the apparent implication that the indenite is
construed referentially is never more than a conversational implicature.
long-distance dependency; (7.6){(7.9) demonstrate this, with the indenite be-
ing inside a relative clause in (7.6) and (7.7), inside the second of two conjuncts
in (7.8), and inside an adverbial clause in (7.9).
If, however, we approach the phenomenon of quantier scope from the per-
spective of incrementally establishing a logical form on a left{right basis, a
dierent range of solutions becomes expressible. Indenites dier from other
quantifying expressions in being able to freely depend on a distinct term of suit-
able type, a freedom redolent of anaphora. A noun phrase { that is determiner
and nominal together { project an annotation on a node of type e; and these
actions may include the projection of constraints on the eventual scope relations
of the form t1 < t2, `term t1 has scope over term t2'. Such scope statements
have to be collected at root node (requiring a formula of type t), where they
must fully determine the semantic evaluation of the resulting formula. Inde-
nite determiners, for example, can be analysed as imposing a constraint that one
argument of the scope relation is a meta-variable, its value to be identied as a
matter of pragmatic choice. Bearing in mind that every propositional formula
is of the form F o(Si : ), with Si a term denoting the index of evaluation for
, we can analyse indenite noun phrases as taking narrow scope with respect
to that term Si when they take wide scope with respect to other quantifying
expressions, if there are any.5 This enables us to say that indenites always
take narrow scope with respect to some term, that term decided by a prag-
matic choice. In (7.12) there is only one choice, this being that the indenite is
interpreted relative to Si, the term representing the index of evaluation:
(7.12) A student fainted.
Notice that, on this view, scopal relations between terms are projected directly
as relations between terms annotating the structure, and not as a relation be-
tween nodes in the structure.
The advantage of this view is that it classies pronoun antecedent construal
and choice of scope construal for indenites together, as involving a pragmatic
choice. Like anaphoric expressions, the scope of indenites is, as we've seen,
not sensitive to island restrictions. Nevertheless, like the construal of anaphoric
expressions, there is strong preference for the scope of the indenite to be es-
tablished on the basis of linear order. Thus mixed quantication sentences are
reported to be ambiguous with two scope-distinguished interpretations if they
contain an indenite following some other quantifying expression:6
5 Expressing the apparent independent scope readings of indenites in these terms is ap-
proximate only, as there are some construals in which the indenite has to be construed as
taking narrow scope with respect to the index, characteristically selected as the second ar-
gument of the T e relation constructed from tense. Pending an explicit account of tense and
modality, we leave this detail on one side. See Farkas 1997 and Kempson 1996 for argu-
ments that construal of indenites may be dependent on temporally sorted variables used to
represent construal of tense/mood.
6 The only exceptions are examples like a certain which may force a wide scope construal
of the indenite (see section 7.3.2.3).
(7.13) mei ge ren mai le yi ben shu [Chinese]

every classifier man buy asp one classifier book
Èvery man read one book.'
In contrast, the interpretation of the indenite relative to what follows is much
more restricted, and subject to cross-language variation. In particular, though
English allows an interpretation of mixed quantication sentences in which an
indenite is interpreted as taking narrower scope than some following quantied
expression, Chinese reportedly does not:7
(7.14) yi ge ren mai le mei ben shu
one classifier man buy every classifier book
Òne man read every book.'
Moreover, it is notable that what looks supercially like a construction-specic
restriction on scope is in fact echoed by anaphora resolution, for pronoun con-
strual patterns along with indenite construal in Chinese in disallowing the
antecedent of a lexical pronoun to be some following quantied expression, even
in a structure where this would be natural, for example with a relative clause
containing a pronoun to be construed as bound by the nominal head (see Kemp-
son and Gabbay 1998):8
(7.15) hushi ta muqin de mei ge ren dou mei
ignored hei mother rel every classifier manj all not
kao jige.
test pass
Èveryonej who ignored hisi mother failed the test.'
(7.16) xiang-xin Mary chong-bai ta de mei-yi-ge ren dou
believe Mary admires himi rel every manj all
cuo-le
mistaken
Èveryonej who believes Mary admires himi is mistaken.'
This similar patterning of pronoun and indenite scope construal as generally
following a left{right restriction suggests that in a language such as Chinese,
both anaphora resolution and indenite scope construal are subject to a similar
restriction that the value of the meta-variable must be selected from the terms
7 It has sometimes been argued, in the face of the reported non-ambiguity of examples such
as (7.14) that the ambiguity of (7.13) is not a true ambiguity, since there is an entailment
relation between the two readings; but, as has been demonstrated by more than one author (see
Winter 1997, Reinhart 1997, Cormack and Kempson 1991), there are many cases where the
dierentiated interpretations are logically independent, so no such conclusion can be sustained
(see the detailed discussion in Jiang 1995 about the relevant data in Chinese).
8 Not quite all our informants reported the asymmetry between (7.14) and (7.13), but if
they reported that (7.14) allowed a reversed scope interpretation, then they consistently also
reported that (7.15) and (7.16) allowed interpretations in which the pronoun is construed as
a bound variable.
which have been introduced previously into the tree, a restriction which is not
encoded in languages such as English.9
The parallelism between anaphora construal and construal of indenites is
sustained even in languages where such a linearity restriction on scope choice
cannot be lexically imposed. For example, in German subordinate clauses, the
noun phrases precede the clause-nal verb. In these structures, an indenite
expression initial in a subordinate clause preceding some quantifying adverbial is
generally not interpreted as within the scope of the following adverbial, whereas
indenites that follow the adverbial are freely interpretable as taking either
narrow or wide scope with respect to that adverbial:10
(7.17) weil ein Assistentsarzt zweimal einen Patienten von
because a doctor twice a patient from
einer gefahrlichen Krankheit geheilt hat. [German]
a dangerous disease cured
`because one doctor twice cured a patient of a dangerous disease'
(7.18) weil zweimal ein Assistentsarzt einen Patienten von
because twice a doctor a patient from
einer gefahrlichen Krankheit geheilt hat.
`because twice a doctor cured a patient of a dangerous disease'
(7.19) weil von einer gefahrlichen Krankheit zweimal ein
because from one dangerous disease twice a
Assistentarzt einen Patienten geheilt hat.
doctor a patient cured
`because from a dangerous disease one doctor twice cured a patient'
So in (7.17) ein Assistentsarzt is reported to be construed as taking wide scope
with respect to zweimal, whereas both einen Patienten and einer gefahrlichen
Krankheit freely allow interpretations with wide or narrow scope with respect
to zweimal. In (7.18) all three indenites can be interpreted in both ways, and
in (7.19) it is einer gefahrlichen Krankheit that is restricted to take wide scope
with respect to the adverbial. However, the restriction does not appear to be
structurally denable, because if two indenites precede the adverbial in such
constructions, the second may take narrow scope with respect to the adverbial
which immediately follows, as in (7.20), which allows an interpretation in which
one doctor twice cured a patient of some dangerous disease (dierent patient
and dierent disease on each occasion):
9 This restriction on the interpretation of indenites as ambiguous only if they follow some
other quantied term is very widespread, though it is not often expressed in the terms sug-
gested here (see Hoji 1986 for a report of similar data in Japanese).
10 These data are, we believe, due to Hans Kamp.
(7.20) weil ein Assistentsarzt einen Patienten zweimal von

because a doctor a patient twice from
einer gefahrlichen Krankheit geheilt hat.
In this very type of circumstance, anaphora resolution also may be determined
against the linear ow of words, allowing an interpretation of the pronominal
ihren in (7.21) as bound by the quantifying expression that follows:
(7.21) weil ein Assistentsarzt ihren Krankheitsverlauf zwei
because one doctor their disease-progress two
Patienten zeigte
patients showed
`because one doctor showed two patients their hospital record'
The claimed parallelism between anaphora and indenite construal is not to
deny the existence of idiosyncratic restrictions. As is well-known, pronoun res-
olution is subject to a locality restriction, which indenite construal is not.
Conversely, there are restrictions on indenite construal that do not apply to
anaphoric expressions. An indenite can only stand in a scope relation to other
expressions in the string: no such restriction applies to anaphoric expressions
(which freely allow so-called indexical construal). There is, furthermore, lexi-
cal variation between types of anaphoric expression, and so, equally, there may
be varying lexical constraints on indenites, both within and across languages.
Nevertheless, modulo such item-specic restrictions, the interpretation of pro-
nouns and indenites alike appears to be driven by system-external pragmatic
considerations.11 There thus seems to be broad cross-linguistic motivation for
analysing indenite construal and anaphora resolution in similar terms, as im-
posing constraints on the construction of some suitable logical term, which in-
volves an essential element of choice. On this view, the ambiguity of inter-
pretation of a string such as (7.11) (repeated here) is the result of a choice
of scope { that the term projected by some condition proposed by Chomsky is
to be construed either relative to three arguments or relative to each student
or independent of either and dependent on the temporal index Si of the root
clause:12
11 It is of some interest that the one structure in English where construal of indenites and
pronouns appear to diverge, the subject position, is an environment where expletive pronouns
are licensed in English. Thus, alongside (i), where the indenite may have narrow scope with
respect to the subsequent quantied expression, there is (ii), with an expletive subject pronoun
which has to be interpreted as identied with the clausal complement within the predicate:
(i) A nurse interviewed every patient.
(ii) It is essential that every patient get well.
We do not pursue this here, since we have not given lexical denitions for expletive pronouns as
in (ii), but we note none-the-less that the position of subject appears to be the only exception
in English to what is otherwise a marked parallelism between pronominal and indenite scope,
suggesting that this pair might be seen as part of the general pattern.
12 In addition there is the choice of temporal index for the LINK ed structure projected by
the relative, but this would not in this case give rise to a distinct interpretation.
Quantication 231
(7.11) Each student has to come up with three arguments that show that some
condition proposed by Chomsky is wrong.
The additional advantage of assuming, more generally, that all quantied
expressions project terms of type e, with scope statements to be accumulated,
is that it makes it possible to deal with quantied expressions in an incremental
process of constructing a logical form despite the fact that scope relations can
be dened only globally. Accordingly, we dene a system in which an individual
determiner projects a lambda term involving a quantifying operator indicat-
ing the quanticational force of the term to be constructed, and the common
noun projects a requirement for some scope statement at some suitable domi-
nating node requiring type t. For some determiners the arguments for this scope
predicate are xed by rule, strictly following order in which the arguments are
presented in the string. Others, specically the indenite determiners, project a
scope statement of which one argument is a meta-variable, which gets replaced
by a pragmatic substitution process. The scope statements, accumulated at a
root node, are then used to assign full scope to the propositional formula at that
node. We now turn to articulating this informally sketched proposal in more
detail.
7.3 Quantication
Even before setting out this account, we have a caveat: it is not our object
to give an in-depth analysis of the semantics of quantication. We are happy
to accept the analyses and formalizations oered in the abundant literature on
this subject. There is also a sense in which we are barely even sketching a
theory of quantication, since we are not addressing the problems of plurals
and the various problems of group and distributive interpretations which those
induce. There is, however, a particular problem that is posed by the left-to-
right constraints of a parsing perspective. We will have to incorporate the
information supplied by quantied noun phrases into the tree before their scope
relations to other noun phrases and to modalities can be determined. This is
a consequence of the fact that we take natural language strings to supply their
information in a sequential order. We therefore want to be able to collect any
restrictions on scope between quantifying expressions as they become available
in the course of a parse through the string. Our analysis of quantication then
falls into two parts. First, we analyse quantied NPs as structured objects of
the entity of type e. Secondly, we assign scope relations to these structured
objects.13 Quantied NPs, like proper names, project values of the Formula
predicate belonging to the domain DF o, of type e. These NP representations
13 The assumption that relative scope statements are collected in the construction of a fully
annotated tree is similar to that dened in Minimal Recursion Semantics (see Copestake et
al. 1998), though in their analysis, the concept of òutscope' (expressed by `<') extends to
all terms, e.g. every outscopes dog, and the progressive collection of such scope statements is
exclusively on a bottom-top basis, with no intended correspondence to a left{right process of
interpretation.
will be formed using variable binding term-operators. For instance, the phrase
some man will be represented by
(; x; Man(x)) 2 DF o:
This quantied NP can be completely described by xing four parameters:
1. The Binder (e.g. ) indicates the mode of quantication (i.e. existential).
2. The Variable (e.g. x) indicates the variable being bound by the binder.
3. The Restrictor (e.g. Man (x)) indicates the binding domain of the vari-
able. By their restrictors, objects are introduced in a discourse; and pred-
ications on these introduced objects are then constructed from items more
standardly seen as falling within the nuclear scope of the determiner.
These three features exhaustively specify quantied NP denotations in isolation.
However, to characterize NPs as occurring within a sentence, a scope statement
is required.
4. A Scope statement is an atomic DU proposition, using the scope relation
< DF o;e DF o;e , of the form
x<y
stating, in this case, that the quantier binding x has scope over the
quantier binding y.
Once these four features for quantied NP representations are xed, the se-
mantic, i.e. truth-functional, interpretation of a clause projection is fully deter-
mined. With a given Formula value of type t,
= [(1 ; x1 ; 1)=x1 : : : (n; xn ; n)=xn ]
where the (i ; xi ; i ) displayed are the projections of all quantied NPs occurring
in a clause, each with a unique variable xi , we can associate a strict partial order:
B = hB; <B i
where B fx1; : : : ; xn g and <B B B is an irre exive, transitive relation on
B. Such an ordering re ects a choice of scope constraints on the terms occurring
in . Consider a basic mixed quantication example such as (7.22), to which
we shall return several times during the course of the chapter:
(7.22) Every dog ate a biscuit.
On the interpretation re ecting the order of the words in the string, we have
the pair:
(B; );
Quantication 233
where
= Ate((; x; Dog(x))(; y; Biscuit(y)))
<B = fhx; yig
giving rise to the same truth-functional interpretation as
8x(Dog(x) ! 9y(Biscuit(y) ^ Ate(x; y))):
The scope relation x <B y; is interpreted to mean that the quantier binding
x has scope over the quantier binding y. So, if we have the reversed scope
constraint
= Ate((; x; Dog(x))(; y; Biscuit(y)))
<B = fhy; xig,
then this has the interpretation
9y(Biscuit(y) ^ 8x(Dog(x) ! Ate(x; y))):
The pairs (B; ) used in this example are of a special kind in two ways:
all terms involved in occur in some scope relation,
the terms are (strictly) linearly ordered by the scope relations (in this
case, trivially so, because there are only two terms).
In this chapter our discussion of quantication will restrict itself to sentences
exhibiting this special structure. All formulae of rst-order predicate logic (and
non-branching quantication in general) give rise to pairs (B; ) of this special
kind.
So, our analysis of quantication requires that declarative units of type t
with formula = S : for S some temporal index, have a label giving a scope
sequence:
h: : : ; hS; x1 ; : : : ; xn i; : : : ; ti :
where the ordering S <B x1 <B : : : <B xn contains at least all bound variables
in plus the temporal index S . We now add a binary relation symbol `<' to
the DU alphabet such that xi < xj is an atomic DU formula for all xi ; xj 2
D [ MV , and we set:
Denition 12 (Truth Denition for Scope Statements)
Æ T [h: : : ; hS; x1 ; : : : xn i; : : : ; ti : ] j= xi < xj i
xi <B xj , or
= : : : Qxi : : : Q0 xj
, Qxi and Q0xj are unique with respect to
the variables they bind, and xi ; xj occur free in .
This disjunctive condition embodies the fact that x has scope over y, either
if x <B y is explicitly recorded in the scope label, or the quantier binding x
in has scope over the quantier binding y in . We will see that the latter
condition makes the statement x < y persistent under restructuring.
Having parsed the string which is (7.22), the decorations of the root node of
the resulting tree description may form the set
fT n(0); T e(S ); F o(Ate((; y; Biscuit(y)))(; x; Dog(x))); T y(t); : : : g [
f(S < x; x < y)g
That is, the scope relations, however projected, appear as atomic propositions
involving the bound variables on the root node. These propositions have been
collected in the course of the parse. Because every term will have been as-
signed a distinct variable, we can identify a term with its variable. The scope
propositions can now be used to algorithmically update the Formula value,
Ate(; x; Dog(x))(; y; Biscuit(y)), into the more familiar shape:
fT n(0); F o(S : 8x(Dog(x) ! 9y(Biscuit(y) ^ Ate(x; y)))); T y(t); : : : g:
In this update the quantier binding x indeed has scope over the quantier
binding y.14
We will now discuss the various components of our analysis in more detail.
7.3.1 Quantied NPs
In this section, we discuss the various sub-parts of noun phrase terms, and their
contribution to the interpretation of sentence-sized units. The structures which
we successively introduce are summarized by the display in gure 7.1, to be
presented in turn. As a way of talking about appropriate constructs, we will
dene some monadic predicates, abbreviating modal DU formulae evaluated at
a type e node:
NP
[h: : : ; ei : APL(P : ; (x; Man(x)))]
DET NOM
[h: : : cn ! ei : P : ] [h: : : ; cni : (x; Man(x))]
VAR RESTR
[h: : : ei : x] [h: : : (e ! cn)i : Z (Z; Man(Z ))]
Figure 7.1: The structure of quantied NPs

14 It is important to realize that a formula and a scope structure B involving all terms
linearly completely determines the truth-functional interpretation. So, once a parse through a
natural language string has set up such a pair, a complete and well-formed representation has
been constructed. Of less importance now is to understand the algorithm which transforms
such a pair into a more familiar logical formula, for instance, one of the typed lambda calculus,
though see section 7.3.2.2.
Quantication 235
NP () df F o() ^ T y(e)
DET () df h#1 iF o()
V AR() df h#0 ih#0 iF o()
REST R() df h#0 ih#1 iF o()
NOM () df h#0 iF o()
Notice that two nodes in the tree projected by a quantied noun phrase are
annotated by a Formula of type e: the top node and the one annotated by the
nominal variable. These two locations in the tree will give rise to two possible
extensions and interactions with the predicate which this term is involved in,
which is generally called the nuclear scope of the quantier. In chapter 4 we saw
that relative clauses give rise to LINK ed tree structures, and that for restrictive
relatives, the head to which the LINK ed structure is connected is the nominal
variable, whereas in non-restrictive relatives, the head is the top node of such
NP congurations.
We will also introduce a monadic predicate DOM , which collects the tem-
poral index plus all nominal variables it dominates:
T n j=M DOM () () T n j=M ?T y(t) ^ (## (T y(e) ^ V AR()) _ T e()):
If DOM is evaluated at the root node of a tree, then all type e variables are
collected by the predicate plus the tense index.
For dierential treatment of indenites with respect to scope, we will also
use DOM +, which refers to the proposition Indef(+) contributed by indenite
determiners:
T n j=M DOM + () () T n j=M ?T y(t) ^ (## (:Indef(+) ^ V AR()) _ T e())
We return to this DOM + in due course, having introduced the eect of the
determiners themselves.
7.3.1.1 Determiners
The Binder is contributed by the determiner in the surface string. Determiners
interact in various ways with common nouns. All determiners have type cn ! e,
but they create a type e formula from a type cn formula in very dierent ways.
One category of determiners is formed using variable binding term-operators.
The indenites and universal quantiers, as do most generalized ones, supply a
binding operator , and are of the form
Xcn (; X )e :
For instance, the projection P (; P ) of the indenite singular determiner a
applied to the nominal (x; Man(x))cn results in the term (; (x; Man(x))e ). So
the determiner a contributes
fF o(P (; P )); T y(cn ! e)g
to the semantic tree, through the lexical entry
IF f?T y(e)g
THEN put(Indef(+)); make(h#1i); go(h#1i); put(F o(P (; P )); T y(cn ! e));
go(h"1 i); make(h#0 i); go(h#0 i); put(?T y (cn))
ELSE ABORT
In addition, singular indenites, like all other indenites, are subject to a par-
ticular form of scope statement, to which we return shortly. Figure 7.2 gives
the general shape of the actions undertaken by this determiner.
fT n(a); ?T y(e); }g
f(a); ?T y(e); Indef (+)g
fh"1 iT n(a); T y(cn ! e); F o(P (; P ))g fh"0 iT n(a); ?T y(cn); }g
Figure 7.2: The action projected by the indenite determiner a. The appearance
of `+' in the label sequence of the top node is a value on the ìndeniteness'
feature.
The denite determiners, on the other hand, are treated like pronouns (of
a more expressive variety); they lift the variable (of type e) out of the cn and
turn it into a restricted meta-variable. For this meta-variable an antecedent to
substitute has to be found:
F o(X U (X )[U= (X )])
r l
where l and r are the left and right projection function respectively.15 For
instance, if man projects (x; Man(x)), then the man projects, by function ap-
plication, a Formula containing a meta-variable Uman(U):
APL(X U (X )[U= (X )] ; (x; Man(x))) = UMan(U) :
r l
Indexed terms are interpreted on a rst-order model K with interpretation func-

tion I as follows: for an individual constant t in DF o and P a monadic predicate,
a constant tP is interpreted by I (tP ) = I (t) if K; I j= P (t), otherwise I (tP ) is
undened. So, for instance, I (JohnMan ) = I (John) if K; I j= Man(John) and
I (JohnMan ) is undened otherwise.
15 Given a pair ha; bi we have l (ha; bi) = a and r (ha; bi) = b.

Quantication 237
The determiner the has the lexical entry:
the
IF f?T y(e)g
THEN put(?9xF o(x); Indef( )); make(h#1 i); go(h#1 i);
freshput(U; fF o(X U (U) ); T y (cn ! e)g);
go(h"1 i); make(h#0 i); go(h#0 i); put(?T y (cn))
r
ELSE ABORT
(where the action freshput() is explained below, it is as the put() action but
introduces a fresh variable). The eect of this denition is that denite noun
phrases such as the man are analysed as anaphoric expressions, with the meta-
variable U of UMan replaced by a pragmatic substitution process. So in the
text John left: the man was clearly tired, we construct for the second clause a
tree whose root node is annotated with F o(T ired(UMan )). This variable U is
replaced by John, where the truth of Man16(John) entails that T ired(JohnMan )
is true if and only if T ired(John) is true.
On this analysis, the problematic uniqueness presupposition attributed to
denite noun phrases (see chapter 1) is not incorporated as a denotational prop-
erty of the term to be constructed, but is a pragmatic process of substituting
some value for the meta-variable, as in the case of pronouns.17
7.3.1.2 Common Nouns
The common noun man contributes the nominal variable and the restrictor as
sub-terms to the nominal :
F o(U; Man(U)) T y(cn)
The lexical entry has as condition a requirement for nominal type cn, and its
action is to introduce an argument daughter with nominal variable of type e:
F o(U) T y(e)
and a function daughter with the restrictor:
F o(X (X; Man(X ))) T y(e ! cn)
Figure 7.3 displays the actions projected by a common noun. The variable of
type e projected by the common noun is passed to the nominal node, its mother.
16 This property is shared with that, this, these, and those.
17 There are two further uses of denite NPs: terms that refer to institutionalized terms
such as (i) and fact that constructions as in (ii), but both of these require separate denition:
(i) You need to see the doctor.
(ii) The fact that I'm tired is irrelevant.
The rst type is idiomatic. The second we suggest, in the light of (iii) and (iv), is a form of
expletive:
(iii) The fact is that I'm tired.
(iv) The received opinion was that I should retire.
fT n(a); ?T y(e)g
fh"0 iT n(a); ?T y(cn); }g
fT n(a); ?T y(e); ?9ySC (y)g
fh"0 iT n(a); ?T y(cn); }g
fh"0 ih"0 iT n(a); T y(e); F o(U)g fh"1 ih"0 iT n(a); T y(e ! cn);
F o(X (X; Man(X )))g
Figure 7.3: The action projected by the common noun man
What happens with the variable present at the nominal node is then determined
by the determiner annotating its sister. As we have seen, the determiner can
either construct an object at the type e node in which the variable is bound, or
it can pass on the variable to the type e node. This variable can enter into the
nuclear scope by anaphoric means, or it can be copied to a LINK ed structure
to prepare for a restrictive relative. Given such a linkage to the variable node,
the Formula value at the common noun node can be updated to create an
extended restrictor.
The variable introduced has to be fresh. We will use a rather primitive
device to achieve this: we will have an action freshput(x; ) which checks a
nite sequence of variables one by one and uses the rst xi not occurring in
the tree to execute put([xi =x]). Obviously, the number of variables has to be
chosen so as to be large enough to accommodate the trees under consideration:
freshput(x; F o()) =
IF f" # F o(x1 )g
THEN IF f"# F o(x2 )g
THEN : : :
: : : IF f"# F o(xn )g
THEN ABORT
ELSE put(F o([xn =x]))
:::
ELSE put(F o([x2 =x]))
ELSE put(F o([x1 =x]))
Notice the variable x is here a name not ranging over F o values: the boldface `x'
does that. This action is now used in the macro projected by common nouns.
Quantication 239
Common noun M
IF f?T y(cn)g
THEN make(h#0i); go(h#0i); freshput(x; F o(x)); put(T y(e));
go(h"0 i); make(h#1 i); go(h#1 i);
put(F o(X (X; M (X ))); T y (e ! cn)); go(h"1 i);
go(h"1 i); hfV AR(x)g; put(?SC (x)); ABORTi; go(h#0 i)
ELSE ABORT
So a common noun creates an argument daughter where it deposits a fresh vari-
able of type e; it creates a function daughter and annotates it with the restrictor;
and nally, it annotates the type e node, immediately dominating it with a re-
quirement for a Scope Statement hfV AR(x)g; put(?SC (x)); ABORTi involving
the nominal variable.18 The next paragraph will explain this requirement.
7.3.2 Scope
As has been discussed, we will render scope relations between projections of noun
phrases as scope statements of the form x < y expressing that the term involving
variable x has scope over the term involving variable y. These statements will
annotate top nodes in the tree { that is, nodes annotated by T y(t) propositions.
We will dene a monadic `SCope' predicate SC , where SC () indicates that
the Formula value annotating the pointed node is involved in scope relations.
Denition 13 The monadic predicate SC is dened by
SC () df NP () ^ " (T y(t) ^ 9y(( < y) _ (y < ))):
As expressed by this denition, SC () holds at a node if the Fo value is of
type e and this value is involved in a scope relation annotating some dominating
node of type t.
Above, we have seen that a common noun projects a requirement ?SC (x) for
the nominal variable x at the top node of the tree projected by the whole noun
phrase. So, in order to end up with tree without requirements, the requirement
for a scope statement has to be fullled by an action annotating some appropriate
node with such scope statement. These kind of actions will be pragmatic ones,
i.e. actions assigning scope information will be elements of the set P .
Because of the left-to-right nature of the parse, the actual implementation
of the relative scope relations of the various NPs introduced in the sentence has
to be deferred until the sentence has been parsed through. Determining the
scope structure of the eventual representation takes place in two stages. First,
the string is parsed, and the representation is annotated with scope statements.
In the second stage, these statements are used to construct a more familiar
quantication structure.
18 The notation `h; ; i' indicates ÌF , THEN , ELSE '.
7.3.2.1 Introduction of Scope Statements

At its introduction, having been culled from the natural language string, a quan-
tied expression will receive an initial description in which the restrictor derives
immediately from the common noun of the quantied NP. Thus, parsing through
Every dog ate a biscuit, our example (7.22), the phrase a biscuit will initially
be represented by (; y; Biscuit(y)) and the phrase every dog by (; x; Dog(x)).
These terms are retained until the tree has been evaluated to the top node.
Associated with each such term, at its type e node, there is a scope re-
quirement ?SC (x), contributed by the nominal, and an annotation Indef( =+)
contributed by the determiner. The scope requirement is fullled once the nom-
inal variable (bound to x) is involved in some scope statement annotating some
appropriate dominating node of type t. This statement indicates the scope of
the element relative to other appropriate expressions in the resulting formula; it
is, in eect, a quantier storage device. The label Indef(+= ) indicates whether
we are dealing with an indenite NP or not. If the NP under consideration is not
an indenite, then the scope it introduces follows the order of occurrence in the
string: if the previous quantied NP had variable x and y is the variable bound
by the current NP, then the statement x < y is added to the rst dominating
node requiring type t.
In the case of indenites, the scope is xed as an anaphoric-like choice. The
scoping statement associated with an indenite quantier Qy is completely free,
apart from the fact that it must depend on something. In the actual parsing
process, this is eected by the indenite projecting a scope statement involving a
meta-variable to depend on. Subsequent pragmatic substitution of this variable
then leads to a complete scoping.
Note that in the lexical entry below, the assignment of an initial scope state-
ment takes place after the determiner and common noun have been processed
and left their marks at the mother node, but before the types of the various
daughters have been compiled to give T y(e):19
19 The determiner xes the kind of scope statement through contributing Indef( =+) to
the type e node, and the common noun contributes the variable. So the type e node is
the rst where all relevant information for scoping is present. This is why scope cannot
be determined by the lexicon. On the other hand, if we allow the determiner to dom-
inate its variable in the way the nominal does, i.e. have a tree description of the form:
fT n(a); F o(P (; P [x=l(P )])); T y(cn ! e)g,
fh"0 iT n(a); F o(x); T y(e)g,
fh"1 iT n(a); F o(yP (; P [y=l (P )])); T y(e ! (cn ! e))g;
then all the information for scoping is xed at the level of the determiner; so it can be pro-
jected by the lexicon, and we do not need the intermediate step via the type e node (with extra
requirements). Such a variable would need a tree node of its own on this alternative analysis,
for this is the only way in which we can force its freshness { note the discrete variable for
both determiner and nominal.
Quantication 241
Scope action for indenites:
IF fIndef(+); ?SC (x); ?T y(e)g
THEN gofirst"(?T y(t));
freshput( (U; (U < x, )
?9y(DOM (y) ^ y < x ^ 8z(y < z ! x < z))))
ELSE ABORT
Notice that, though the selection of some meta-variable U for an indenite to
depend on is free, the choice domain consists of the type e Formula values
satisfying DOM (), and so has to be selected from the terms introduced as part
of the interpretation of the string, and not, like anaphora, more freely.
Scope action for non-indenites:
IF fIndef( ); ?SC (x); ?T y(e)g
THEN IF fDOM + (y); 9z(DOM + (z) ^ y < z)g
THEN ABORT
ELSE put(y < x)
ELSE ABORT
Notice that with respect to scope, non-indenites are blind to the variables
contributed by the indenites (this is eected by using the DOM + predicate).
7.3.2.2 Evaluation of the Scope Statements
After the tree has been evaluated to the top node, this node is annotated with
F o()^T y(t) plus scope statements. These statements can be used to iteratively
(and algorithmically) restructure the formula to give it a form that is more
familiar to us. The restructuring is guided by the linear, connected and maximal
S < x1 < : : : < xn chains in the set of scope statements. As we have assumed,
for every formula, there is only one such sequence. The end result will be a tree
with a top node annotated by F o() such that the semantical interpretation
of re ects the desired scope relations between the quantied NPs. The basic
restructuring move is of the form:
Denition 14 (Q-Evaluation Rule) Given a formula in which the terms
(1x1 1)A; : : : ;
(nxn n) occur, the Q-evaluation rule has the form
h: : : ; hS; x1 : : : ; xn i; : : : ; ti : [xn n =xn ]
h: : : ; hS; x1 : : : ; xn 1 i; : : : ; ti : f x ();
n n n
where for x occurring free in and S a (temporal) index, the values fx (),
for 2 f; ; Qg, and fS () are dened by:
fx () = 8x( ! )
fx () = 9x( ^ )
fQx () = Qx( )()
fS () = (S : )
So the Q-evaluation rule consumes one variable, the last, from the scope se-
quence in order to restructure the formula. Iteration of this procedure will lead
to S being the only variable left. This variable is then xed to the front of
the formula. We stipulate, in fact, the following incremental relation between
atomic formulae (see remark 6 of chapter 9)
[x =x] F o fx ():
These incremental updates lie at the basis of the following Quantier Rules:
Q1 ffffTTn(na()a;)T; Ty(yt()t;)S; S<<xx;1:;:::: :; x; xn 1<<xxn ; ;FFoo((8[x n =xn ]); : : : ; }g : : : g
xn ( ! )); : : : ; }g : : : g
1 n 2 n 1
and
Q2 ffffTTnn(a(a);)T; Tyy(t()t;)S; S<<xx1; :; :: :: :; x; xn 1<<xxn ; F; Fo(o(9[x n =xn ]); : : : ; }g : : : g
1 n 2 n 1 x n ( ^ )); : : : ; }g : : : g
and
Q3 ff ffT n(a); T y(t); S < x1 ; : : : ; xn 1 < xn ; F o([Qxn =xn ]); : : : ; }g : : : g
T n(a); T y(t); S < x1 ; : : : ; xn 2 < xn 1 ; F o(Qxn ( )()); : : : ; }g : : : g
and
ffT n(a); T y(t); hS i; F o(); : : : ; }g : : : g
Q4 ffT n(a); T y(t); h i; F o(S : ); : : : ; }g : : : g
These rules tell us now explicitly how to restructure F o values in order to satisfy
scoping requirements. Again, the interpretation of (7.22) as:
hS; y; xi F o(Ate((; x; Dog(x)); (; y; Biscuit(y))))
can be restructured to
hS; yi F o(8x(Dog(x) ! Ate(x; (; y; Biscuit(y)))))
by Q1, and this again can give rise to
hS i F o(9y(Biscuit(y) ^ 8x(Dog(x) ! Ate(x; y))))
by Q2. This, nally, leads by Q4 to
h i F o(S : 9y(Biscuit(y) ^ 8x(Dog(x) ! Ate(x; y))))
Notice that, by denition 12, the statement y < x remains true under the
restructuring, because removal of y <B x from the scope relation is compensated
by the fact that now the quantier binding y has scope over the quantier
binding x.
This unravelling process applies to yield bound variable interpretations of
pronouns, as in the sentence
(7.23) Every dogi indicated hei was hungry.
Quantication 243
Encountering the pronominal he in a parse through this sentence, an antecedent
has to be chosen and substituted for the variable U contributed by he. In this
case, the antecedent has been the term (; x; Dog(x)) as it has been constructed.
Having parsed the last word of the sentence, the decoration that results at the
top node (ignoring tense) is
fT n(0); T y(t); F o(Indicate((; x; Dog(x)); Hungry((; x; Dog(x)))); : : : g
The rule Q1 then works as it should in giving
8x(Dog(x) ! indicate(x; hungry(x)))
Compare this to (7.24) under the interpretation indicated:
(7.24) Every dogi who indicated hei was hungry, : : :
In (7.24), the pronominal is construed to give rise to a restrictive relative. The
substitution of the pronoun's meta-variable by the nominal variable x will ensure
that the variable x, introduced by the common noun dog, is linked to the top
node of the LINK ed structure annotated by
fT n(a); T y(t); F o(Indicate(x; Hungry(x))); : : : g:
By the incorporation process of chapter 4, this creates a complex nominal:
fT n(a); T y(cn); F o(x; Dog(x) ^ Indicate(x; Hungry(x))); : : : g:
Restructuring by rule Q1 then results in the desired
8x((Dog(x) ^ Indicate(x; Hungry(x))) ! : : : ):
These various rules will then interact to capture the fact that if a pronoun
in a relative within a quantifying expression is construed as identied with
some other term external to that expression as in (7.25), the only possible
interpretation is one in which the indenite is construed as taking narrow scope
with respect to that other term, with the unravelling process constrained by the
variable structure inside the Fo value:
(7.25) Every dog choked on a biscuit he ate.
Example (7.26) requires a dierent treatment:
(7.26) Every dogi ate a biscuit. Hei indicated hei was hungry.
Here the term (; x; Dog(x)) that is introduced on encountering the noun phrase
every dog, needs to be preserved (in a sense to be made clear in section 7.3.3),
in order to serve as an antecedent for the pronominal in the second sentence.
But this initial sketch at least indicates how such a term-based analysis can
yield interpretations of pronouns in both the restrictor and the nuclear scope of
a quantifying expression, as in (7.23){(7.25).
7.3.2.3 Creation of Scope Relations

Since scope relations are projected by determiners and common nouns, they
allow idiosyncratic lexical variation, either within an individual language or be-
tween apparently equivalent terms across languages. Variation in scope can be
dened for any item which amongst its actions decorates nodes of type e: that
is, either determiners or verbs. Determiner variation, as we've already seen,
projects variation on scope choice, and this may vary across languages. So, for
example, Chinese precludes any interpretation of an indenite on some following
term. This restriction, which is more stringent than English, is straightforwardly
captured by dropping the meta-variable in the characterization of the scope ac-
tion (which allows waiting on an instantiation), in favour of a standard variable
which must nd an instance on the spot.
IF fIndef(+); ?SC (x); ?T y(e)g
THEN gofirst" (?T y(t));
hfDOM (y)g; put(y < x ^ 8z(y < z ! x < z)); ABORTi
ELSE ABORT
An example of variation between quantifying elements within a single language
is the eect of adding the adjective certain to the indenite article in English.
The sequence a certain imposes the requirement that the phrase it introduces
may not be dependent on some term within the same clause as the indenite.
So in (7.27), a certain historical problem cannot be construed as ranging over
as many historical problems as students:
(7.27) Two professors insisted that every student in their class studied a
certain historical problem.
It can, however, be construed distributively with respect to two professors.
The second source of scope variation may be more surprising: verbs may
project requirements on the formulae annotating the nodes they select for. For
example, the so-called middle verbs are much more restricted in their interpre-
tation than regular transitives; and weigh in (7.28) can only be interpreted with
the term in the predicate as taking narrow scope with respect to the subject,
and not vice versa:
(7.28) A cake I made weighs two kilos.
Another example is English double object constructions. These are not identical
to indirect object constructions for they dier in their potential for scope vari-
ation from the prepositional phrase variant. The rst term of a double object
construction, (the indirect object) cannot depend on the subsequent expression
(the direct object). Whereas (7.29) is ambiguous, (7.30) is unambiguous, with
the rst noun phrase only able to be construed independently of the second:
(7.29) John showed a book to every student.
(7.30) John showed a student every book.
Quantication 245
Leaving on one side the full lexical projection of verbs such as show, the appro-
priate requirement on the local node is nevertheless straightforward to state as:
?8xy(DIR-OBJ(x) ! (IND-OBJ(y) ! :x < y))
where DIR-OBJ() h#1ih#0iF o() and IND-OBJ() h#1 ih#1ih#0iF o().
Here we assign both dative and double object constructions a tree structure as
displayed in gure 7.4
f?T y(t)g
SUBJ f?T y(e ! t)g
DIR-OBJ f?T y(e ! (e ! t))g
IND-OBJ f?T y(e ! (e ! (e ! t)))g
Figure 7.4: Double object construction tree

Notice how capturing this variation between two semantically equivalent
sentences is straightforward once the scope statements are formulated indepen-
dently of the tree itself, a notable advantage of this style of analysis.
7.3.3 Term Reconstructions
In our model, noun phrases are introduced in the representation by the intro-
ductory sketch (; x; ) 2 DF o of type e consisting of a binder (projected by
determiners), a nominal variable x and a restrictor (the latter two projected
by common nouns). These objects are not given a semantics, because they dis-
appear under the evaluation of the semantic tree, to be replaced by the more
familiar quantiers. However, the essential part is not that the terms disappear,
but that the scope information is incorporated into the formula itself. And the
restructuring that takes place during the absorption of scope information does
not have to eliminate the term. On the contrary, the scope information may
be embodied in the terms themselves. The terms that arise in the process have
been studied in logic in the so-called epsilon calculus, for instance, and there
they do have a transparent semantics. In such a calculus there are no quanti-
ers, but only terms with an internal structure20complex enough to contain all
information required for a semantic evaluation.
We will illustrate the intuition behind the semantics by returning to (7.22)
(repeated here):
20 The epsilon calculus was dened by Hilbert (1939) as a formal system within which the
arbitrary names familiar in predicate logic natural deduction could be subject to formal study.
See also Prawitz 1965, Leisenring 1969, Meyer-Viol 1995.
(7.22) Every dog ate a biscuit.

In the rst-order translation,
8x(Dog(x) ! 9y(Biscuit(y) ^ Ate(x; y)));
interpreting one reading of (7.22), the locations occupied by the variable y can
be fully described by the triple:
(; y; (Biscuit(y) ^ Ate(x; y)))
This description species the mode of quantication (existential), the variable
being bound (y) and the binding context, (Biscuit(x) ^ Ate(x; y)). Now if we
replace the variable y by its description (abbreviated here to ax, the subscript
`x' indicates the free occurrence of x in that term) and remove the corresponding
quantier, we get the formula:
8x(Dog(x) ! (Biscuit(ax) ^ Ate(x; ax )))
From this formula we can always retrieve the original quanticational structure;
that is, no information is lost when we eliminate the existential quantier
in favour of an epsilon term with a complex internal structure. In the same way
we can now remove the quantier 8x by replacing x by the description (a tau
term):
(; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax)))
where indicates the mode of universal quantication. Abbreviating this term
by b, this gives the quantier-free formula:
Dog(b) ! (Biscuit(ab ) ^ Ate(b; ab ))
where ab stands for:
(; y; (Biscuit(y) ^ Ate(b; y)))
This quantier-free formula can now be evaluated to produce the same truth-
functional behaviour as the quantied formula that was the starting point. That
is, for every Model K; I we have:
K; I j= Dog(b) ! (Biscuit(ab) ^ Ate(b; ab )) i
K; I j= 8x(Dog(x) ! 9y(Biscuit(y) ^ Ate(x; y)))
This process of term construction has been formalized in a logical system (for
discussion see Meyer-Viol 1995), which has the following two rules concerning
the terms:
9x(x) (; x; (x))
(; x; (x)) 8x(x)
Quantication 247
Addition of these two rules to a standard system of (classical) rst-order pred-
icate logic does not add to its classical power.21 Notice that, using these rules
in combination with the standard quantier rule, we have:
f8x(Dog(x) ! 9y(Biscuit(y) ^ Ate(x; y)))g `
fDog(b) ! (Biscuit(ab ) ^ Ate(b; ab )))g:
where the abbreviations b and ab are as dened above.
In our set-up this semantics holds only after the scope relations have been
incorporated into the terms; a semantics for the formula tout court is not den-
able until after we have reached an empty scope sequence. This means that not
all constellations of formula and scope sequences lead to formulae which make
semantic sense. For instance,22
h: : : h i : : : ; ti : Dog(; x; W alk(x))
has no proper semantics, but in a dierent scope environment:
h: : : hxi : : : ; ti : Dog(; x; W alk(x))
the formula can be further restructured to:
h: : : h i : : : ; ti : Dog(a) ^ W alk(a)
where a = (; x; (Dog(x) ^ W alk(x))). Now we have a semantics, because
9x(Dog(x) ^ W alk(x)) $ (Dog(a) ^ W alk(a)):
In order to construct the appropriate terms (b and ab respectively) from the
initial structures, for example (; x; Dog(x)) and (; y; Biscuit(y)) together with
the scope statements, rather than from the predicate logic formula, we give the
following update rules which should be compared to those of denition 14.
Denition 15 (Term Restructuring) For in which the variable x occurs
free
fx () = [a=x] ! [a=x]
where a = x( ! )
fx () = [b=x] ^ [b=x]
where b = x( ^ )
fQx () = ( [c=x])([c=x])
where c = Qx(( )()).
21 It is a so-called conservative extension of rst-order predicate logic.
22 Here we give the display strictly as a node on the tree (with the Type and other labels,
labelling the Formula ), since it is necessary to be able to refer to the absence of a scope
sequence.
As we have now seen several times, the parse of the string Every dog ate a biscuit
results in a tree the top node of which is annotated by
(S < x < y) F o(Ate((; x; Dog(x)); (; y; Biscuit(y)))); T y(t):
The term-restructuring rules now allow us to generate by the second rule the
structure:
(S < x) F o(Biscuit(a) ^ Ate((; x; Dog(x)); a)); T y(t)
where
a = (; y; (Biscuit(y) ^ Ate((; x; Dog(x)); y)))
Notice that (; x; Dog(x)) is a sub-term of a. This can be seen as dening
(; x; Dog(x)) < a. Then applying the rst rule gives:
(S ) F o(Dog(b) ! (Biscuit(ab) ^ Ate(b; ab ))); T y(t)

where
b = (; x; Dog(x) ! (Biscuit(c) ^ Ate(x; c)))
c = (; y; (Biscuit(y) ^ Ate(x; y))).
The scope statement, x < y has disappeared, because the term b with variable
x occurs as a sub-term of a with variable y.
Notice that in computing the tau and epsilon terms from the initial struc-
tures, the restrictors of the terms and of the predicate have to be successively
restructured in the course of the evaluation. Every dog is restructured as:
(; x; Dog(x)) F o (; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax))))
where ax is (; y; (Biscuit(y) ^ Ate(x; y))); some biscuit is restructured as:
(; y; Biscuit(y)) F o (; y; (Biscuit(y) ^ Ate(b; y)))
where b is (; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax)))); and the predicate
Ate(x; y) is restructured as:
Ate(x; y) F o Dog(x) ! (Biscuit(y) ^ Ate(x; y)):
As is clear from the above procedure, the global nature of the quantiers is
re ected in the internal complexity of the terms that are introduced. That is,
an atomic formula like Dog(b) has complex internal structure hidden in the
term.
Quantication 249
7.3.4 Applications { E-type Anaphora
The advantage of using such epsilon terms in linguistic analysis is that they
denote directly the concept of witness set, often invoked in the analysis of E-
type pronoun construal (see van der Does 1996 among others, and Meyer-Viol
1995 and van Heusinger 1997 for analyses of E-type pronouns as epsilon terms).
The idea is that, for instance, (7.31)
(7.31) A man is leaving.
is represented by:
Leave(; x; Man(x));
where (; x; Man(x)) is the denotation of a man. Now, by the appropriate sub-
clause of denition 15, the predicate Leave(x) is incorporated in the restrictor
of the term to give:
(; x; (Man(x) ^ Leave(x))):
This term, supposedly, is the correct antecedent for he, in (7.32):
(7.32) A man is leaving. He is smoking.
The second sentence of (7.32) is represented by:
Smoke(; x; (Man(x) ^ Leave(x)))
and thus the restrictor can again be extended to create a term representing a
leaving, smoking man. In this way, terms can be built up to be used as complex
antecedents. Notice how the construction of the term is possible only after
the completion of the interpretation of string.
Similarly, in (7.33):
(7.33) Every dog ran out. He barked.
The rst sentence of (7.33) is represented by
Run(; x; Dog(x));
and from this form the predicate can be incorporated in the restrictor of the
term, but now to give (; x; (Dog(x) ! Run(x))). Continuation by the second
sentence in (7.33) is represented by
Bark(; x; (Dog(x) ! Run(x)))
where the new predicate should now extend the consequent of the implication
to give the term
(; x; (Dog(x) ! (Run(x) ^ Bark(x)))):
Instead, the rule as we have formulated it gives the incorrect
(; x; ((Dog(x) ! Run(x)) ! Bark(x))):
In order to get the right result, we can consider a more subtle restructuring rule:

fx () = ( ! ( ^ ))[a=x] if = !
( ! )[b=x] otherwise
where a = x( ! ( ^ )) and b = x( ! ).
In the second sentence of (7.26) (repeated here):
(7.26) Every dogi ate a biscuit. Hei indicated hei was hungry.
the variable associated with the pronominal hei is substituted by the `condi-
tional object' (; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax)))) which has been pro-
duced by the previous sentence. Now, the above rule allows a restructuring to
incorporate the new information in the (projection of the) antecedent of the
pronominal:
fT y(t); ( b ); F o(Indicate(b; Hungry(b))) : : : g
fT y(t); ( ); F o(Dog(c) ! (Biscuit(ac) ^ Ate(c; ac) ^ Indicate(c; Hungry(c))))g
where
b = (; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax ))))
c = (; x; (Dog(x) ! (Biscuit(ax) ^ Ate(x; ax )) ^ Indicate(x; Hungry(x))))
And the Formula value the rule produces has the correct truth conditions as
a representation of (7.26) as indicated. In the latter characterizations, we are
recreating the eect of Accommodation rules posited in Discourse Representa-
tion Theory (see Roberts 1989, Kadmon 1990), but, unlike in that framework,
this same eect is here achieved by general principles of term construction and
not by special stipulation.
With this characterization of such terms as re ecting the full content of the
root formula, we have directly to hand the explanation of E-type eects of the
construal of head-internal relatives in Japanese:
(7.34) Hiroto wa Mary ga sanko no ringo o muita no o
HirotoT OP MaryNOM threeGEN appleACC peeledNom;ACC
tabeta.
ate
`Hiroto ate all the three apples which Mary peeled.' [Japanese]
In this type of example, the completed term projected from sanko no ringo
o assigned as an annotation to a node within the LINK ed structure is also
assigned (by the LINK Introduction rule) as the corresponding term projected as
an argument of the predicate in the primary structure (see chapter 4). Because
the completed term re ects the full content of the established propositional
formula, albeit itself of type e, the interpretation of the relative indicated is the
only possible interpretation of the relative.
The consequence of these dierent but equivalent modes of display of content
of quantied formulae is that the algorithm translating scope requirements into
scope structure can take two forms.
Quantication 251
First, we can transform the pair [(1 x1 1)=x1 : : : (n xn n )=xn] and B into a
formula 0 with standard quantiers and without structured terms. For a pure
quanticational formula, the scope statement has the simple interpretation
T n j=M xi < xj () 9 = : : : Qx1 : : : Q0 xj 2 DF o ; T n j=M F o(),
where Qxi and Q0xj are unique with respect to the variables they bind,
and xi ; xj occur free in .
That is, a scope statement xi < xj expresses that the quantier binding xi
occurs to the left of the quantier binding xj in the representational formula.
On this analysis, the initial projection of terms of type e proves to have been
for convenience only.
Secondly, we can transform the pair [(1 x1 1)=x1 : : : (n xn n )=xn]) and B into
an updated, quantier-free formula 0 [(1x1 10 )=x1 : : : (n xn n0 )=xn]:
[(1 x1 1 )=x1 : : : (n xn n )=xn )] F o 0 [(1 x1 10 )=x1 : : : (n xn n0 )=xn ];
where both the predicate and the arguments are updated in the fullment of the
scope requirements. Now the scope requirement xi < xj has the interpretation
T n j=M xi < xj () 9 = [(j ; xj ; j0 )=xj ; (i ; xi ; i0 )=xi ] 2 DF o;
T n j=M F o(),
where (j ; xj ; j0 ) = a[(i ; xi ; i0 )=y] for some a which has y free.
That is, xi < xj holds at a node with Formula if the unique term with variable
xi occurring in is a sub-term of the unique term with variable xj occurring
in .
In this latter case, the NP projections remain representations of type e,
although they undergo several successive updates between their introduction in
the tree and their shape in the eventual, nished tree.
We leave these options, merely noting their relative merits. The rst alter-
native retains the eventual formula in a familiar predicate logic format, making
use of the variable binding term-operator notation only as a notation of con-
venience for purposes of computing the formula, albeit not its nal form. The
second alternative retains a quantier-free formula throughout, so that the in-
ternal structure of the tree re ects not merely the process of computing the
resulting formula but also the internal structure of that formula. Its disadvan-
tage is the very considerable internal complexity of the terms, but, given the
need for such terms in characterizing E-type anaphoric eects and the general
representationalist stance, this hiding of the semantic structure within the term
may be appropriate. Indeed, the striking advantage of constructing such terms is
that they are then automatically available as antecedents for some subsequent
pronoun construal, and the regular substitution action can apply: nothing
particular needs to be said.
Overall in this chapter, we have set out to provide an account of quantica-
tion satisfying the syntactic desideratum that noun phrases project arguments
to the predicate. The motivation for this was the need to provide an analysis
of indenites as lexically projecting an underspecied scope statement. The
decision has a number of advantages. It makes it possible to provide a basis
from which E-type anaphora can be explained as part of a uniform pragmatic
account. Moreover, the compilation of epsilon terms with complex internal
structure means that these anaphoric eects can be expressed without any ad-
ditional statement. It also provides a natural basis for characterizing lexical id-
iosyncrasies constraining scope which are quite unstatable in terms of a general
storage mechanism. More generally, it provides a relatively natural re ection
of the way in which choices, made on a left{right basis, aect scope of terms
within some logical formula. And the phenomenon, as so described, provides a
further instance where pragmatic actions feed into an algorithmic process { here
that of establishing the logical representation of the content of a string. This
account thus gives one more example of a structural process of interpretation
that follows the left{right ow of the parsing process and involves interaction
between computational and pragmatic actions.
8
Re ections on Language
Design
8.1 The Overall Perspective
Throughout the previous chapters, we have been outlining and arguing for a
formal model of language which purports to be at one and the same time a
model of an incremental left{right process of interpretation for natural language
strings, yet also a basis for explaining structural properties of language, and
hence a grammar formalism. We now look back on what we have done in order
to be able to re ect on the relevance of the results for general linguistic theory.
Our starting point in this endeavour, in chapter 1, was to set out the prob-
lems for semantic accounts of anaphora by way of background motivation for ex-
ploring in some detail a representational account of the process of interpretation.
In case after case of pronoun{antecedent pairing, we found that characteriza-
tion of pronominal construal through a semantic characterization of the relation
between the expression and the object denoted was far from straightforward.
What appeared to be needed was some representational form of explanation,
analysing a pronoun as projecting a skeletal indication of some resulting inter-
pretation to be replaced by a term selected either from what is made available
in the build-up of interpretation of the sentence under evaluation or, more gen-
erally, from implications drawn from the previous discourse.
In chapters 2{3 we dened a modular system which, across a left{right
sequence of words, induces a structure from which some lambda term is compiled
representing an interpretation of the string. The progressive update of partial
tree structures re ects three properties central to natural language processing:
(i) The parsing process is goal-directed, the goal being to establish some
logical form as representation of a propositional content for some sentence.
(ii) The information needed to establish such interpretation is progressively
accumulated across the left{right sequence until the goal is achieved.
253
254 Re ections on Language Design
(iii) The specication of interpretation provided by a sequence of words may

be less than the interpretation assigned to that sequence, and in such
cases interaction with pragmatic actions is needed to update the lexical
specications provided.
The system is explicitly dynamic, dening a sequence of transitions across par-
tial tree structures, each node having imposed on it requirements, some of these
requirements leading to the development of other nodes, each with one or more
requirements imposed. The two primary concepts of tree growth were:
growth of tree structure by addition of nodes and xing of tree relations,
growth of decorations on a node with progressive replacement of require-
ments by annotations.
These were implemented through three types of action, each with the property
of mapping some partial structure to another: computational, lexical and prag-
matic. Of the computational actions, it was the introduction of an underspeci-
ed tree relation which gave us the novel account of long-distance dependency
that forms the central analysis of this book, in which an unxed node is intro-
duced from some top node requiring type t, and eventually merged with a xed
position in the emergent tree structure. Relative to this dynamic perspective,
long-distance dependency is not an isolated property of underspecication in an
otherwise fully determined system (as in LFG { Kaplan and Zaenen 1989); nor
is it an imperfection of what is otherwise a `perfect' computational system (as in
the Minimalist programme { Chomsky 1995). To the contrary, it exemplies a
core property of a system specically adapted for parsing { that of a projection
of some partial specication which requires later updating.
In the lexicon too, the essential dynamics creates a new perspective on lexi-
cal content. Words are no longer assigned to some terminal node in a structure.
Rather, a word projects an idiosyncratic sequence of actions updating a partial
tree description as a contribution towards the overall goal of enabling a logical
formula of type t to be derived. It is, that is, dened to be a contributor to
the process of establishing a representation of one possible content of the sen-
tence. This allows us a very much increased richness in the type of information
projected by a lexical item. It may project decorations on more than one node
or construct additional nodes. This procedural concept of lexical content can
distinguish between words projecting annotations of the same type in terms of
other actions that they may initiate. Verbs, for example, can be dened in dif-
ferent ways for dierent languages; and despite sharing the same logical type,
they may dier with respect to additional actions, providing a lexical basis for
characterizing clause-internal word-order variation. We can also distinguish be-
tween the contribution made by a word in two discrete environments despite
projecting a single Formula value, leading to a cluster of compound actions for
a single word with slightly dierent eect. As we saw in chapter 3 and sub-
sequently, this is a common pattern. Wh expressions in English provide one
of many examples. They are assigned a disjunctive characterization, the two
The Overall Perspective 255
alternatives diering in that there is a projection of an additional feature Q
only if the wh expression annotates an initially unxed node. An example very
much later was the denition in chapter 7 of dierent scope constraints for verbs
such as give, depending on whether they were projecting an indirect object or a
double object construction. Such disjunctive characterizations do not have the
status of ambiguity as in other frameworks, projecting as they do but a single
Formula with additional decorations: they are merely a re ection of ancillary
context-relative properties of a lexical item. Finally, we can straightforwardly
dene co-occurrence between particular pairs of words, with one of a correlated
pair of words projecting a requirement for the formula projected by the item
with which it is correlated, that second item then being dened to provide the
appropriate annotation. This gives us a means of expressing local discontinuities
of a number of dierent sorts: agreement phenomena, expletives of one sort or
another, verb{particle constructions, and so on. The exibility for lexical char-
acterizations provided by the procedural perspective and the concept of partial
descriptions is very considerable.
Pragmatic actions, the third type, also play a part in determining well-
formedness, even though the implementation of any one such action is relative
to system-external constraints. The pragmatic actions dened were substitution
operations updating (in eect replacing) a place-holding formula with some se-
lected term (the identity of which is not determined by the system). This gave
us our representationalist account of anaphora construal, with the anaphoric
expression dened to project an incomplete specication for a Formula expres-
sion rather than a xed term. Notice how, in this respect, anaphora is like
long-distance dependency in providing an instance for which the parser takes
some projected characterization and updates it: as in the case of structural un-
derspecication, it is not some surprising irregularity, but a denitive property
of a system developed for establishing some increasingly rich description of an
interpretation.1
The system, as developed in chapters 2 and 3, was applied to English, where
the correlation between top{down projection of structure and left{right parsing
is straightforward to dene, and also to Japanese, where the correlation is much
less obvious. In the latter case, the ability to articulate underspecied tree
relations at non-nal stages of the parsing process was essential, enabling us
to capture the tension between the apparently at sequence of constituents in
such languages and the resulting congurations that they project (see the long-
standing debate as to whether such languages are congurational or not { Hale
1983, Kiss 1991, Diesing 1992, Speas 1990, Maracz and Muysken 1988 etc.).
In chapter 4 we started to see the benets of such a dynamically dened
system in application to a particular class of data { relative clauses. We dened
one more type of tree relation, the LINK relation between one tree structure
and another, and a concept of handing on information from one structure to
1 Underspecication might also be allowed in other labels. See Marten 1999 for a proposal
that lexical items, specically verbs, have an underspecied type specication which is only
established in the presence of a full set of adjunct noun phrases (or, as in English, prepositional
phrases).
another. With this extension, we then dened the process of constructing se-
quences of tree structures. Relative clauses were taken to project a LINK ed
tree independent of the primary structure except in so far as there was an im-
posed requirement on the development of the LINK ed tree from its top node
that the formula to result at that node must have as a sub-term a copy of the
formula annotating the node from which this otherwise independent tree was
developed. Given the dynamics of the system already developed for individual
tree growth, this constraint can be satised in one of a number of ways in the
emergent LINK ed structure.
This interaction of transfer of information from one tree to another and the
dynamics of tree growth for an individual tree enabled us to capture variations
both within a single language and across languages. The rst major result was
that the pattern of behaviour of pronouns in languages such as Arabic, where
resumptive pronouns are obligatory in non-subject positions in relative clauses,
was naturally achieved by dening the LINK transition for that language in
terms of an imposed requirement on the top node of the LINK ed structure for
a copy in that structure of the head formula, and not as some annotation on an
unxed node (as in English). The resulting characterization of the Arabic data,
both for relative clauses and for clitic left-dislocation structures, was achieved
without any structure-specic stipulation about the pronoun itself. It then
emerged that the concept of LINK ed structure applied equally to head-initial
and head-nal relative clauses, with minor variation in the rule formulation.
Within this general picture, the so-called head-internal relatives are also un-
problematic (unlike in other frameworks): in these relatives, the term in the
LINK ed structure is constructed in the normal way, duly providing a value for
the subsequent head formula by a simple process of copying. The result overall
was a unifying relative clause typology of considerably greater generality than
other analyses of relatives.2
In chapter 5 this style of analysis was applied to what, at rst glance, seems
a totally dierent type of phenomenon { the dierent forms of wh expressions
in questions, spanning wh-initial expressions, wh-in-situ expressions, and the
supercially very dierent phenomenon of wh -expletives displayed in some lan-
guages. In other frameworks, wh expletives constitute a scoping problem, the
expletive element `stretching' the scope for the full form of the wh expression
which follows. In these frameworks, the analysis requires additional stipula-
tion, though advocates of dierent analyses disagree as to how much additional
stipulation is needed. In the present system, no extension of the framework is
needed. It is simply that a lexical expression can be anticipatory for a second
expression by providing the annotations projected by the second expression as
a requirement. Indeed, the concept of having lexical expressions in a language
whose role is anticipatory is entirely natural within this framework, opening the
way for a general theory of expletives encompassing categories of all types (see
Perrett 2000 for an analysis of verb-chaining phenomena in Hadiyya in terms
2 The only analysis which comes close to the breadth of this characterization is that of
Kayne 1994; see chapter 4, n.42.
The Overall Perspective 257
of an expletive form of the tense marking). The analysis of wh expletives was
broadly successful at the descriptive level, covering the same data as other anal-
yses, but furthermore incorporating other data (specically the non-was forms
of expletive), an integration which is new. As in the case of the analysis of
relatives, the result is a typology incorporating the major type of wh question
words and their position, according to whether they project an annotation to
an unxed or a xed node, or a requirement on subsequent tree development.
Crossover phenomena, which were the subject of chapter 6, involved the
interaction of all the analyses so far set out { construal of pronouns, relatives
and wh questions. Its primary interest lay in the demonstration of how these
analyses, independently set up and motivated in previous chapters, could be
seen to yield an integrative account of what had previously been analysed as
a heterogeneous set of phenomena. The distinguishing property of our expla-
nation is its critical reference to the process of interpretation { the possible
construal of pronouns depended on what information is available at interme-
diate points in the interpretation process and what consequences follow from
any particular selection of a value. This was the heart of the analysis of the
crossover phenomena, the dierences between strong and weak crossover in in-
dividual languages turning on the details of what was made available by the
LINK transition, the node within the structure annotated by the wh expres-
sion or pronoun, and interaction with general transition rules. The additional
novelty of the account of weak crossover phenomena was its exploration of the
consequence of allowing interaction between computational and pragmatic ac-
tions. Dening pronoun construal as substitution of some arbitrary formula,
without any system-internal constraint on how that selection might be made,
means that the system itself is extremely liberal with respect to the data, for the
unacceptability of interpretation for a given string may be due to considerations
entirely external to the system itself. In the case of weak crossover data, we
argued that the lack of robustness of weak crossover eects in questions was a
consequence of such system-external pragmatic constraints, suggesting that the
phenomenon is not due to a structural restriction at all.
In chapter 7 we looked at quantication as a challenge to our general commit-
ment to a left{right form of incrementality of build-up of interpretation, for the
global nature of quantier scope seems to deny the possibility of sustaining an
incremental approach to natural language understanding. However, we made a
separation between scope statements and rules for determining the denotational
content of quantier expressions, and then dened a way in which constraints
of relative scope statements can be gathered in an incremental way as the tree
and its annotations are set out. This characterization makes possible the artic-
ulation of lexical constraints on scope determination which are inexpressible in
any system in which the scopal properties of quantifying expressions are dened
through a general process such as quantier movement or quantier storage. Fi-
nally, we demonstrated, albeit brie y, how this account of quantier construal
could be used to project bound variable and E-type forms of anaphora con-
strual. The particular signicance of this account of quantication is that it
provided another case where the pragmatic action of Substitution must feed an
algorithmic process, that of determining the nal logical form projected by the
string, hence requiring interaction of computational and pragmatic actions.
Once this way of looking at natural language data becomes familiar, whole
sets of phenomena suggest analyses in similar terms. We might analyse verb
raising (the verb-second phenomenon in Germanic languages) in terms of pro-
jecting some locally unxed node to be annotated by the verb. Clitic raising
also, we might analyse as the projection of an unxed node, this time for the
pronoun to annotate. Clitic doubling might be analysed as projection of a re-
quirement by the pronoun, rather than an annotation. And so on. Many of
these analyses have been only hinted at here, but, seen as dierent manifes-
tations of the progressive update of structural information, not a single one
of these phenomena should be taken as an indication of ìmperfection' in an
otherwise `perfect' system (Chomsky 1995). To the contrary, they display the
natural language system for what it is { a vehicle for updating partial descrip-
tions through from some initial goal to a logical form in satisfaction of that goal.
In short, they re ect the dynamics intrinsic to the parsing process.
Overall, it has to be conceded that what we have achieved in formal terms is
rather little; for we have barely provided enough detail to see how a full fragment
could even in principle be dened. In particular, on the formal side, we have
not provided a full characterization of control of the construction process and
pointer movement. And from the data point of view, several important types
of phenomena are missing: we have done no more than sketch an account of
tense construal, there is no characterization of connectives, hence of complex
constructions such as conditionals; and there is no characterization of non-nite
constructions or any of the so-called A-chain phenomena. As a construction of
the interpretation process too, the model provided dened is also not more than
partial.
The decision to apply the framework by individual sketches was deliber-
ate, however, for each of the problems we chose to address in some detail is a
puzzle, given current assumptions. To our knowledge, there has been to date
no attempt within any of the standard frameworks to provide a general ac-
count of pronoun construal which encompasses discourse construal,3 interaction
with long-distance dependency phenomena, and resumptive uses. The exis-
tence of head-internal relative clauses is acutely problematic for all standard
accounts of relative clauses, dened as they are in terms of the head (or its
associated relativizing operator) binding a position within the relative. W h
expletive phenomena pose another puzzle for syntactic theories, highly prob-
lematic for Minimalist assumptions. Weak crossover phenomena have remained
an acknowledged `mystery' since Postal (1993), and no attempt has been made
to relate these data to the phenomenon of strong crossover (with the exception
of Dowty (1993) working from categorial grammar assumptions). Yet, relative
to these new assumptions, the data have emerged as a consequence of artic-
ulating the dynamics of the goal-directed interpretation process, with natural
3 This, despite the fact that the changing interaction between anaphora and grammatical
processes is a well-known basis for language change. This is a well-known and traditionally
much-studied phenomenon, re-emphasized by Givon (1976).
Underspecication and the Formal Language Metaphor 259
cross-language typologies determining the limits of variation. The individual
sketches gathered together thus display the potential for new insights to be
gained by adopting the very dierent perspective proposed here { that a for-
mal characterization of the left-to-right process of building up representations
of interpretation over a sequence of words itself provides the basis for explaining
syntactic properties of such strings without need of any further level of repre-
sentation. Hence, coming round full circle to the outset of the book, we take the
model we are articulating to constitute a formal reconstruction of the common-
sense view that knowing a language involves knowing how to use it to recover
some intended interpretation.
8.2 Underspecication and the Formal
Language Metaphor
It is now time to take a step back and, on the strength of these empirical results,
to reconsider some of the general issues raised initially in chapter 1. There are
three major ways in which this framework is novel (though with respect to the
rst of these, see Hausser 1989, 1999).
(i) Structure is projected on a left{right basis across the sequence of words.
(ii) Structural concepts of underspecication are licensed, so that not all tree
descriptions constitute complete trees.
(iii) The concept of syntax is shifted to incorporate the dynamics of progres-
sively building up incomplete structures.
The rst property is self-evident: the projection of a logical form is not on a
purely bottom-up basis, but through applications of rules for unfolding a tree in
a goal-directed, top-down way as driven by the sequence of words. The second
and third properties are two sides of the same coin, the one a consequence of
the other. With a concept of underspecication incorporated into syntactic
dscription, the denition of a dynamic tree-update process becomes essential.
Unlike in other frameworks, the characterization of underspecication of in-
terpretation for a natural language expression and its resolution has not been
articulated within a system of semantic evaluation of some antecedently dened
structure. To the contrary, in our system, the diversity of construal of natu-
ral language expressions in context results from a constructive term-building
process for which natural language expressions provide input procedures. It
is not semantics, then, which is taken as the basis of the unfolding dynamics
of natural language understanding. It is syntax that has been made dynamic,
with the remit of syntax redened as covering the projection of sequences of
partial tree structures. Phenomena are described not in terms of some complete
structure or structures, but in terms of transitions across partial and growing
tree descriptions. Accordingly, the explanation of anaphora and the diversity
of interpretation of natural language expressions in context is now analysed in
structural terms as resulting from a process of substitution which is part of the

system of tree growth.
8.2.1 English is not a Formal Language
In making this change, the objects to which structure is ascribed have also
changed. Structures are articulated as the basis for interpretation; but these
structures are no longer trees whose terminal elements are natural language
expressions. The elements of the string may project a decoration (either an-
notation or requirement) for one or more nodes, an unxed node, or a node in
a LINK ed tree; or they may exhaust their contribution completely in creation
of tree structure. The consequence of this change of perspective is that the
annotations of terminal nodes in the resulting structure no longer re ect the
order of the words in the string. The change has involved a departure from
the assumption that has been the guiding methodological assumption of lin-
guistic theorizing for the last thirty years since Montague's Ènglish as a formal
language' (1974a) that natural languages can be characterized like formal lan-
guages, assigning strings of the language a syntactic and a semantic structure
that are homomorphically related.
A critical distinction between natural languages and these interpreted formal
languages lies in the underspecication of natural language expressions vis-a-vis
their interpretation in any given context, leading to perceived ambiguity for
all sentences which demand context-dependent construal. Indeed, ambiguity is
widely taken to be a distinguishing characteristic of natural languages, a prop-
erty which, by denition, is excluded in the formal languages that constitute
the pattern on which linguistic theories have been constructed. Despite this
mist, and the recognition that underspecication in the lexically dened in-
terpretation of elements is central to the analysis of natural language content,
the mould of the formal language metaphor in the articulation of grammars for
natural language has not been disturbed in any of the major formalisms. Lin-
guistic theories have maintained the concept of a formal system in which a set
of syntactic rules denes the set of possible sentences for a given language, as-
signing each such string a structural analysis (possibly involving more than one
level of representation), with some assignment of content. There is considerable
variation within individual theories as to the relation between syntactic and se-
mantic characterizations of a string and the form that semantic interpretation
might take. Nevertheless, it is assumed that context4 dependence is the burden
of semantic theories to explain, rather than syntax. The consequence is that
there is no basis for exploring possible forms of interaction between general dis-
course processes such as anaphora resolution and grammar-internal processes.
4 In accounts within the Principle and Parameters and Minimalist paradigms, in which the
exclusion of semantics from the computational system ensures that the problem of context
dependence is not addressed, the formal language metaphor is nevertheless retained, with
liberal use of indices enriching the object under analysis, so that the gap between string and
assigned interpretation is eliminated, or at least substantially reduced, and the LF assigned
to such indexed strings remains the interface level with some denotational semantics to be
dened.
Underspecication and the Formal Language Metaphor 261
By denition, syntactic phenomena are characterized by abstraction from any
application to which the natural language system might be put, and any phe-
nomena which appear to display such interaction are analysed as corresponding
to two independent subsets of data, those which can be characterized system-
internally and those whose properties are seen as a consequence of more general
discourse-based principles.
To the contrary, the present system was devised as a model of the process of
natural language interpretation, seen as the progressive building of a structured
representation. And there was a commitment, also, given the need to explain
the systematic cross-linguistic ambiguity displayed by all anaphoric expressions
(see chapter 1), to articulating a single process of anaphora resolution. The
characterization of syntactic phenomena emerged in the development of this
framework as a by-product of dening a representationalist account of the in-
terpretation process. However, the consequence of recognizing the system as a
model of natural language itself, with the triple of computational, lexical and
pragmatic actions each playing a role in determining well-formedness, is that we
abandon the assumption that natural languages are to be dened as interpreted
formal languages. According to this new perspective, it is no longer the case
that sentences are made up of words in a fully dened structure, relative to
which model-theoretic contents are dened, and independent of any application
to which such a system might be put to use. The phenomenon being character-
ized as the content of natural language expressions is not what the expressions
are used to make assertions about. Rather, it is the objects constructed in
parsing that constitute this content { the progressively developed partial logical
forms which the words in combination make available.
Notice how phenomena that are characterized in other frameworks as struc-
tural properties of a string in some static conguration, whether a word or a
phrase, are now seen in terms of an overall process of tree development and their
contribution to that process. We do not refer to the structural properties of a
string or properties of a natural language expression within a xed hierarchical
conguration, but rather to the contribution made by a word or sequence of
words to a goal-directed process of tree growth as dened over strings. Equally,
we do not articulate some model-theoretic content for a sentence as a function of
the content of each individual part of that string; rather, we dene the content
of words in terms of transitions from one partial logical form to another. It is
only the logical forms resulting from this process which have to be assigned a
model-theoretically dened content. In sum, the essence of all explanations is
the transition from one decorated partial structure to another. The dynamics
of transition between the input and output structures is the heart of the ex-
planation. On this view, the phenomenon of underspecication of expressions
such as pronouns vis-a-vis their interpretation is not some aberrant departure
within an otherwise regular formal language system. The skeletal nature of
their interpretation taken independently of context is, to the contrary, a re ex
of the fact that a natural language parser is a set of principles for progressive
and goal-directed processing of linguistic input.
The principle of compositionality which acts as a constraint on the denition
of how words in a string contribute to its overall interpretation has to be mod-

ied in the light of this new perspective. Because from our perspective, words
project actions, the compositionality principle constrains the relation between
the actions induced by strings of words and those associated with the individ-
ual elements (recall that the goal of the interpretation process is to construct a
logical form annotating the root node of a tree having used all the words in the
string { i.e. leaving the empty string, e, as the remainder):5
if (T n; w1 ; : : : wk ) )LCP (T 0 n0 ; e), then there are T1 n1 ; : : : ; Tm nm such
that
(T n; w1; :::wk ) )LCP : : : )LCP (T1 n1; w2 ; : : : wk ) : : : )LCP : : :
: : : )LCP (Tm nm ; wk ) )LCP : : : )LCP : : : (T 0 n0 ; e).
Compositionality of the formula that results can then be dened in familiar
ways, expressible as a bottom-up characterization of the content of formulae
inhabiting the nodes of the resulting tree. This characterization of composition-
ality for natural language, divided here into two quite dierent concepts, is of
course incomplete without a denition of a semantics for such resulting struc-
tures; and it remains quite open as to what form of semantics to dene for the
resulting logical formulae.6 But with this structural concept of compositional-
ity for natural language strings, we can articulate a new concept of a grammar
formalism for natural languages: a grammar for a natural language denes a
mapping from sequences of words onto sets of sequences of partial decorated
trees.
8.3 Well-formedness and Availability of
Interpretations
It might seem that in advocating a grammar in terms of possible sequences of
partial trees we have con ated the concepts of grammar and parser, and in so
doing have left ourselves no space within which to articulate any distinction at
all between grammaticality (well-formedness) and acceptability. However, this
isn't so, for we have a clearcut distinction between the grammar formalism and
a model of utterance interpretation in context. The formal mechanism dened
is not a parser in the sense of being a device which selects one interpretation
over a multitude of others; for it merely denes the space of possible parse se-
quences for strings of a language. Indeed, we cannot be said to have dened a
parser in the orthodox sense at all, since we have not addressed this so-called
selection task, which is taken to be the primary challenge for any computational
5 See section 9.3.2 for further discussion.
6 Nothing precludes dening a semantics for the resulting formulae which involves a mapping
from these formulae onto situation-theoretic constructs, or onto DRT congurations, or onto
expressions in (some extension of) Dynamic Predicate Logic. However, it is notable that the
motivation for such less orthodox semantics is very considerably reduced, once the dynamic
element is incorporated into the syntactic system.
Well-formedness and Availability of Interpretations 263
or psychological model of the parsing process. To transform the current formal-
ism into a parser in this more orthodox sense, we would have, minimally, to
dene the higher-level structures needed to represent a means of evaluating al-
ternative interpretations as they become available, with some articulation of the
relationship between formulae annotating the nodes in the tree and the `chunks'
of world knowledge which these terms make available, on the basis of which some
appropriate choice can be made. But this we have not done: we have not taken
it to be part of our remit to articulate how hearers actually make choices as the
information they process unfolds within a given overall context, on the grounds
that this is the remit of pragmatics. As Sperber and Wilson (1995) argue in
detail, this is the central task which a pragmatic theory has to address; but
the construction of a pragmatic theory was not our goal. Rather, what we set
out to do was to dene the unfolding structures relative to which a modelling of
the selection task could be formally addressed. In so far as this more restricted
goal is successful in characterizing properties of natural language, we take this
decision to be vindicated, conrming that this delimitation of responsibilities
between a formal model of linguistic structure and a pragmatic theory of ut-
terance interpretation is appropriate. Indeed, we should note at this juncture
that the general properties of natural language processing that we have elected
to model correspond very directly to informal assumptions made by Sperber
and Wilson (1995), making a hearer-directed theory of pragmatics such as Rel-
evance Theory the most natural direction in which to turn for such a pragmatic
account.7
In separating the grammar formalism from a model of the selection task
facing a hearer in parsing a string, we retain the means of distinguishing well-
formedness from acceptability. The characterization of well-formedness of a
sentence turns on there being some set of actions licensed by the system that
makes available at least one pairing of a string with an interpretation. Accept-
ability of a sentence, on the other hand, is a psychological concept involving
speakers/hearers and how they judge the possibility of pairing a presented string
with an interpretation. As is familiar, sentences may not be acceptable to an
informant for reasons which are irrelevant to the articulation of properties of
the linguistic system. In presuming on some pragmatic theory to explain the
selection task in terms of general constraints on the retrievability and manipula-
tion of string{interpretation pairs, our system can re ect the distinction between
well-formedness (= grammaticality) and acceptability, like any other framework.
The only dierence between our system and others in respect of the relation to
the data is that, in dening a parsing-directed system, we dene a concept of
7 Though we have not explored the problem of disambiguation, it is of interest to note that
upon the presumption of the proposed framework as the background grammar formalism, the
ambiguity problem becomes very dierent. According to the current formalism, the sequence
of parsing actions involves the potential interaction between computational, lexical and prag-
matic actions at every stage in the interpretation process, so that pragmatic considerations
leading to choice of interpretation can be made at any step in the sequence of transitions. The
result is that the number of interpretations out of which selection has to be made, though
large enough, is not some calculation on the total list of possible interpretations, but, at any
one step, involves a choice out of whatever alternatives remain outstanding at that point.
well-formedness which is more liberal with respect to the set of strings licensed,
since whenever an informant characterizes a string as acceptable in virtue of
there being a string{interpretation pairing, we would characterize such a string
as well-formed (unless there was strong evidence to the contrary). So, to take
one type of example prominent in chapter 6, we would characterize (8.1) as well-
formed, on the grounds that it is acceptable under certain circumstances:8
(8.1) The man that Sue said he needs a holiday is chairing the meeting.
In short, the relationship between theory and data is much as in other frame-
works. Despite the fact that we are promoting a grammar formalism which sets
out a dynamic architecture for parse sequences, it is none-the-less possible to
dene a clearcut concept of well-formedness and describe a further concept of
acceptability which re ects some interaction between a grammar formalism and
pragmatic constraints on retrievability of interpretation.
8.4 Universals and Language Variation
A further re ection, given a broad Principles and Parameters or Minimalist
perspective, is that the analyses given for any one phenomenon have, disap-
pointingly, invoked dierent forms of rule in each language considered. To take
one example, far from there being a single LINK transition rule capturing what
is common to relative clauses, to the contrary, there may be more than one LINK
transition rule in even a single language (as in English). And although the pair
Introduction (which introduces requirements for daughter nodes) and Prediction
(which licenses the construction of such nodes) might seem good candidates for
a set of universally applicable rules, not all language-particular analyses have
made use of them. Japanese, in particular, was dened as inducing the full se-
quence of constituents from the absolute root by lexical action. And in English,
these rules were used only to introduce the subject and predicate nodes. Even
lexical items of the same syntactic category do not necessarily project the same
sequence of actions. Though transitive verbs, for example, have been assumed
to project a formula of type e ! (e ! t), they vary freely with respect to other
accompanying actions they induce. In some languages they have been dened
to initiate a set of actions from a requirement ?T y(t), as in Arabic (in Japanese
even building rst an unxed node of that requirement), in other languages
from a requirement ?T y(e ! t), as in English; and, as suggested in passing
8 One problem in evaluating judgements of acceptability in borderline cases is that infor-
mants may make dierent judgements according to whether they place themselves in the role
of hearer or in the role of speaker. Examples such as (8.1) are treated as unproblematic in
language use, occurring regularly in conversational English, but nevertheless may be reported
as a form which the informant as speaker might avoid. All such cases would be included by us
in the set of data to be characterized as well-formed, unless to do this would lead to internal
inconsistency in the system, as in the case of agreement mismatches which all informants
agree are unacceptable despite being understandable:
(i) The men is seeing my mother.
Unlike (8.1), examples such as (i) do not occur regularly in conversational English.
Universals and Language Variation 265
in chapter 5, we might dene German verbs as initiating a set of actions from
a condition ?T y(e ! (e ! t)) (presuming on Introduction and Prediction to
induce the full template of structure for a predicate and its arguments). More-
over, the characteristic pattern of lexical entries is a disjunctive specication,
projecting dierent actions in dierent environments.
A follow-up question re ecting on this rich lexical diversity might well be:
in what sense does the syntactic explanation oer anything more than a het-
erogeneous list of dierent forms across dierent languages? The answer to this
question lies in the constraint that the process of natural language interpre-
tation is by denition goal-directed. The interpretation of a natural language
string involves goal-directed construction of an annotated tree structure, guided
by the sequence of words, the goal being to yield a composite lambda term of
type t constructed out of other such terms (not expressions in the string under
interpretation). The form of that goal, its initial statement as a goal to be
achieved, and the format of the increasingly rich tree descriptions which consti-
tute the transition from the assigned goal to its achievement are all determined
by the particular tree-description language, together with the selected systems
for representing node decorations, being a combination of LOFT, a variant of
the lambda calculus (with term-binding operators) for expressing formulae, and
other labelling devices. To express the dynamics involved in such a sequence of
transitions, the vocabulary for describing node decorations includes both anno-
tations and requirements. Moreover, the decision to adopt a modal logic as the
tree-description language, with h"i, h#i, hLi, hL 1i, h#i, #, h"i, ", hDi, hU i,
means that requirements can be dened which are locally imposed on a certain
node but may nevertheless be fullled by annotations on some node which is
only related globally to the node at which the requirement is introduced. In
addition, individual annotations may be partial, requiring substitution by some
xed value in all such cases to yield a completed tree. Thus, for example, the
language for representing Formula values has meta-variables U, WH, etc., and
the tree-description language has h#i, hDi, hU i, these being devices for express-
ing concepts of incompleteness of annotation along the dimension provided by
the particular label, F o and T n, respectively. Finally, given the gap between
the imposed goal and its achievement, the processes which develop tree struc-
tures are dened so as to impose further requirements, project annotations, and
compile these so that the goal can be attained. The framework, that is, dictates
that natural languages be described as sets of strings interpreted by an incre-
mental left-to-right process monotonically constructing partial decorated trees
to yield a logical form; and the structural properties of a language are dened
in terms of the contribution made to this process by words in the language in
combination with general transition rules. It should be noted that there is no
unrestricted process of type construction, so that there are only a nite number
of Type annotations available for node decoration. In other words, there is no
general process of type inference, as in categorial grammar formalisms.
What varies within the space of possibilities provided by these assumptions
is how the goal imposed on the interpretation of a particular string is achieved
in any one language, and, on this, there are no additional a priori restrictions.
All lexical items express idiosyncratic input to the goal-directed construction

process, and, since rules express transitions not projected from the lexicon,
languages dier in the particular form of general transitions they allow. This
gives us a very dierent perspective on language variation than is provided
by the Principles and Parameters framework, for variation across languages,
on the present view, is restricted neither to the lexicon nor to a particular
type of constituent.9 There is no restriction on whether some daughter node is
built by Prediction or by some lexical action, no restriction on whether some
LINK ed node is built by general rule or lexical action, and so on. Each of these
possibilities may be realized in dierent languages, or in a single language. In our
model, English, Japanese and Arabic all vary in the LINK transition processes
they make available, and each has more than one such transition, whether in the
form of lexical items or as an additional LINK Introduction rule.10 Languages
also vary in the types of action associated with lexical items. Thus verbs,
albeit all projecting some n-place predicate, nevertheless vary across languages
as to their triggering condition and other actions. Indeed, the characteristic
pattern of lexical items is that they may have slightly dierent eects in dierent
environments, and are dened as projecting a core set of actions with additional
actions in specic contexts. This permissiveness gives rise to the phenomenon
we saw in chapter 5 in looking at German expletive wh forms, that a language
may have variants which share the same set of words, and project from them
the same output structures, but nevertheless dier in the way they project those
structures, and hence dier in the strings they license.
The advantage of this exibility is that we have a very ne-grained approach
to variation, both within an individual language and between languages. None-
the-less, a basis for characterizing universal properties of natural language is
preserved. The properties of natural language systems that are universal are
the form of the resulting structures and the general dynamics of the process of
tree growth. The variation lies in the particular transitions that lead from the
linear sequence of words in the individual system onto such structures.
8.5 On Knowledge of Language
In closing, now, we tentatively indicate what we take to be issues of primary
interest to philosophers in what we have set out in this book. We have a frame-
work for dening a grammar as a process of parsing language strings. There
are two consequences to this departure from the orthodox assumption that ar-
ticulation of language is independent of the tasks of speaking/hearing to which
it might be put. The rst is that parsing is seen as the basic property of a
language system. The concept of a language system is similar to the Fodor-
ian concept of an input system (J.A. Fodor, 1983), but, in being made up of
9 Within the Principles and Parameters framework it has been argued that parametric
variation is restricted to functional categories (see Ouhalla 1991).
10 Given the means to dene varying forms of controlled Modus Ponens, in principle we
could also have dierent corresponding forms of Prediction and Introduction.
On Knowledge of Language 267
computational, lexical and pragmatic actions, is not encapsulated in Fodor's
sense. As in Fodor's system, nothing we have dened bears any relation to the
production process. Production has to be characterized as in some way para-
sitic on the process of parsing. This is arguably not implausible; but at this
juncture, we leave the debate to others, merely noting that a re-evaluation of
the competence{performance distinction is demanded in view of the asymmetry
between the relation between the grammar and parsing on the one hand and
the grammar and production on the other.
The second consequence of this departure from the orthodox perspective is
the need to reconsider how we should view our language capacity. We have
taken a language to be a set of procedures, partly dened by general principles,
partly dened by idiosyncratic lexical specication. The overall characterization
takes the form of an action metaphor. This account provides no grounds for
seeing a grasp of language as involving in any sense a set of structural axioms
which a language-user must know independent of any application or use: our
so-called `knowledge' of language is simply the capacity to construct appropriate
decorated structures as interpretations of stimuli of a particular sort. This is, of
course, only half the story, for the explanation of how the language capacity is
related to language use must emerge in conjunction with the pragmatic theory
with which such an account needs to be paired. We leave all these as open
questions, noting merely that the story to be told will necessitate a philosophy
in which language competence is seen as directly related to language use.
9
The Formal Framework

9.1 Introduction
This chapter will outline the formal framework within which the treatment of the
previous chapters is to take place. In particular, it will supply the formal cadre to
the informal introduction of chapters 2 and 3. As is the case in these chapters, we
deal rst with the declarative part of the model, the decorated partial trees
in which the results of the parsing process accumulate, and subsequently with
the procedural part of the model, the actions which transform one decorated
partial tree into the next one.
The basic abstract model in which to describe the declarative and procedural
interaction in a dynamic system is that of Propositional Dynamic Logic (PDL).
This model consists of a non-empty set of states, and two sets of variables to be
interpreted in these states: a set of propositional variables { which may `hold'
at states, and thus can represent the declarative side of a dynamic system, and
a set of action variables { which may hold at pairs of states, namely all pairs
where the action can take the rst state to the second, and so can represent the
procedural side of a system. A P DL model is thus given by a triple
PT = (P T; Vp ; Va )
where P T is the non-empty domain of states,1 Vp assigns propositional variables
from AT OM to states, and Va assigns action variables from ACT to pairs of
states.2
Vp : P T 7! P (AT OM ) and Va : (P T P T ) 7! P (ACT ).
In a familiar notation, for T ; T 0 2 P T and atomic formulae , we will set
T j= to mean 2 Vp (T ) and T < T 0 to mean 2 Va (hT ; T 0 i). Declarative
information 2 AT OM is now said to hold for state T in PT if T j= , and
procedural information is encoded in its re ection in declarative information
1 Where, to avoid an overload of notation, we let P T stand for `partial trees'.
2 The symbol `P ' denotes the powerset operator.
268
Introduction 269
where for 2 ACT and a formula `based on' AT OM , we have T j= [] if
for every T 0 2 P T : if T < T 0, then T 0 j= .
As usual, Vp can be lifted to assign to elements of P T the formulae of a
language DU , the closure of AT OM under the Boolean operations, quantiers
and modalities, and Va can be lifted to assign to pairs of states in P T the actions
in A(ACT ), the closure of ACT under the regular operations, plus test, by the
standard P DL truth denition.
To t our model in this mould, we have to keep in mind that the actions in our
model are generally to be interpreted as constructions that contribute to a
logical form, and that these forms constitute some `maximum' for these actions.
This motivates the following specialization of the standard P DL model.
Specialization 1: a partially ordered set of states
On the domain P T we introduce a binary relation , the development relation,
where is a (discrete) partial order3 with a unique minimal element T and
a set of maximal elements P T+ such that for all T 2 P T there is T+ 2 P T+ :
T T T+ . The set P T+ represents the objects to construct (e.g. logical
forms), and the elements of P T are partially constructed such elements. The
actions of P DL are now seen as constructions. In other words, for all 2 A,
< :
That is, the states in P T+ are the objects to construct. The actions incremen-
tally construct these states, starting from T .
If we consider a path T : : : T : : : T+ 2 P T+, then the parts of the
maximal states4 under construction are exactly the persistent features; that is,
the formulae such that T j= and T T 0 implies T 0 j= . There may also
be such that T j= which do not persist in this way. Such formulae represent
transient features of states and may function to control actions, but they are
not parts of the objects under construction.5
The maximal elements in P T+ (the set of logical forms) divide the (parsing)
3 An order is discrete if, whenever T T 0 ; T 6= T 0 , then there is a T 00 : T T 00 T 0 such
that for every T 000 such that T T 000 T 00 , we have T 000 = T or T 000 = T 00 : if an element
has a successor, then it has an immediate successor.
4 The parts of the maximal states are trees decorated by declarative units.
5 Partial trees are not the basic elements of our model: the basic elements are the partial
trees plus a pointer; that is, partial trees with one selected node. The elements T 2 P T
now become equivalence classes consisting of collections of all pointer positions over one and
the same tree. On a non-empty domain P P T = fT n j T 2 P T; n 2 T g we place two binary
relations and , where is a (discrete) partial order as above, and an equivalence relation
(between two pointed partial trees which dier only in the pointer position), such that for all
2 A,
< ( [ ):
That is, we consider a P DL model with actions that are incremental with respect to some
partial order on the domain of states. All actions move along , except pointer movements,
which go along . We can identify the elements of P T with the objects over P P T , where
this set is dened by O(P P T ) = f[T n] j T n 2 P P T g. Objects are related through their
elements: [T ] [T 0 ] i for all T 00 2 [T ] there is a T 000 2 [T 0 ] such that T 00 T 000 .
Moreover, if T T 0 , then [T ] [T 0 ] .
270 The Formal Framework
paths in P T starting from T1 into `correct' and ìncorrect' ones: correct pathways
are those that have continuations reaching P T+; incorrect pathways are then
the ones that don't. A formalization of this goal-directedness is our second
addition to P DL.
Specialization 2: requirements
The addition of requirements to the P DL model with a partially ordered domain
of states consists in the addition of a (third) function V? besides Vp and Va ,
V? : P T 7! P (DU )
where we set T j=? if 2 V? (T ), which is intended to mean that is `required'
at T (note that this function is not dened by induction on AT OM ).
The function V? satises the following constraint:
T j=?
which implies
T 6j= ,
if T T 0 and T 0 j=?, then 8T 00 (T T 00 T 0 ) T 00 6j= ).
So a requirement for at T persists along until a T 0 is reached satisfying
. Note that may itself be non-persistent or persistent. Moreover, all non-
maximal states T (i.e. T 62 P T+) must have outstanding requirements. This is
what goal-directedness means in our model:
fT 2 P T j V? (T ) = ;g P T+
On the other hand, the minimal state cannot be without requirements. The
transition from state to state must be driven by the need to full requirements:
V? (T ) 6= ;
Apart from an ordering on the P DL states and the addition of a third `valuation'
function, our parsing model specializes the general P DL framework further, in
that the states themselves are structured objects (decorated partial trees), and
the basic actions are not arbitrary, but are specic to the states of our model.
Specialization 3: structured states and specic actions.
Abstractly, the parsing process of a natural language string w can be viewed as
establishing an association between w and a set S of terms (formulae) in some
logical language, the possible interpretations of w. In any parsing framework,
this interpretation is likely to be the result of applying some (set of) basic
operation(s) O to the representations contributed by the words of string w. In
fact, the representation is generally a normal form under these operations.
This means that the association between string w and set of formulae S can be
broken down into associations between the individual words of w and sets of
representations labelling the frontier of a binary tree the top node of which
is labelled by S :
Introduction 271
O(X3 ;O(X1 ; X2 ))
X3 O(X1 ; X2 )
X1 X2
For instance, we may take X1; X2; X3 to be feature structures and the operation
O to be unication, or X1 ; X2 ; X3 to be lambda terms and O Application, or
X1 ; X2 ; X3 to be labelled categorial expressions and O Application: Modus
Ponens, or X1; X2; X3 to be DRS s and O Merging. In all these grammatical
frameworks the construction of terms in a particular parsing process can thus be
represented as the growth of decorated binary trees. In this book we will take
the operation O to be function application in a typed lambda calculus, and the
objects of the parsing process, that is the elements of P T+ , will be terms in this
calculus together with some labels; but it is important to keep in mind that the
choice of the actual representation language is not central to the parsing model
developed here.
So the maximal states in P T+ in our model will take the form of labelled
formulae of a typed lambda calculus. The labels record, among other things,
type information and linguistic features like tense, agreement, etc. A formula of
the lambda calculus will be represented by a decorated binary tree as follows.
The sentence John read a book, represented by the formula Read(John, Some(x,
Book(x))), can be seen as resulting from the term
APL(APL(xy [Read(y )(x)]; APL(P (SomeP ); (x; Book (x))); John))
by -reduction. The obvious tree structure of this term we will exhibit in the
form of a bracketed formula:
[0 [0John] [1[1 xy[read(y)(x)]] [0 [1P (someP ) [0(x; book(x)) ]]]]]
where 0 represents the argument and 1 the function daughter. Tree and decora-
tions can be considered independently by pulling them apart6 (gure 9.1). This
supplies the abstract point of view of the following sections.
[|0 [0 ] [1 [1 ]{z[0 [1 [0 ]]]]]}
T ree
|f00:john, 011:xyread, 0101:{zP (someP ), 0100:(x; bookx)g}.
Decorations
Figure 9.1: A term as a labelled tree

On such a tree, the destructor APL, which applies a function to its argument,
can be seen as a modal operator, where tree node n veries APL(1 ; 2 ) if 2 ,
6 We use function application to model the relation between head and complement. We
use update functions to represent the relation between head and adjunct.
the function, holds at the function daughter, and 1 , the argument, holds at the
argument daughter of n. Now, -reduction APL(1 ; 2 ) = , can be represented
by (instances of) the axiom scheme APL(1 ; 2) ! , relating the propositions
on function and argument daughters of n to propositions at n (see Blackburn
and Meyer-Viol 1994).
The minimal element T in this set-up is a one-node tree decorated by a
requirement for an annotation of a formula of type t { that is, a requirement for
a proposition. The parsing process which maps T eventually to a decorated
tree in P T+ creates on the way decorated partial trees. These will constitute
the domain P T of our model.
Now, we will view parsing as the construction of a (set of) decorated binary
trees, logical representation(s), guided by the input of a natural language string.
This can be modelled as (incremental) growth of a (set of) decorated tree(s).
This is modelled by the relation on the states. In this book we will assume
that this construction takes place while traversing the natural language string
from left to right and in a word-by-word fashion. Every word maps a partial
decorated tree to an extension, a decorated partial tree in which it can be em-
bedded. The input string licenses the growth of a particular collection of trees
in the forest of possible trees. From left to right, every word in the string may
induce a specic extension of the current (set of) partial tree(s) (sentence-initial
John, The), or it may more generally constrain the possible completions of the
current partial tree (John who : : : , John is going to project some argument in
the tree constructed by the relative clause). There are various components to
this growth relation: the growth of tree structure, the growth of tree decorations,
and the disappearance, i.e. fullment, of requirements.
First, a tree structure may be extended by growing new branches, extending
old ones, etc. But the tree structure may also be underspecified: it may be
underdetermined whether some daughter is a function or argument daughter {
thus we may have [n : : : [# : : : ] : : : ] representing the relation of immediate
dominance without having more specic information. Also, the tree may be
underdetermined in the fact that only knowledge about dominance is involved
{ denoted by [n : : : [ : : : ] : : : ]. In such a partial tree, the fact that one node
immediately dominates another one does not (yet) mean that the second one is
a function daughter or an argument daughter of the rst one, and that one node
dominates another one does not mean that these nodes are related by a sequence
of immediate dominance steps: this holds only in the (complete) binary trees in
which it can be embedded.
Secondly, in the course of the traversal of the natural language string the for-
mulae decorating the nodes of the tree under construction may grow in number
and also in information content. That is, they may be updated and, in particu-
lar, variables may be substituted.
Thirdly, the growth of a partial tree is not arbitrary. The string John read
: : : leaves us with the structure
[0 [0 ], [1[1],[0 ]]] f00 : John; 011 : xyRead(y)(x)g:
Declarative Structure 273
This structure is a partial decorated tree that cannot be reduced to a lambda-
free term: at least one argument is still required to be able to reduce the tree to
a term in the representation language. That is, this partial tree restricts the set
of possible completions to the one where an object to read is supplied (see gure
9.2). Here is where the function V? comes into play. On the way to the complete
[|0 [0 ],{z[1 [1], [0]]]} f| 00 : John; 010 : xy[read
{z
(y)(x)]; 011 :?T y(e)g}
T ree Decorations
Figure 9.2: Partial tree with requirements

term we must create partial objects. In order to construct a term l in stages, we
consider the term structure of l and represent l as a binary tree annotated by the
appropriate sub-terms. Within the set of all such structures there will now exist
annotated partial trees that cannot be extended to structures corresponding to
actual terms. Thus, if we represent the construction of a term as a process of
annotated tree formation, then it must be represented as a process directed
towards the annotated partial trees that correspond to actual terms. This goal-
directedness we represent by the notion of a requirement function V? , which can
annotate the nodes of a partial tree. Now the basic relation of development
of such structures is that of constructing extended tree structures which full
requirements. Ideally, the requirements are fullled if and only if the partial
tree represents a well-formed term, an element of P T+.
We take the natural language string w to be a sequence of instructions
to construct the term that is to be the representation of the content of w.
Here is where the actions in A(A) have their role. Because the actions in our
model derive directly from the natural language under consideration and are
intimately connected to the decorated trees they create, we will not base the
action language of our model on a set of arbitrary action variables, but we will
dene a set ACT of constants specic to decorated trees. That is, for every tree
relation we have an action creating a new node standing in that relation to the
current node and an action moving the pointer to that node; for every formula
of our language we have an action which decorates the current node with that
formula; and nally we have an action merging two nodes { that is, identifying
two nodes and accumulating the decorations of both.
9.2 Declarative Structure
9.2.1 Feature-decorated Tree Construction
In this section we will give abstract representations for trees and decorated
partial trees which can develop in a goal-directed way. On these trees we will
then interpret terms represented as decorated trees and the growth of terms as
development of these trees.
9.2.1.1 T -structures
The decorated binary trees that will serve as our representations of logical forms
form a subset of the decorated partial trees. These, in turn, are special kinds of
decorated T -structures. In this section we formalize the structural parts of the
decorated structures, i.e. bare T -structures, partial trees and binary trees; the
next section, 9.2.1.2, will concentrate on decorations, and section 9.2.1.3 will
then deal with the structures supplied with decorations.
Denition 16 (T -structures) A T -structure is a tuple
T = hT; 0; 1 ; L; # ; ; D i
where T is a nite non-empty set and all of the i are (possibly empty) binary7
relations on T .
So, a T -structure consists of a non-empty domain on which there are a number
of binary relations dened. Without further ado, these relations do not struc-
ture the domain as a binary tree. This will have to be taken care of by a next
denition.
We let I = f0; 1; L; #; ; Dg, and I (as usual) will be the set of all nite se-
quences of elements of I (including the empty one). For j 2 I and i 2 I we set
ji =j Æ i .
Structural similarity of T -structures is measured by T -morphisms. A function
f mapping T into T 0 is a T -morphism from T -structure T into T -structure T 0
if for all n; m 2 T , for all i 2 I : n i m ) f (n) i f (m). A function f is a
T -isomorphism from structure T to T 0 if f is a bijection from T onto T 0 and
both f and f 1 are T -morphisms.
Denition 1 (Binary Trees) A T -structure T is an (ordered) binary tree if
hT; # i is a directed, non-circular graph such that T has exactly one ele-
ment without # predecessors, and every element with # successors has
exactly two such successors,
0 and 1 are partial functions such that # = (0 [ 1 ) and
(0 \ 1)= ;,
is the re exive, transitive closure of # .8
A T -structure T is a binary linked tree if T is the disjoint union of a nite set
of binary0 trees such0 that n L m, for m 2 Ti 2 T , implies that n 62 Ti and
n L m for all m 2 Ti . We dene D as follows: n D m i there is a
0 k ! such that n i1 m1 : : : i 1 mk 1 i m, where for all 1 j k:
ij 2 f#; Lg (D may relate nodes from distinct binary trees).
k k
The set of T -structures that are binary linked trees will be denoted by BT .
7 The symbol `' will be used to indicate the end of a denition, remark, etc.
8 We will also nd use for the relations 0 and 1 which are the re exive and transitive
closure of 0 and 1 respectively.
Remark 1 Notice that in linked binary trees we have
0 ; 1 #D and L D :
From right to left we have increasingly `specied' tree relations. The most
specic are the 0 and 1 relations. Both these relations are immediate dominance
relations (#), and immediate dominance relations again are dominance relations
() which, in their turn, are D relations, possibly involving more than one tree.
The 0 and 1 daughter of a node will be called the argument and function daughter
respectively. The result of applying the formula annotating the 1 daughter of n
to the formula at the 0 daughter is deposited at n itself.
Apply(Function, Argument)
0 1
Argument Function
For n 2 T , the local tree (global tree) of n is the smallest set T R such that n 2 T R
and if m 2 T R, n j m or m j n, j 2 f0; 1; #; g (j 2 f0; 1; #; ; L; Dg) then
m 2 T R. So the local tree of a node n consists of the binary tree it is a node of.
Now we are set to introduce the basic structural objects of our model: partial
trees. These are the structures underlying the partial logical forms the parser
constructs.
Denition 2 (Partial Trees) The set P T , of partial trees, consists of all T -
structures T such that there is a T -morphism mapping T to an element of BT ,
a binary linked tree.
So a partial tree is a T -structure that can be embedded in a binary linked
tree. We are going to conside the partial trees dynamically as `states' in a
construction process. This construction process will take place at the nodes of
a partial tree; that is, we need a tree together with a pointer indicating some
node.
Denition 3 (Pointed Partial Trees) A pointed partial tree (ppt) T n con-
sists of a pair of a partial tree T 2 P T and a node n 2 T , the domain of T . The
set of all ppt will be denoted by P P T . The System of ppts, PPT , is the tuple
PPT = hP P T; fij i 2 I gi
where T n i T 0 n0 if T = T 0 and n i n0.
Partial trees are related by the basic relation of extension, development or
growth.
Denition 4 (T -extensions) A ppt T 0 n0 T -extends the ppt T n, notation
T n T T 0 n0 , if there is a T -morphism f mapping T to T 0 such that
f (n) = n0 .
Notice that T is a pre-order on P P T (i.e. it is re exive and transitive). This
relation on pointed partial trees can be lifted to the partial trees; for T ; T 0 2 P T
we set T T T 0 if for all n 2 T there is a n0 2 T 0 such that T n T T 0n0.
Remark 2 (Examples of T -extensions) If T n T T 0n0 , then it may be that

T 0 extends the frontier of T or constructs elements above the top node.
A A A E
T and T
B C B C B C D A
D E B C
It can also mean that adds to T a completely unconnected partial tree,
T0
creating, in eect, a set of two disjoint trees. Moreover, in T 0 nodes of T may
have been unied (the f0 -images of two distinct elements of T coincide, under
T -morphism f : T 7! T ).
A A
0 T (*,0)
[B ] C [B; D] C
*
D
Tree growth need not involve addition of tree nodes. In a T -structure T
some ùnderspecied' tree relation may be further specied. For instance, a
pair n; m 2 T such that n D m may be mapped by f to a pair such that
f (n) f (m) and later, skipping # , by some g, to a fully specied relation
g(f (n)) 0 g(f (m)):
A A A
D T (D,*) T (D,*,0)
B C B C B C
Compare this sequence to the following growth stages:
A A A
0 T (*,0) T (D,*,0)
B C B C B C
Here, no actual growth takes place, although each new T -structure is a proper
extension of the previous one. In contrast with the previous growth sequence,
only conclusions of bookkeeping activity are added. Once we have T n 0 T m,
then it is a matter of bookkeeping to add the pair hT n; T mi to the relations,
# ; and D . So, independent of annotations on tree nodes, there is a notion
of growth or development towards complete structures.
Remark 3 (Partial Trees in Normal Form) The partial trees in P P T will,
in our model, all occur in normal form. This form has to do with the under-
specied tree relations #; and D . Consider the situation where we have a
node with a `dangling' daughter and a second node with a daughter.
(X ) A (Y ) A
U #
B C
Then there are three normal extensions embedding (X ) and (Y ):
(Z1) [A; B ] (Z2) A (Z3) A
# (#,U) #
C or [B; C ] or U C
U
B
That is, either the dangling node B unies with one of the nodes A or C , or it
comes to dangle below the bottom node.
Normal form constraint I Suppose we have T n1 j T n2 for j 2 f; Dg.
If tree T is in normal form and we have T n1 i T n0 for i 2 f0; 1; L; #g, then
there is some T n3 such that T n1 i T n3, and one of the following three is the
case: T n2 = T n1, T n2 = T n3, or T n3 j T n2.
When a structure is in normal form, then an underspecied tree relation is
pushed as low is possible. So, a structure of the form
(Z ) A
* #
C B
is not in normal form. But notice that Z is an -lower bound of the normal
forms Z1; Z2 and Z3. We will consider P P T only to contain trees in normal
form. By proposition 1 clause 2 below, this does not exclude any completions
to full binary trees, for the three possibilities exhaust the possible embeddings
of (X ) in an eventual complete binary tree.
With respect to the relation #, there is the following constraint. Given the
structures
(Y ) A (Z ) A
# and 0 1
B D E
there are only two possible embeddings in normal form
A A
(#,0) 1 or 0 (#,1)
[B; D] E D [B; E ]
Normal form constraint II If a tree node T n1 has a function daughter and
an argument daughter, then the relation T n1 # T n2 entails either T n1 0 T n2
or T n1 1 T n2.
That is, in case there are function and argument daughters, then # is not an
underspecied tree relation, but an underspecied description of a fully specied
relation. So, the structure
A
# 0 1
B D E
is not in normal form. Notice that every complete binary tree is in normal form
{ that is, satises both constraints.
A pointed partial tree T n is a pointed binary tree if the set fn0 2 P P T j
T n T n g (i.e. the sub-tree of T n with top node n) is a binary tree. These
0
pointed binary trees, elements of P BT , represent a complete sub-term structure
of some logical form under construction. These objects are the goals of our
parsing construction. Thus, the growth of structure of logical forms is handled
in a structure described by the triple
hPPT ; T ; P BT i:
where PPT is a proper subset of the set of all pointed partial trees, and the set
P BT consists of T -maximal elements; that is,
1. for every T n 2 P P T , there is some T 0n0 2 P BT such that T n T T 0 n0,
2. if T n 2 P BT and T n T T 0n0, then T 0 n0 T T n.
Proposition 1 Every structure T n in P P T can be identied with a nite set
of Nor(T n) (of normal forms), the elements of which satisfy constraints I and
II, such that
1. T n is a -lower bound of Nor(T n),
2. if T n 00 T000n0 and0 T 0 n0 2 P BT , then there is some T 00n00 2 Nor(T n) such
that T n T n ,
3. if T 0n0 satises constraints I and II and T n T 0n0 , then there is T 00n00 2
Nor(T n) such that T n T n .
00 00 0 0
For instance, Nor(Z ) = fZ1; Z3g. (Notice that Z3 Z2, but Z2 6 Z3.)
9.2.1.2 F e-structures
The bare pointed partial tree structures are going to be decorated with informa-
tion in the form of labels and formulae. We will dene these decorations in the
general format of feature structures.
A feature structure is a labelled decorated directed graph. The elements of
feature structures will be called points, and to say that feature structures are
labelled directed graphs simply means that one can envisage the points being
linked by arrows bearing a label. The most important constraint on feature
structures concerns these labelled arrows: there is no point in any feature struc-
ture from which two distinct arrows bearing the same label emerge. In eect,
labelled arrows are representations of partial functions; these partial functions
are usually called features. As to the decorations, we envisage the points of fea-
ture structures being adorned with pieces of information of linguistic interest.
Let's make this precise. We assume signature hF ; D; !i where F and D
are non-empty nite or denumerably innite sets, the set of features and the
set of decorations respectively, and ! : F 7! P (F [ D) is a function satisfying
8f 2 F : !(f ) 6= ;. The function ! assigns the sub-features or values a feature
can take.
Typical elements of F , as used in this book, might be Ty, Fo, Tn, standing
for `type', `formula' and `tree node' respectively, and more linguistic features
like number, person, and tense; while typical elements of D might be e ! t
(2 !(T y)), Walk(Man) (2 !(F o)), 010 (2 !(T n)), and the more usual singular,
plural, 1st, 2nd, 3rd and present. A computational linguist would probably
depict a feature structure as follows:
2 3
Type e!t
4 Formula xW alk(x) 5
Tree Node 01
Here the left column displays elements from F , and the right column elements
from D. Such two-dimensional objects are called Attribute Value Matrices
(AVMs).
As in the case of T -structures, our target decorations, complete feature struc-
tures, declarative units, are special kinds of partial feature structures which are,
in their turn, special kinds of F e-structures.
Denition 5 (F e-structures) An F e-structure of signature hF ; D; !i is a triple
F of the form hU; fRf gf 2F ; fQdgd2D i, where U is a non-empty set; for all f 2 F ,
Rf is a binary relation on U that is a partial function; and for each d 2 D, Qd
is a unary relation on U .
The above AVM can be displayed as an F e-structure of the following shape

RT y RTn RFo
e!t 01 xW alkx
A function g mapping F e-structure F to F0 is a F e-morphism if for all u; u0 2U
if uRf u0, then g(u)Rf0 g(u0), and if u 2 Qd, then g(u) 2 Q0d. We say that F e-
structure F0 extends structure F, notation F F F0 , if there is an F e-morphism
mapping F into F. A function g is an F e-isomorphism from structure F to F0
if g is a bijection and both g and g 1 are F e-morphisms.
For instance,

RT y RT n F RT y RTn RFo F RTy RTn RFo
U 01 e!t 01 e!t 01 xW alkx
In the second F e-structure, a meta-variable U is updated to the type e ! t,

and a third feature dimension is added, but without a value; in the third F e-
structure, such a value is supplied. An F e-structure F is a feature structure
if every feature has a value which is either a (sub)-feature or some decoration:
so F e-structure F is a00 feature structure0 if uR00 f u0 implies that either there is
g 2 !(f ) \ F and a u 2 U such that u Rg u , or there is a d 2 !(f ) \ D and
u0 2 Qd . We let F S be the set of all F e-structures which are feature structures.
In order to accommodate the construction of binary trees. we have dened
partial trees to represent the stages on the way to the complete product. For
the same reason, we will introduce partial feature structures as approximations
to complete feature structures.
Denition 6 (Partial Feature Structures) An F e-structure F is a partial
feature structure if there is a feature structure F0 such that F can be F e-
morphically embedded in F0 .
Remark 4 (Examples of F e-extensions) Each of the following examples in
AVM notation highlights a dierent notion of partiality, with a corresponding
notion of incrementation.
Expansion exemplies incrementation in its most literal form:

Formula John
Formula John F Type e
Substitution of a value for a variable should not count as a violation of incre-
mentality (even though, in a literal sense, a formula is replaced by another one.)

Formula U F Formula John
::: ::: ::: :::
Inferences incorporate computations we allow on feature values. We will make
very restricted use of this possibility, and every application will have to be ex-
plicitly licensed.
2 3 2 3
Formula Formula
4 Type t 5 F 4 Type t 5
::: ::: ::: :::
where ` !
Type Reduction and Normalization can also be rendered in this fashion,
2 3 2 3
Formula AP L(xW alk(x); John ) Formula Walk (John )
4 Type ModP on(e ! t; e) 5 F 4 Type t 5
::: ::: ::: :::
Operations on decorations include a partial join operation tF and a meet op-
eration uF .

Formula U Formula John
Lab a tF Type e
2 3
Formula John
=4 Type e 5
Lab a

Formula U
Lab a uF Formula John
= Formula U
Just as the pointed partial tree structures in P P T can reach maximal elements
in P BT , so the partial feature structures can be developed into declarative units
DU F S . The elements of DU are feature structures that are in some sense
`complete'. For instance, we have the stipulation that no value of the Formula
feature can occur without a value on the Type feature, so [F ormula John] is an
element of F S but not of DU . The declarative units have in our set-up the role
that sort-resolved and totally well-typed AVMs play in HPSG. Consequently we
can describe the space of partial feature structures by the triple
hP F S; F ; DU i
where P F S is a set of partial feature structures such that hP F S; F i is a meet
semi-lattice and DU P F T is a set satisfying
for every structure F 2 P F S there is F0 2 DU : F F F0 ,
F F F0 for F 2 DU implies F0 F F.
9.2.1.3 FT-structures
In this section we will dene the full structures consisting of pointed partial tree
structures the nodes of which are decorated by F e-structures.
Denition 7 (Feature-decorated T-structures) By a system of (nite, or-
dered, binary) feature-decorated PPTs (fds) (of signature hF ; D; !i) is meant
a tuple hPPT ; (Z; z)i where PPT is a system of pointed partial trees, Z is a
function that assigns to each element T n of P P T a nite, point-generated par-
tial feature structure (of signature hF ; D; !i), and z is a function that assigns
to each node T n of P P T a point z(T n) 2 Z (T n) that generates z(T n).
So at a tree node T n, an F e-structure has the form:9
9 F e-structures in general may have endpoints further than once removed from the gener-
ator. These would represent nested features.
z (T n)
Z (T n) =
a1 a2 ::: :::
Now, apart from extension of tree relations, formalized by T , we can also con-
sider extensions of node decorations. This notion will simply lift the the relation
F , dened for F e-structures, to the nodes decorated by these structures.
Denition 8 (F e-extensions) A ppt T 0 n0 F e-extends T n, notation T n F
T 0 n0 , if there is a F e-morphism g mapping Z (T n) to Z (T 0 n0 ) such that
g(z (T n)) = z (T 0 n0 )
So T n F T 0n0 implies Z (T n) F Z (T 0 n0). Now a tree node T 0n0 F e-extends
T n if the decoration of T n is updated. The relation F will represent the ow
of computation on decorations, i.e. feature values. In general, the values of
labels and formula features decorating the various nodes in one tree are not
unrelated but are constrained by co-occurrence principles. This means that
some combinations of (atomic) decorations are simply not represented in the
class of possible decorations, and some combinations of these decorations cannot
co-occur as decorations within the same tree. Thus, updating the value on one
feature dimension may require change of other values. In particular, if the same
value occurs at various nodes in some tree, then updating this value at one
location must mean updating it everywhere. We will formulate the basic notion
of value sharing:
Denition 9 (Value Sharing) Tree nodes T n and T 0 n0 share values on the
features f F , notation T n f T 0 n0, if and only if T = T 0 and for all f 2 f
and b 2 !(f ) we have
z (T n)Rf b () z (T 0 n0 )Rf b:

This relation of value sharing between nodes we will want to make invariant
under extensions. That is, if we replace a meta-variable U by a proper value
at some pptm T n, then every occurrence of U at every node in ppt T must be
simultaneously replaced by that value.
Combining pointed partial tree systems and feature-decorated pointed par-
tial trees gives us our target structures.
Denition 10 (Decorated Pointed Partial Trees) A structure of decorated
pointed partial trees (dppt) relative to signature hF ; D; !i is a quintuple
hP P T; fij i 2 I g; (Z; z ); ; F BT i;
where
1. hP P T; fij i 2 I gi is a system of pointed partial trees,
2. hP P T; (Z; z)i is a feature-decorated T -structure,
3. =T \ F ,
4. F BT = fT n 2 P P T j T n 2 P BT; Z (T n) 2 DU g.
Here F BT is the set of Feature-decorated pointed Binary Trees, and the order-
ing satises the following constraints:
a. for all f F , if T n T 0n0 and T n f T m, then there is T 0m0 such that
T m T 0 m0 and T 0 n0 f T 0 m0 ;
b. for all T n 2 P P T there is an T 0n0 2 F BT such that T n T 0n0;
c. if T n 2 F BT and T n T 0 n0, then T 0n0 T n.
So the growth relation incorporates both growth of tree structure and growth
of feature decorations.10
10 There are various forms of interaction possible. For instance, typical rewrite steps in-
volving types would be:
[T ype (e; e ! t)] F [T ype t] and [T ype e] F [T ype (e ! t) ! t].
These rewrite steps are known, for instance, from the Lambek Calculus. The typing at tree
nodes will re ect their combinatorial properties, and in particular their status as function
or argument. Thus, from our perspective, there is an essential dierence between the two
rewrites above. The tree structures on which we will evaluate our languages will be ordered
binary trees, and this ordering will consist in the presence of a function daughter (1 ) and an
argument daughter (0 ). This distinction must correspond to the types decorating the nodes.
The second rewrite, however, transforms an argument into a function; that is, an update along
F interacts with tree structure, and thus with the relation T .
In the situation,
0 1
T y (e) T y (e ! t)
no update of T y(e) to T y((e ! t) ! t)) is possible; that is,
T y (t)
0 1 6 0 1
T y ((e ! t) ! t) T y (e ! t) T y ((e ! t) ! t) T y (e ! t)
due to the con ict between the decorations at the nodes and the labelling of the arcs. On the
other hand, suppose we have created daughters only with the immediate dominance relation,
then the update is possible. That is,
# # # #
T y (e) T y (e ! t) T y ((e ! t) ! t) T y (e ! t)
T y (t)
1 0 1 0
T y ((e ! t) ! t) T y (e ! t) T y (e ! t) ! t)) T y (e ! t)
Moreover, identity of label and formula values is preserved under updates.

In particular, update of a variable to some proper term is uniform over the
dppt. Every pointed partial tree decorated by a partial feature structure can be
extended to a full-blown binary tree decorated by a declarative unit.11 Moreover,
once a structure has developed to a structure in F BT , our representation of
complete structures, logical forms, then it has reached its nal form.12
9.2.1.4 The Language DUF
We are now prepared to introduce the language we will use to talk about trees
decorated with feature structures. The language is called DUF , and is the
language DU layered over the language LF denedF in a very simple way. We
simply choose the atomic DUF formulae to be L prime formulae and then
make the DUF will-formed formulae on top of this base in the usual way. As
a result, we've given an ìnternal structure' (viz. a modal structure) to the
propositional symbols of DUF .
It is clear that, as we have dened them, feature structures are multi-modal
Kripke models; and indeed the language LF we now introduce will be the obvious
modal language for talking about them, with the Rf serving to interpret its
modalities, and the Qd interpreting its propositional symbols. However, we
shall continue to call these entities feature structures.
The language LF (of signature hF ; D; !i) contains the following items: all
the elements of D (which we will regard as propositional symbols), a truth-
functionally complete collection of Boolean connectives, and all the elements of
F (which we will regard as one-place modal operators). The set of ws of LF is
the smallest set containing all the propositional symbols (i.e. all the elements
of D) closed under the application of the Boolean and modal operators (i.e. the
elements of F ). Thus a typical w of LF might be the following:
hTyi(e ! t) ^ hFoixW alk(x) ^ hTni01:
This represents the AVM
2 3
Type e!t
4 Formula xW alk(x) 5
Tree Node 01:
Such13Attribute Value Matrices are essentially a perspicuous notation for LF
ws. Our parser constructs entities called declarative units which are in eect
11 This excludes e.g. structures where the relation cannot be resolved due to incompatible
decorations with all accessible nodes.
12 Feature-decorated pointed partial trees are an example of bred semantics { the denition
bres a tree over a collection of feature structures. Finger and Gabbay (1992) have proved a
number of useful results for such systems.
13 Standard AVMs of the form

agreement [person 3rd]
case genitive
nite sets of LF formulae of modal depth 1,
fhLa1 ia1 ; : : : ; hLan ian ; hF oig;
or, in AVM notation, 2 3
La1 a1
6
6
... ... 7
7
6 7
4Lan an 5
Fo
These consist of a nite set of labels annotating a formula. Because in this book
we mostly consider declarative units of modal depth 1, we will generally display
a declarative unit as
fLa1(a1 ); : : : ; Lan(an ); F o()g
instead of the modal variant. For modalities of depth 1, an atomic DUF formula
consists simply of a formula of the form Lai(l), where Lai is some label predicate
in F and l 2 DLa ; that is, l is some appropriate value, or l 2 MV , where MV
is a set of meta-variables, or l = y, where Lai(y) will mean that the label Lai
i
is undened. And, generally, these will be the atomic formulae we will use.
We will, however, give a more general denition which allows decorations with
arbitrary feature structures.
Denition 11 (Atomic DUF Formulae) Let F = hF ; D; !i be a (feature)
signature, let Lai be a letter called a label or a feature identier which can
apply to values a 2 DLa = D \ !(Lai), and let F o be a Formula identier
which applies to values in DF o. Let MV be a set of meta-variables. Now, the
i
set of all atomic DU propositions, consists of all prime LF formulae.

A prime LF formula is an LF formula of the form
hPi1 i : : : hPi ia;
n
where each Pi is either a label identier Lai or a Formula identier F o, and

a 2 MV [ fyg [ (D \ !(Pi )); Pi 2 !(Pi 1 ); : : : ; Pi2 2 !(Pi1 ). Given a subset
j j
of labels L F and their possible values DL = D \ !(fl 2 F j l 2 Lg), we can

n n i
x the language DUL by xing AT OML.

On the basis of the atomic formulae we will now dene the language DUF .
Denition 12 (The Language DUF ) For # 2 f#0 ; #1; #; #; L; Dg and l 2
D [ MV , x 2 V AR, the set DUF of formulae belonging to the Language of
Finite Trees over feature signature F is dened by
::= AT OMF j ? j > j } js j ^ j _ j 7! j h#i j f#g j h# 1 i
j f# 1 g j 9x([x=l]) j (x)([x=l]) j hLF i j [LF ] j 2 j 4.
exhibit a nesting of features. In our modal language, a nesting of modal operators would be
hagreementihpersoni3rd ^ hcaseigenitive:
Typically, here 3rd 2 !(person); person 2 !(agreement) and genitive 2 !(case). Currently
we have found no reason to nest labels and/or features, but the framework as dened here is
able to handle this nesting.
In this denition, the bold variables x; y; z should not be confused with their
non-bold variants. The latter occur in the values of the F o feature, the elements
of DF o, in the eventual logical form.
In the previous chapters we have used the connectives and operators
!; 8x; [#], and [# 1 ]. Here we will introduce them by denition.
! df 2( 7! ) 8x df 2(x)

: df 2 s [#] df 2f#g
[# 1] df 2f# 1g
Denition 13 (Declarative Units) Let DU (a) be a maximal set fLa1(a1 );

: : : ; Lan (an ); F o()g where a = a1 : : : an , ai 2 DLai ; 2 DF o , and La1 : : : Lan
are all the features in F . Let DU (x) = Du1(x) : : : Duk (x) be a xed enumera-
tion of all non-empty subsets of DU (x). The set Dui(x) will always denote the
same set in x and thus V Dui(x) will denote a specic formula and always the
same one. A declarative unit is a maximally consistent Duk (a) { that is, a sub-
set of DU (a) that is an F maximum in DU . Given the xed enumeration, the
sets Dui(x) and Dui(y) will denote the same set of label/formula predicates Lai
applied to variables in x and y respectively. Thus V Dui(x) = V(Dui (y)[x=y]).
To avoid notational overload, we will generally consider Dui(x) itself to be a
conjunction of formulae.
9.2.1.5 Truth
The truth denition that follows is divided into three parts. First, we will
dene truth for atomic DUF formulae. This denition will involve a twist to
guarantee persistence of atomic formulae in an environment where these may be
updated. Secondly, we will give a purely classical truth denition for the Boolean
connectives, the quantiers and the tree modalities. Here we will introduce
connectives and operators that do not occur in chapter 2. The formulae thus
introduced are interpreted completely on the individual partial trees without
reference to their possible extensions. Thus they will generally not persist
along the relation of extension or growth.
To represent the persistent formulae of chapters 2 and 3, we will, thirdly,
introduce a modality `2', interpreted along , such that 2 represents a formula
persistent under growth. By means of this modality, we will be able to define
the persistent connectives, quantiers and modalities of chapter 2.
Denition 14 (Truth for Atomic DUF Formulae) For
2 AT OM = AT OMF [ f>; ?; }g;
the valuation VP 2 P (AT OM )P P T is dened by:
if 2 AT OMF , then 2 Vp (T n) i there is a T 0 n0 : T 0 n0 F T n
& Z (T 0n0); z(T 0 n0) j=F .
where j=F (dened below) is the truth relation for feature structures. For
any feature structure F (i.e. hU; fRf gf 2F ; fQdgd2D[MV ; i) and any point
u 2 U:
F; u j=F d i u 2 Qd ;
F; u j=F y i :9d : u 2 Qd ;
F; u j=F hf i i 9u0 (uRf u0 and F; u0 j=F ):
if = >, then 2 Vp (T n).
if = ?, then 62 Vp (T n)
} 2 Vp (T n) i for all n0 6= n 2 T : } 62 V (T n0 )
Remark 5 The formulation of truth for atomic propositions has this form in
order to guarantee persistence of the relation `j=' along in a context where
these propositions may be updated. For instance, if

Z (T n) = Formula John
::: :::
then it may still be that hPPT ; T ni j= F o(U), because

Formula U Formula John
::: ::: F ::: :::
The left AVM may decorate some T 0n0 such that T 0n0 F T n. So, if the node
T 0 n0 has been introduced such that it satises F o(U), then this formula persists
along even if we substitute some value, e.g. John.
The verum symbol > holds at all nodes, the falsum symbol ? holds at no
node. The atomic proposition } represents the pointer and holds at { points
to { exactly one node in every partial tree.
Notice that the pointer } assigns to every partial tree T the unique pointed
partial tree T p, where p is the unique n 2 T such that } 2 V (T n).
There are no surprises in the truth denition for the standard logical constants,
connectives, quantiers and modalities.
Denition 15 (Truth Denition for DUF ) Given a dppt PPT = hP P T; fij
i 2 I g; (Z; z ); ; F BT i, an T n 2 P P T , and # 2 I = f#0 ; #1; #; ; #; L; Dg , then
the relation `j=' between the tuple hPPT ; T ni and a formula 2 DUF is given
by the following clauses:
hPPT ; T ni j= i 2 AT OM & 2 Vp (T n),

hPPT ; T ni j= Id(l; l0) i l; l0 2 D and l = l0,
hPPT ; T ni j=s i hPPT ; T ni 6j= ,
hPPT ; T ni j= ^ i hPPT ; T ni j= and hPPT ; T ni j= ,
hPPT ; T ni j= _ i hPPT ; T ni j= or hPPT ; T ni j= ,
hPPT ; T ni j= 7! i hPPT ; T ni 6j= or hPPT ; T ni j= ,
hPPT ; T ni j= 9x i there is a l 2 D : hPPT ; T ni j= [l=x],
hPPT ; T ni j= (x) i for all l 2 D : hPPT ; T ni j= [l=x],
hPPT ; T ni j= h#i i there is a T 0 n0 : T n # T 0 n0 & hPPT ; T n0 i j= ,
hPPT ; T ni j= f#g i for all T 0 n0 , if T n # T 0 n0 then hPPT ; T n0 i j= ,
hPPT ; T ni j= h# 1 i i there is a T 0 n0 : T 0 n0 # T n & hPPT ; T 0 n0 i j= ,
hPPT ; T ni j= f# 1 g i for all T 0 n0 : if T 0 n0 # T n then hPPT ; T n0 i j= .
This completes the denition of the operators that are evaluated without refer-
ence to anything other than the current partial tree structure.
The set F BT of fully decorated complete binary tree structures, the repre-
sentations of logical forms, can be addressed with the following modalities:
hPPT ; T ni j= [LF] i for all T 0 n0 2 F BT such that T n T 0 n0 ,
hPPT ; T 0 n0 i j= .
hPPT ; T ni j= hLFi i there is a T n 2 F BT such that T n T 0 n0 &
0 0
hPPT ; T 0 n0 i j= .
Notice that the universal operators in f7!; (x); f#g; f# 1gg are generally non-
persistent. That is, if hPPT ; T ni j= 7! and T n T 0n0, then it is not
necessarily the case that hPPT ; T 0n0i j= 7! . We get the persistent variants
of these operators by using the following denition. We rst set
hPPT ; T ni j= 2 i for all T 0 n0 if T n T 0 n0 then hPPT ; T 0 n0 i j= ,
hPPT ; T ni j= 4 i there is a T 0 n0 such that T n T 0 n0 : hPPT ; T 0 n0 i j= .
Thus, spelling out the denitions, we have the well-known interpretation of the
strong connectives and operators:
hPPT ; T ni j= : i for all T 0 n0 2 P P T such that
T n T 0 n0 : hPPT ; T 0 n0 i 6j=
hPPT ; T ni j= ! i for all T n0 2 P P T such that
0
T n T 0 n0 , if hPPT ; T 0 n0 i j= then hPPT ; T 0 n0 i j= .
hPPT ; T ni j= [#] i for all T 0 n0 such that T n T 0 n0 and all
m 2 T 0 such that T 0 n0 # T 0 m it holds that hPPT ; T mi j= :
hPPT ; T ni j= 8x i for all T 0 n0 such that T n T 0 n0 ,
for all l 2 D : hPPT ; T 0n0i j= [l=x]:
The DUF formulae built from elements of AT OM using only non-universal
operators and connectives plus strong connectives and operators will constitute
the set P DUF of Persistent DUF formulae.14 The facts at T n are given by the
function FA dened by FA(T n) = f 2 P DUF j hPPT ; T ni j= g.
14 Note that the language used in chapters 2 and 3 falls completely within this persistent
fragment.
Remark 6 (Persistence and Non-persistence) The formulae constructed
from persistent atomic formulae using the non-universal and strong universal
operators and connectives will persist under this truth denition. The atomic
(feature logical) formulae themselves are persistent according to this truth def-
inition. Atomic formulae are allowed to change incrementally along F , but
this change is compensated by the `backward-looking' truth clause: if an atomic
DUF formula holds at a ppt T n, then all successors will have a predecessor
where this formula holds and this guarantees persistence. Thus, if T n T 0 n0,
then FA(T n) FA(T 0n0). Persistent features represent the construction pieces
put into place; they are contributions to the eventual logical form.
As we are developing a model of a process, a dynamic system, we will also
nd reasons to consider transient aspects of partial trees: actions are to be
taken with respect to the momentary state of a tree. These transient aspects
are represented by non-persistent formulae.
For instance, we will use classical negation when we formalize IF THEN ELSE
actions over partial trees. Classical negation is not persistent.
hPPT ; T ni j=s () hPPT ; T ni 6j= :
The proposition s records the absence of at the node T n. If this absence
were a proper feature of the construction, it would have to persist (i.e. : would
have to hold). The following principles hold for classical negation in interaction
with strong negation:
_ s
:(^ s)
: $ (> !s)
The last is again equivalent to : $ 2 s , the denition of strong negation.
Notice that these are valid, and thus persistent, schemas, even though s
generally is not. Of interest is also the proposition s:, or, equivalently, 4,
which states that not all developments of the current node exclude ; so it
expresses the situation in which the current node has a successor where holds,
a (generally) non-persistent state of aairs.
A second example of a non-persistent construction is hLFi for arbitrary
, with the interpretation: there is a tree in F BT accessible where holds.
However, by denition 10.4b the formula hLFi> is persistent.
When we discuss the goal-directed side of the model, we will introduce the
typical non-persistent features { namely requirements. By their very nature,
requirements are non-persistent in a successful process.
A special class of non-persistent formulae is that of control features, which
can be deposited at nodes of a partial tree to be removed `later'.
Denition 16 (Control Features) A control feature C is a propositional con-
stant. The set CF containing all control features is disjoint from AT OM , and
we introduce a valuation Vc 2 P (AT OM [ CF )P P T with the stipulation that
for all T n: Vc(T n) \ P DUF = Vp (T n) and for C 2 CF :
hPPT ; T ni j= C () C 2 Vc (T n):
Notice that this denition allows control features to disappear from valuations
along the relation . Moreover, because CF and AT OM are disjoint sets, the
denition of FA(T n) does not include control features.
If we allow non-persistent features, then we must be clear when we may
and when we may not use them. We may not use them to denote aspects of
the tree that are contributions of rules or lexical actions; it would go against
the principle of incrementality to use them for this purpose. On the other
hand, we may use them as conditions guarding actions. Actions may test for
features of the current structure, as there need not be any assumption that these
features are permanent. We will allow classical negation, but only conned to
test conditions for actions.
Remark 7 (Partiality and the Maximal Element) Because T n may

be a pointed partial tree, it may be that hPPT ; T ni j= h#i holds whereas
hPPT ; T ni 6j= h#0 i _ h#1 i. This is, by denition, impossible on complete bi-
nary trees. But because the partial tree T n can, by denition, be embedded in
a complete binary tree T 0n0 2 BT (by denition, for every T n 2 P T there is a
T 0 n0 2 BT : T n T 0 n0 ), we do have:
(*) if hPPT ; T ni j= h#i, then hPPT ; T ni j= 2 4 (h#0i _ h#1i).
To see this, consider that 2 4 is satised by a node T n, if every -extension of
that node has an -extension satisfying . This means that is consistent with
the set of all formulae satised by T n, its theory, and with the theory of every
extension of T n. Now, every partial tree can be embedded into a binary tree no
matter how we extend that partial tree (to another partial tree!). So, because
the modal formula h#i is persistent along , and because h#i ! (h#0i_h#1 i)
holds on binary trees, the conclusion of (*) follows.
Analogously, we have:
if hPPT ; T ni j= h#i, then hPPT ; T ni j= 2 4 ( _ h#ih#i).
In
PPT = hP P T; fij i 2 I g; (Z; z ); ; F BT i
the set F BT of maximal elements represents the set of all logical forms, the end
products of a parse through a grammatical string of words. The elements of
F BT are `nished' structures. This is the import of denition 10.4c. It is easy
to see that this guarantees
T n 2 F BT ) 8 : hPPT ; T ni j= 2 4 ! :
So the principles of nite binary trees that hold everywhere with the prex 24,
hold without this prex in the elements of F BT . That is, the elements of F BT
are classical structures, and
PPT j= [LF]( _ :):
By denition 10.4b, every partial object in P P T can reach an object in F BT .
That means the function LF 2 P (F BT )P P T , dened by
LF(T n) = fT 0 n0 2 F BT j T n T 0 n0 g;
has a non-empty value set for every T n 2 P P T . The function LF (Logical
Form) assigns to every partial logical form its possible completions.
We have also as a validity the schema :[LF] ! [LF]:, because the suc-
cessors of a node include the elements of the LF set assigned to that node, and
a node in this set has only itself as successor.
The set of formulae such that [LF] holds at T n will of course include the
persistent facts that already hold at T n, but generally there will be a non-empty
set REST (T n) = f 62 F A(T n) j hPPT ; T ni j= [LF]g. The set REST (T n)
contains all facts that separate T n from a complete object; because the facts in
REST (T n) hold at all completions, they must, in some sense, be ìmplied' by
the facts at T n (and some general principles). In the section on goal-directedness
we will be considering only models in which the sets REST (T n) for T n 2 P P T
are nitely based; that is, there is a nite set REST (T n) such that for all
2 REST (T n) we have j= (i.e. ! is universally valid). If the set
V
REST (T n) is nitely based, then T n essentially lacks only a nite number of

facts. We can then associate these facts with T n as requirements on that node.
Remark 8 (Internal and External Modalities) In a partial tree we can
dierentiate between internal and external tree relations. For instance, it may
be that T n 0 T n0, whereas T n 6# T n0. This is captured by the fact that, on
partial trees, h#0i does not entail h#i. However, even on partial trees it makes
sense to consider an immediate dominance relation, which then encompasses the
0 daughter, the 1 daughter and the # relation. In the same vein we can consider
an external variant of the h#i modality. The denitions are straightforward.
We shall introduce external versions of the underspecied tree relations, using
the symbols `#' and `#' with the denition:
hPPT ; T ni j=# i hPPT ; T ni j= h#0 i _ h#1 i _ h#i,
hPPT ; T ni j=# i hPPT ; T ni j= _ ## _ h# i.
So these modalities have the standard LOFT interpretations. We now have
h# i ! 24 # and h#i ! 24 # .
9.2.2 Goal-directedness
Now we introduce requirement functions, in order to deal with the goal-directedness
of the parsing process, the development of pointed partial tree models towards
fully decorated binary trees.
9.2.2.1 FTR-structures
Denition 17 (Requirement Functions) Given a dppt PPT , a requirement
function V? : P P T 7! P (P DUF ) over PPT is a function assigning nite sets
of persistent DUF formulae (i.e. formulae in P DUF ) to elements of P P T

satisfying
1. V? (T n) \ FA(T n) = ;,
2. T n # T 0 n0; & V?(T n) = ; ) V? (T 0n0) = ;,
3. T n T 0n0 ) V? (T n) (V? (T 0 n0) [ FA(T 0n0 )),
4. T n T 0n0 ) (V? (T n) = ; ) V? (T 0n0) = ;).
A requirement15 function assigns persistent formulae to be achieved to nodes
in the model. Note that V? (T n) may be an inconsistent set. By (1) facts
and requirements are mutually exclusive. By (2), if V? assigns requirements to
a daughter of a node, then it must assign requirements to that node itself. In
case T n T 0n0 holds, we have by denition Vp (T n) Vp (T 0n0)16 : facts persist.
This need not hold for requirements: by (3) these persist only until they are
satisfied. Moreover, by (4), if a tree node has no requirements, in which case
we will call it a static node, then this remains the case along -developments.
So this denition allows the number of requirements to increase along , but
only if there are requirements; once a node is without requirements, then it
stays that way. It is natural to consider the developments of a dppt where all
its requirements have been satised. At such a development, supposedly, it is
the complete sub-term that has been created, of which the current dppt is a
partial approximation. We will introduce some notations dened in terms of
the requirement functions.
Denition 18 (Success Functions) For given function V? , we let a success
function LFV? be a function P P T 7! P (P P T ) such that LFV? (T n) = fT 0n0 2
P P T j T n T 0 n0 : V? (T n) FA(T 0 n0 )g.
The elements of LFV? (T n) are all developments of T n in which the requirements
associated with T n are satised (have become facts). So, all requirements at
T n are fullled i T n 2 LFV? (T n) (note that this implies V? (T n) = ;). If
V? (T n) has only persistent formulae (i.e. 2 V? (T n) implies $ 2), then
LFV? (T n) is an -closed set, and T n T 0 n0 entails LFV? (T 0 n0 ) LFV? (T n).
Finally, we dene an internal version of the complete structures in F BT :
F BTV? = fT n 2 P P T j T n 2 LFV? (T n)g:
In this subset of P P T , all pointed decorated trees are collected which have no
outstanding requirements.
Denition 19 (Decorated Pointed Partial Tree Models)
A decorated pointed partial tree model (pptm) is a tuple
PPT M = hPPT ; V? i;
where
15 See remark 10 for requirements for non-persistent features.
16 Also FA(T n) FA(T 0 n0 ).
PPT = hP P T; fij i 2 I g; (Z; z ); ; F BT i is a dppt structure and
V? is a requirement function over PPT such that LFV? (T n) F BT for
every T n.
In a decorated pointed partial tree model, the fact that no requirements are
outstanding guarantees that the structure is complete (i.e. an element of F BT ).
We have F BTV? (T n) F BT . Notice that the presence of requirements does not
entail that the decorated pointed partial tree under consideration is incomplete
(in the sense of being outside F BT ).
At every tree node we now have the set of atomic facts, elements of AT OM
and a set of requirements, elements of DUF . The atomic facts represent a partial
declarative unit to which we now have to add a nite set of required formulae.
We can extend the AVM notation to accommodate requirements in the following
way.
2 3
Lai U
6
6 Laj aj 7
7
4 ::: ::: 5
1 : : : ; n
And the typical Merge operation t can be illustrated by the following typical
application:
2 3
2
Lai U
3 2
Lai ai
3 Lai ai
6 Laj aj 7 6 Lak ak 7
6
6 Laj aj 7
7
6
4 ::: :::
7
5 t 6
4 ::: :::
7
5 = 6
6 Lak ak 7
7
Lai (ai ); h#0 i Laj (aj )
4 ::: ::: 5
h#0 i
Here the Merge operation on the standard part comes down to applying tF .
What the operation t adds is the absorption of requirements by facts.
9.2.2.2 The Language DUF (?)
We will now introduce some operators into the language to exploit the increased
structure of our feature-decorated growing partial tree structures. The language
DUF (?) extends DUF with some constructions dealing with requirements:
DUF (?) = DUF [ fF ; G ; ? j 2 P DUF g [ f?;g
We interpret these new operators as follows,
Denition 20 (Truth Denition for DUF (?))
For 2 DUF we have the standard truth denition. If 2 P DUF , then
hPPT M; T ni j=? i 2 V? (T n),
hPPT M; T ni j=?; i V? (T n) = ;,
hPPT M; T ni j= F i there is a T 0 n0 2 LFV? (T n)
such that hPPT M; T 0 n0 i j= ,
hPPT M; T ni j= G i for all T n 2 LFV? (T n)
0 0
it holds that hPPT M; T 0n0i j= .
A proposition 2 DU (?) is valid on PPT M, notation PPT M j= , if

hPPT M; T ni j= for every T n 2 P P T . Proposition is universally valid,
notation j= , if PPT M j= for every PPT M.
The set of requirements of a node is a finite list. This is why we can check for
both presence and absence of a formula as a requirement.17 The modalities F
and G make reference to the developments of a node where all its requirements
are satised. Notice that if LFV? (T n) 6= ; is equivalent to hPPT M; T ni j= F >,
so the requirements at T n are consistent with the facts there. By denition 17
the following four schemas hold (respectively):
!s?
?; ! [#]?;
? ! ( _?)
?; ! 2?;
Because F BT V? (T n) F BT , we have the schema
G( _ :)
We know that
(_?) ! G
for, by persistence, facts persist in all developments; and, by denition, a re-
quirement can only disappear by becoming a fact. When we also demand the
converse
G ! (_?);
then every that is neither in F A(T n) nor in V?(T n) is witnessed by some
T 0 n0 2 LFV? (T n) for which we have
hPPT M; T 0 n0 i j= ::
Now we do not have the principle F > { not every partial logical form can be
completed { because there may be requirements involved that are inconsistent
with the facts or with other requirements.
Remark 9 (Non-persistence of Requirements) By adding requirements to
our logic, we have included non-persistent propositions. That is, if ? holds at
some node, then there is no guarantee that this will remain the case when we
move along . Requirements are not meant to persist, but are control elements
guiding the construction process. But if a requirement disappears, then it is
only to be exchanged by a (persistent) fact. That is,
_?
is persistent for every . And notice that the four requirement schemas above
are all persistent.
17 Notice that ? holds only if is explicitly listed as a requirement. We do not take
consequences of requirements also as requirements, because the set of requirements may be
inconsistent (among themselves or with the facts).
Along the relation we can reach all developments of a node, the successful
and unsuccessful ones. Of particular interest, however, are the developments
which can be successfully completed and the invariances pertaining to them.
Remark 10 (Requirements for Non-persistent Features) It is notewor-
thy that a control feature C (denition 16) is a non-persistent propositional
constant, and requirements for these features have not been dened. In par-
ticular, clause 3 of denition 17 does not work, because facts may disappear.
A requirement for a non-persistent feature should persist until that feature ap-
pears, whether it is a persistent or a non-persistent feature that is required. We
can, however, introduce a requirement function for both non-persistent and per-
sistent features V?C : P P T 7! P (DUF ) and replace clause 3 of that denition
by the following:
If T n T 0 n0 & 8T 00 n00 (T n T 00 n00 < T 0 n0 ) V?C (T n) V?C (T 00 n00 ))
then, V?C (T n) (V?C (T 0n0) [ f 2 DUF j hPPT M; T 0n0 i j= g):
(Here, of course, T n < T 0 n0 means T n T 0n0 and T 0n0 6 T n.) Thus a re-
quirement for a non-persistent feature C cannot just disappear. This can only
happen when C becomes a fact. As a fact, it need no longer persist. Notice
that for persistent features (elements of P DUF ) the function V?C comes down
to V? . For 2 DUF we can then set
hPPT M; T ni j=? i 2 V?C (T n):
9.2.2.3 The Uses of Requirements
Consider all terms of the typed logical forms, the representations of the formal
meanings of the natural language strings under consideration. The elements of
this set are the structures which the natural language strings have to construct.
We represent these terms as trees by their applicative structure: that is, a sub-
term (x; ) is represented as a binary branching point with x annotating
the function daughter, the argument daughter, and AP L(x; ) the point
itself. At the nodes of these representations we hang empty sets. These empty
sets represent empty sets of requirements. This we now turn into the set F BTV?
of some pointed partial tree model PPT M, by adding all partial versions that
can be extended to elements of the F BTV? . Now, in order to guarantee that
the F BTV? F BT (i.e. that the set of representations of terms in our typed
lambda calculus includes the set of trees with no outstanding requirements),
every abstraction of such a term to a partial object has to be compensated by the
introduction of a requirement. Starting, for instance, from the term Read(John,
(a,x, Book(x))), we can create a partial term by abstraction over John. This
abstraction is not a term of our typed lambda calculus, so it should not belong to
F BT . By denition, this means that there must be some unfullled requirement
associated with it. But this does not have to be a requirement for exactly John.
It may (and, in fact, will) be merely a requirement for type e. So here is where
the invariances of the process come in.
Given a specic feature of the logical forms we are interested in, the rst
question is now always: can we devise a system of requirement introduction
such that the fullment of all requirements annotating a given tree corresponds
exactly to a completing of this tree to a term of our typed lambda calculus with
the desired feature? We will give three examples.
Binary trees As a rst example, we will consider a requirement strategy to the
eect that the set F BTV? coincides with the set of binary tree structures. That
is, no underspecied tree relations of the form or # are left in F BTV? that
are not the re exive transitive closure of the immediate dominance relation
or the union of argument and function daughter relation, respectively. The
idea is to introduce a tree node label, a monadic predicate T n with values in
DT n = fa x; x j a 2 A; x 2 f0; 1; Lg g. That is, a value of this predicate is
a nite sequence of elements of f0; 1; Lg possibly preceded by a constant from
a set A. When trees are constructed in the parsing process, in general it is
not known whether a description that starts o as a top node will remain so
(and thus be the root node of the eventual tree). This is why we introduce a
new node invariably with address a 2 A, satisfying a formula T n(a), where a
is a constant not yet occurring in the construction. In interaction with the tree
modalities various constellations are expressible. So, given the formula T n(a),
expressing the location of a node in the tree under construction, we can x:
T n(a0) $ h"0 iT n(a),
T n(a1) $ h"1 iT n(a),18
and we can x the root node of a tree as follows:
T n(0) $ ["]?:
(Signicantly, we do not internalize the underspecied modalities h#i and h#i
as values of the T n predicate.) The tree node formula T n(0) holds at a node if
it is a top node and remains so throughout all developments. (Note the use of
the falsum { Àt every node above the current one ? holds.' As ? is satised
by no node at all (denition 15), this means that there are no nodes above the
current one.)
Now, when we introduce a tree node with an underspecied relation to the
source node, we add the requirement for 9xT n(x) to the node with underspec-
ied address:
[?T y(t)] ) [?T y(t)]
[?T y(e); ?9xT n(x)]

18 We will also need T n(aL) $ hL 1 iT n(a) when we consider LINKed trees.
The point is that this requirement can only be satised when the node it dec-
orates is merged with one that has a fully specied relation to the top node of
the tree: only in that case will there be a l 2 DT n such that T n(l) holds at
the node. So if we introduce tree nodes with underspecied tree only if they
are accompanied by requirements for values on the T n predicate, then in all
elements of F BTV? that we can reach, the underspecication with respect to
tree relations will have been resolved.
Pronouns An analogous mechanism allows us to introduce meta-variables or
place-holder values with the requirement that they must be substituted in
order for the term to be complete.
[?T y(e) ] ) [ F o(U); T y(e); ?9xF o(x) ]
An object of type e has been supplied, but a new, weaker requirement has
taken its place. Only a development that supplies a concrete value for the
meta-variable U in F o(U) can end up in F BTV? .
Resumptive pronouns On the other hand, we may introduce the tree node
[?F o(Mary) ]
which contains a requirement that can be fullled only by a concrete value for the
F o predicate. Such a value, a copy of a value already introduced, cannot, in our
model, be derived from the sentence string except by means of a pronoun (as
a copy instruction). In parsing the string Mary, John likes her, the pronominal
it can induce the copying of any accessible value: this is the contribution of the
pronoun. But only if it copies the value from the expression Mary in such a
string will the requirement for the exact formula value that has been introduced
be fullled: this is the contribution of the goal-directed framework. A pronoun
occurring in such a constellation is a resumptive one in our model.
9.2.3 The Structure of Goal-directed Partial Tree Models
In this section we will deal brie y with some algebraic features of PPT Ms. We
will discuss some of the entities inhabiting them.
Remark 11 (Some Notational Conventions) In dealing with a pptm T n
we will use the following notations. In a bracketed notation we will use
T [n ]
to represent T n if Vp (T n) = and V? (T n) = , and we want to focus on the
facts and requirements of node n. We will use the notation
T [n ]=n
to represent T n with Vp(T n) replaced by and V? (T n) replaced by . So
T n = T [n Vp (T n) V? (T n)]=n:
The notation
(T [ f[n1 : : : : : : ] : : : [n : : : : : : ]g)n
k
will denote the pptm T with nodes [n1 : : : : : : ] : : : [n : : : : : : ] added and the
pointer now at node n.
k
For T n 2 P P T M , if we use the notation T nj , then this will denote a node

T m such that T n is a pptm, j 2 f0; 1; Lg and T n j T m. The same will hold
if we use the notation [nj ].
For N T , the notation
T =N
will stand for the tree with domain fn 2 T j n 62 N g and the tree relations of T
restricted to the new domain.
Finally, whenever we list a tree by giving its domain and its tree relations
explicitly,
(fn1; : : : nk g; ni # nj )nh;
then the displayed elements are all elements, and the displayed relations are
all relations. So, in the P P T above, n1 : : : nk are all tree nodes, and ni # nj
is the only tree relation.
Denition 21 (Morphisms and Equivalences)
We call pptm T and T 0 V? -morph (Vp -morph), notation T
=V? T 0 (T =V
T ), if there is a T -isomorphism f : T 7! T , such that for all m 2 T :
0 0 p
V? (T m) = V? (T 0 f (m)), (Vp (T m) = Vp (T 0 f (m))).

We set T =RV T 0 if both T =V? T 0 and T =V T 0 and T = T 0 i the
tree structures are isomorphically related.
p p
We call pptm T and T 0 equivalent, notation T T 0 , if both T T 0 and

T0T.
We call T ; T 0 LF-equivalent, notation T =LF T 0 , if T = T 0 and LF(T ) =
LF(T ).0 19
These relations give rise to equivalence classes of the form [T n]X = fT 0n0 2
PPTM j T n =X T 0 n0g.00 We lift to these classes by the stipulation [T n]X
[T n ]X i for every T n00 2 [T n]X there is a T 000n000 2 [T 0n0 ]X such that
0 0
T 00 n00 T 000 n000 .
0
Notice that it may be that T T but not T = T . For instance,
0
N1 = [n ; T y(e)] [n ; T y(e)] = N2
[n ; T y(e)]
These pptm are not T -morphic, but either can develop into the other, and mov-
ing between them does not change the set of possible (successful) developments
(N1 =LF N2).
19 For the denition of `LF', see remark 7.
9.2.3.1 Minima and Atoms
A -minimal element T n of PPT M is a one-element domain with empty tree
relations and empty Vp (T n) and V? (T n).
f[n ; ;]gn:
The domain cannot be empty, for, by denition, n 2 T . The minimal element
is the = equivalence class of these elements (there is only one).
The -atoms are the one-element domains T n with empty Vp (T n) and
V? (T n) = fAg for some A 2 DUF (?).
f[n ; A]gn;
where the tree domain T = f[n; A]g. The atoms are single tree nodes with an
atomic requirement. An atom of a special kind will be the Axiom:
Axiom f[n; T y(t)]gn;
where the requirement is for an object of type t, i.e. a proposition. This Axiom
will be the starting point of every parse sequence.
Note that
(T [n ; A]) (T [n A ;]=n);
and not the other way round: requirements lie below facts. The -atoms will
be the starting point of the parsing process; every node of the eventual tree has
come into being as an atom of the form [nT y(t)], [n T y(e)], etc. The typical
atoms are nodes with a requirement for some type.
Note also, for arbitrary A 2 DUF (?),
T [n ?A] T [n A]=n:
All question marks heading a requirement can be got rid of.
9.2.3.2 Joins
The lowest upper bound of T n and T 0n0, denoted by T n t T 0n0 if it exists,
constitutes a -lowest00pptm T 00 n00 above both pptms, such that there are T -
morphisms f : T 7! T ; f : T 0 7! T 00 satisfying f (n) = f 0(n0 ) = n00 and for all
0
m 2 T; m0 2 T 0 such that f (m) = f 0 (m0 ):
Vp (T 00 f (m)) = Vp (T m) tF Vp (T 0 m0 ),
V? (T 00 f (m)) = (V? (T m) [ V? (T 0 m0 ))=FA(T 00 f (m)).
For example:
T [n A A ^ B ] t T [n B A ^ B ]=n = T [n A; B ;]=n:
Notice that N1 = [n A ^ B] and N2 = [nA; B] behave dierently, have dierent
-successors, (N2 [n A B ]; N1 6 [n A B ]), although we have LFV? (N1 ) =
LFV? (N2 ). For facts annotating tree nodes, the join operation may involve
update of feature values. For requirements, the join operation takes the
requirements of both except for those that are satised by facts.
Because of the demand for the T -morphisms f : T 7! T 00 ; f 0 : T 0 7! T 00 mapping
T n and T n0 to T 00 n00 that they satisfy f (n) = f 0(n0 ) = n00 , if the join exists,
then the images of the two models must share a node.20 This gives us essentially
three kinds of information merge.
1. We can extend a pptm downwards, attaching new nodes to the current
frontier.
This will be used to incorporate subcategorized information (transitive
verbs create a node for a two-place relation and for an object node below
a node requiring a predicate node), and also to add a determiner-requiring
node and a nominal-requiring node below a node requiring type e in case
the natural language string supplies a determiner as decoration for a node
where a formula of type e is required.
2. We may add nodes above the current top node, extending the pptm
upwards.
This will be used in case a clause has been built which turns out to be
a relative or complement clause (as in Japanese). And it will be used in
genitive constructions where a noun phrase sequence turns out to be a
sub-sequence of a determiner (as in English).
3. We may merge nodes occurring in one and the same pptm.
This will be used for incorporation of information from the lexicon, inessen-
tially, and for the resolution of long-distance dependencies essentially.
The incorporation of information from the lexicon can be handled by using
the join operation. If we want to incorporate atomic as a fact and as a
requirement at node T n, then
T n t f[n ]gn
does the trick. Notice that if the requirement already holds at T n, then the
result is the same as
T n t f[n ;]gn:
For instance, the accusative case on some noun phrase contributes a requirement
?h"iT y(e ! t) to the type e node. If the node in question is indeed a daughter
of a node of type e ! t (as will happen in parsing him in John hits him) then
the requirement evaporates. On the other hand, if this node does not occur
20 Both the denitions for join and the meet given here can be extended to partial trees by
the stipulation: if # 2 fu; tg, then T #T 0 () T p#T 0 p0 , where p; p0 are the nodes in T
and T 0 respectively, such that T p j= } and T 0 p0 j= }.
in such a daughter position (as in parsing Him in Him, John hit), then it will
remain as a requirement until the node has found its position.
Moreover, if F F F0 , i.e. F e-structure F0 is the result of an update of
F e-structure F, then
T [n F ] t f[n F0 ;]gn = T 0 [n F0 ];
where T 0 is the pptm T with every occurrence of F replaced by F0. By denition
10.4a, if more than one of its nodes is decorated with F, then T has no -
successors where only the occurrence of F at T n has been replaced. This kind
of update will take place when substituting, for instance, John for the variable
U in F o(U): Note that
T n t T m;
if it exists, identies the nodes n and m in one and the same tree T . This
cannot be if both n and m are related by fully specied relations (i.e. by
i ; i 2 f0; 1; Lg) to the top node: the position of at least one of them must be
underspecied. In the case of domination this can (and will) be used to unify
nodes. The pptm
T [n ] t T [n 0 0 ];
if it exists, xes an underspecied node in T . Notice that the displayed kind
of unication is typical for T in normal form. If T n and T 0 n0 are in normal
form (remark 3), then their lowest upper bound, if it exists, need not itself be
in normal form. On the other hand, the set Nor(T n t T 0 n0) can represent this
lower upper bound. A normal lowest upper bound of T n and T 0 n0 { that is, an
element in Nor(T n t T n ) { need not be unique.
0 0
9.2.3.3 Meets
The highest lower bound of T n and T 0 n0 , denoted by T n uT 0 n0 , is the -highest
pptm T n such that there are T -morphisms f : T 00 7! T and f 0 : T 00 7! T 0
00 00
such that f (n00) = n; f 0(n00) = n0 and:
Vp (T 00 n00 ) = Vp (T n) uF V (T 0 n0 )
V? (T 00 n00 ) = (V? (T n) \ V? (T 0 n0 ))
[ (V? (T n) \ FA(T 0 n0 )) [ (V? (T 0 n0 ) \ FA(T n))
So the meet of two tree nodes is annotated by the F e-meet of the F e-structures
of the constituent nodes, and has the requirements that appear on both, plus
the requirements of one node that are satised by the other. For instance,
T [n ; ; ; ] u T [n ; ; ]=n = T [n ; ; ]:
Highest lower bounds can be used to deal with varieties of ellipsis and with
structural parallelism in general. In the text John hates Bill; and so does Sue
it is the lowest upper bound of the representations of John hates Bill and Sue
hates Bill which can be used as the background against which the string so does
Sue is interpreted.
We can dene the notion of a similarity measure to capture structural par-
allelism.
A similarity measure of structure between elements of F BT can be dened by
simT n (T 0 n0 ; T 00 n00 ) meaning that pptm T 0 n0 is more similar to T n than T 00 n00 .
The denition is as follows:
simT n (T 0 n0 ; T 00 n00 ) () T n u T 00 n00 T n u T 0 n0 :
As should be, no element is more similar to T n than T n itself. According to
this measure, for instance, the structure projected by Sue hates Bill is more
similar to that projected by John hates Sue than the structure projected by Sue
likes Bill, and this again is more similar to the rst tree than the one induced
by John walks. A structural similarity class with respect to T n
[T 0 n0]T n = fT 00n00 2 F BT j simT n (T 0n0 ; T 00n00 ) & simT n(T 00n00 ; T 0n0)g
gives a class of decorated binary trees which are structurally similar with respect
to a given target structure. So Jane hates Bill is in the same similarity class
with respect to John hates Bill as Sue hates Bill is.
Another use of the meet operation is the following. Consider the development
sequence
A A A
0 1 T (0,*) 1 T (0,*,D) 1
B C B C B C
Here, it is clear that the addition of information is a matter of mere bookkeep-
ing. It represents a taking into account, making explicit, of the information
present in the previous structure. The structure in which all implicit informa-
tion has been made explicit, can be dened using the u operation. Let the BK
(BookKeeping) function be given by
BK (T n) = ufT 0 n0 2 F BT j T n T 0 n0 g;
where Vp (uD) = uF fVp (T n) j T n 2 Dg, and V?(uD) = f 2 DUF (?) j 9T n 2
D : 2 V? (T n) & 8T n 2 D; 2 V? (T n) [ FA(T n)g. The function BK maps
a structure to the common meet of all developments in which all requirements
have been satised.
9.2.4 Tree Descriptions
Our (theoretical) parser will not work on trees themselves as representations of
partial terms of some logical language, but on descriptions of partial trees.
These descriptions consist of nite sets of DUF (?) formulae. Tree descriptions
will be the computational approximations of partial trees. The approximations
of tree nodes are nite, splitting, P R-consistent sets of DUF (?) formulae. Here
P R is a set of principles incorporating the Logic of Finite Trees (LOFT) as
well as axiomatizations of the individual predicates or features. For instance,
by standard logical consistency we have no description containing both and
:; by LOFT principles no description contains h#0 i and h#0 i: (`there is
only one argument daughter'); and by axiomatization of the label predicates,
no consistent description of a node contains both T y(e) and T y(t): types are
supposed to be mutually exclusive. We will use the notation `P R if formula
can be computed from formula set using the principles in P R and the usual
derivation rules.
Tree nodes represent partial, dynamic, declarative units: a declarative unit
is a labelled formula, which we represent by a nite set of atomic DUF formulae
fLa1(a1 ) : : : Lan (an ); F o()g satisfying some notion of completeness. A partial
declarative unit is simply a subset of some complete one, and a dynamic partial
declarative unit is a partial one with requirements added. That is, it consists of
a nite number of atomic formulae from AT OM formulae and a nite number
of requirements ? 2 DUF (?) (a valuation Vp and a requirement function V?).
The object is to get rid of the requirements from this set by satisfying them.
Descriptions of these partial dynamic declarative units are generalized sets
of formulae which may include conclusions of this basis; that is, ^ 2 d implies
; 2 d, and _ 2 d implies 2 d or 2 d, and ; ! 2 d implies 2 d.
The tree relations between tree nodes will hold between node descriptions
in virtue of tree node formulae which are generated by a monadic tree node
predicate T n (having values in DT n = fa n j a 2 A [ feg; n 2 f0; 1; Lgg for
some non-empty set A), and by closure under Boolean and modal operations.
Denition 22 (Node Identier Language) For index set I = f#0; #1; #;
#; "0; "1; " g, the language ADR is the smallest subset of DUF (?) satisfying
1. If a 2 DT n [ MV then T n(a) 2 ADR,
2. if 2 ADR, then hii in ADR for i 2 I or i = L.
By means of the node identier language we can now formulate the notions of
node descriptions and tree descriptions.
Denition 23 (Node and Tree Descriptions) A node description d is a -
nite P R-consistent set of DUF (?) formulae satisfying:
1. 9a 2 DT n : T n(a) 2 d,
2. s 2 d ) 62 d,
3. ^ 2 d ) ; 2 d,
4. _ 2 d ) 2 d or 2 d,
5. ; ! 2 d ) 2 d,
6. 9x 2 d ) [l=x] 2 d for some l 2 D,

7. 2 2 d ) 2 d,
8. ?; 2 d ) :9 :? 2 d.
Let D be a nite family of node descriptions, and let for i 2 f0; 1; L; #; ; D; U g,
the relation Di D D be dened by
d Di d0 () 9hiit 2 d \ ADR : t 2 d0 or 9t 2 d \ ADR : hi 1 it 2 d0 :
Then D is a tree description if
9. T n(a) 2 d; T n(b) 2 d0 ) (a = b () d = d0 ),
10. 9!d 2 D : } 2 d,
11. 2 d; Dd Di D0 d0 ; hii ! 2 d0 ) 2 d0,
12. hii 2 d ) 9d0 2 D : Dd Di Dd0 & 2 d0;
13. [i] 2 d & d Di d0 ) 2 d0 ,
14. V? (d) = ; & Dd Di D0 d0 ) V? (d0 ) = ;.
So every node of a tree description contains an ADR formula. Whenever it is
clear which tree description D we are dealing with, we will drop it as a super-
script of the tree relations i . A node description is explicit about conjunc-
tions, disjunctions and implications. However, if we want to constrain a node
description with, for instance, a non-splitting disjunction, then we have to use
a requirement, e.g. ?( _ ), and consider the developments of the description
where the requirements have been fullled. We can dene T -morphisms on de-
scriptions preserving the tree structure as we have done for T -structures. Thus
we can carry over the various notions of equivalence, as dened in denition 21,
from T -structures to tree descriptions.
Denition 24 (Developments of Node Descriptions) For nite sets d; d0
of DUF (?) formulae we dene d0 is a development of d, notation d d0, by
d d0 () f? 2 DUF (?) j? 2 dg [ f2; G 2 DUF (?) j 2; G 2 dg
(d0 [ f? j 2 (d0 \ DUF g):
We set D D0 i for all d 2 D there is some d0 2 D0 such that d d0 .
Denition 25 (Pointed Tree Descriptions) The space of pointed nite tree
descriptions SP is
SP = hD; ; fij i 2 I g; Li;
where, for i 2 I; j 2 I [ fLg:
1. D = fDd j D is a tree description; d 2 Dg,
2. Dd D0 d0 () d d0 & D D0 ,
3. Dd i D0 d0 () D = D0 & d i d0,
4. Dd L D0 d0 () D \ D0 = ; & 9t 2 ADR : hLit 2 d; t 2 d0 and d0 has
no i predecessor in D0, i 2 f0; 1; Lg.21
5. hLi 2 d ) 9D0 d0 2 D : Dd L D0d0 & 2 d0,
6. [L] 2 d ) 8D0d0 (Dd D0 d0 & D0 d0 L D00d00 ) 2 d00 ).
Notice that tree relations are preserved along . This is because the tree rela-
tions between nodes are determined by the ADR formulae they contain and
implies inclusion (of tree node formulae).
Remark 12 (Consistency Assumption) We proceed now on the assump-
tion
If Dd 2 D and a node description d0 is a P R-consistent extension of d, then
there is a tree description D0 such that d0 2 D0, D0 d0 2 D and Dd D0d0 .
At least for the LOFT principles the consistency lemma guarantees this assump-
tion. Given this closure property, LOFT can then determine tree structure. For
instance, h0i; h0i ! h0i( ^ ), makes the basic relations functional. For,
given a d in D containing the premises, then this node can always be extended
to one containing the conclusion. This extension is a node in an extended tree
description, and there the denition of a tree description guarantees that some
node contains ^ along the 0 relation.
Example 1 We will show in one particular case how being a tree description
is involved in the determination of tree structure. Suppose we have the sets
A = fT n(0) : : : g
B = fh"0iT n(0)X g
C = fh"0 iT n(0); :X g
We have A 0 B and A 0 C according to our denition of 0, but the three
sets cannot be nodes in some binary tree, for such a tree has only one 0 daughter,
and B and C are incompatible node descriptions. So the set T D = fA; B; C g
should not satisfy denition 23. Here is the reasoning why this is so. By the
denition of satisfaction of descriptions, we can conclude
T D; A `P R h#0 iX & T D; A `P R h#0 i(:X ):
By propositional logic, this gives T D; A `P R h#0iX ^ h#0i(:X ). Now LOFT
has the functionality axioms for # 2 f#0; #1; "0; "1g,
h#i ^ h#i ! h#i( ^ ):
21 The empty intersection of D and D0 is guaranteed by the fact that each description is
set up with a tree node (description) containing T n(a) for some fresh constant a. Eventually
T n(i) may be added to (the description of) this node for i that is some (denitely not unique)
sequence in f0; 1; Lg , but this does not alter the presence of the unique T n(a).
These axioms guarantee that a node has (essentially) only one argument daugh-
ter and function daughter. So we have T D; A `P R h#0i(X ^ :X ), from which
LOFT allows us to conclude T D; A `P R ? in contradiction with the consistency
constraint. So T D is not a tree description.
Notice that the tree relations between two node descriptions are completely
determined by the presence or absence of specic formulae in these descrip-
tions themselves: this tree relation is determined independent of larger tree
descriptions of which these two node descriptions are elements. It may turn
out, however, that the tree description of which node description ND is an ele-
ment determines the position of ND to be more specic than is witnessed inside
ND itself. Consider the following tree description
A = fT n(0) : : : g A
B = fh"iT n(0); X g # 0
C = fh"0 iT n(0); :X g B C
We have A # B, and A 0 C according to denition 23. But there is more
information contained in this tree than is witnessed inside the node descriptions.
By the LOFT principle for # 2 f#; "g,
h#i $ (h#0 i _ h#1 i);
we have that the daughter (mother) relation is the union of the argument
daughter (mother) and the function daughter (mother) relation. We know that
A 0 B or A 1 B ; and, by the functionality axioms, we cannot identify B and
C , thus we must have A 1 B .
T D; fh"iT n(0); X g `P R h"1 iT n(0):
This relationship between node description and tree description will be an im-
portant economy principle, since there will characteristically be information
derivable from the tree description which is not represented in the node descrip-
tions. But this is not all. Consider, nally, the tree description T D consisting
of the nodes A; B; C and D:
A = fT n(0); : : : ; T y(t)g A
B = fh" iT n(0); T y(e); X g * 1
C = fh"iT n(0); [#]?; T y(e); :X g 0
D = fh"1 iT n(0); T y(e ! t)g B C D
In this description, node B has only an underspecied location. We have
A B , A 0 C , and A 1 D, according to denition 23. Again, there is
more information contained in this tree than is witnessed inside the node de-
scriptions. In fact, it must be the case that A 1 D B. For, identifying the
underspecied node B with A or C gives inconsistent sets, and it must dangle
below either C or D. But C can have no daughters, so only D is left to dominate
B . The LOFT principle
h# i $ ( _ h#ih# i)
Procedural Structure 307
regulates this propagation of what we might call the long-distance modalities.
This now allows us to derive
T D; fh"iT n(0); T y(e); X g `P R h" ih"1 iT n(0):
Adding this conclusion to the node description can be seen as a restructuring
of a tree with unxed tree relations.
A
C D
*
B
9.3 Procedural Structure

In the previous sections we have dened the goal-directed space PPT M =
hP P T; ; V?; F BT i representing the space of partial logical forms, all connected
to full logical forms (in F BT ) by a relation of development or growth (). This
is the declarative side of the natural language parser we are formalizing. This
space is described by the language DUF (?).
Now we will introduce the second half, the procedural side of the picture,
the actions which drive the Axiom towards a logical form.
The course from Axiom to F BT cannot follow just any path available; it
must be a path licensed by the natural language under consideration within
the context of some language-external mechanisms. A natural language deter-
mines, 0through its actions, a set of transitions L P P T M P P T M , such that
hT n; T n0 i 2 L implies T T 0 ; that is, the transitions respect incrementality.
These transitions then determine straightforwardly an ordering L such
that for all T n 2 F BT : Axiom T n () Axiom L T n. That is, the paths
a language licenses lead to all logical forms.
The way a language licenses a path through PPT M is by contributing
actions which map a pptm to an -related pptm. A language contributes three
kinds of actions. Firstly, lexical actions, projected by the words of the language;
secondly, computational rules, determined by general principles of the language,
and thirdly language-external actions, like Substitution, the resolution of scope
ambiguity, or inferential activity on the basis of semantic content { the F o
values { of the (partially created) tree. That is, a transition T n L T 0n0 must
be licensed either by a word of the natural language under consideration, by
general principles of that language, or by general external principles. The next
section will introduce the actions from which these language contributions are
constructed.
9.3.1 Actions over Goal-directed Partial Tree Models

We will add to the language DUF (?) a set ACT of action constants and to our
models a function Va : (P P T M P P T M ) 7! P (ACT ) which assigns the new
constants to actions. Actions are partial, multi-valued functions which map the
trees in their domain to enriched ones. The actions are adapted to our model
by being incremental.
An action constant determines a binary relation on PPT M through
< = fhT n; T 0 n0 i j 2 Va (hT n; T 0 n0 i)g:
By abuse of notation, we will set (T n; T 0n0) to denote 2 Va (hT n; T 0 n0i).
Denition 26 (Incremental Actions) An action such that (T n; T 0 n0 )
implies T n T 0 n0 will be called a construction, and if (T n; T 0n0 ) implies
T n i T 0 n0 , then will be called a (pointer) movement. A construction such
that (T n; T 0n0) implies T n = T 0n0 and T n F T 0n0 will be called an update.
An action is incremental if (T n; T 0n0) implies T T 0. So, an incremental
action is composed of constructions and movements.
We can now extend the truth denition to incorporate action modalities:
Denition 27 (Truth Denition for Action Modalities) For ACT , a set
of incremental actions with 2 ACT , Va : (P P T M P P T M ) 7! P (ACT )
a function assigning subsets of ACT to pairs in P P T M P P T M , and 2
DUF (?), the action modalities [] and hi are interpreted respectively by:
hPPT M; T ni j= [] i for all T 0 n0 2 P P T M
such that (T n; T 0n0) we have hPPT M; T 0n0i j=
hPPT M; T ni j= hi i there is some T 0 n0 2 P P T M
such that (T n; T 0n0 ) and hPPT M; T 0n0i j=
Remark 13 (Non-persistence) The procedural side of the model deals with
actions which take decorated partial trees as transient phenomena that may
acquire and lose features in the course of development. We may want to execute
an action on a partial tree structure at a certain point in its development0 0but
not at some later stage. For instance, we may want to attach a node T n to
a given node T n with the relation of domination, i.e. T n T 0n0, but only
if T n does not (already) have daughters. Being without daughters is a feature
that may disappear under development, so if an action has to be sensitive to
this state of aairs, then it must recognize non-persistent conditions { that is,
control features, as we have called them in denition 16.
The fact that actions may be triggered by a structure T n without being
triggered by all of its developments means that the action modalities dened
are generally not persistent. For instance, hPPT M; T ni j= hi does not in
general imply that hPPT M; T 0n0i j= hi for T n T 0n0.
The incremental actions we will use are built from basic actions in a set ACT
by the regular operations plus an IF THEN ELSE construction to give a set
A(ACT ) of actions.
Apart from the abort and stay put actions, the set ACT allows creation of
a new tree node connected by a tree relation to a given a node, moving from
a given node to another node along a tree relation, adding atomic annotations
or (arbitrary) requirements to a given node and, nally, merging of underspec-
ied nodes with more specied ones. Notice these four kinds of actions are
incremental.
Denition 28 (Actions over PPT M) Let ACT , the set of basic actions con-
sist of
ACT = fABORT; 1; make(#); put(); go(#); merg(#)g;
where
ABORT: the ABORT action is dened by,
ABORT(T n; T 0n0) ) hPPT M; T ni j= ?
(i.e. there are no pairs standing in the ABORT relation).
1: This rst action is the identity action, `stay put', is given by
1(T n; T 0 n0 ) ) T n = T 0 n0 :
make(#): For # 2 fh#0 i; h#1 i; h#i; h"i; h# i; h" i; hLi; #; "; # ; " g the action
make(#) adds to T n one new node related to n by # where, if
# =#, then #=# [ 0 [ 1, and analogously for the other exter-
nal modalities (see remark 8).
make(#)(T n; T 0 n0 ) ) T 0 n0 = T n t (f[n ; ;]; [n0 ; ;]g; n # n0 )n:
(For this notation, see remark 11.) Notice that this action creates the
node if it is not already there. If it is, it may still add a pair to some
tree relation: it may create a new connection between already existing
nodes. For, instance, it may be that T n 0 T n0, but not T n T n0.
So make(h#i)(T n), does not introduce new nodes, but adds the pair
hn; n0 i to the relation .
put(): The action put() adds the proposition to the node. Let
AT OM [f? j 2 DUF (?)g be a nite set, and let F = AT OM \
and R = f j? 2 g (note = R [ F ). The action put() is
dened by:
put()(T n; T 0 n0 ) ) T 0 n0 = T n t T [n F R]=n:
Note, put(?)([n ]) = [n ] and put(?)([n ]) = [n ; ].
So if 2 Vp (T n) [ V? (T n), then put(?)(T n) = T n = put()(T n).
Moreover, if F , i.e. is the result of an update of , then

put( )([n ]) = [n ].
The restriction of the put action to atomic propositions and require-
ments is essential for it to count as a construction. A formula of,
for instance, the form h#0i cannot be added to a tree node, for it
requires a tree context to be true, and this context need not be there,
whereas neither atomic propositions nor requirements require such a
context.
go(#): For # 2 fh#0 i; h#1 i; h#i; h"i; h# i; h" i; hLi; #; "; #; " g the action go(#)
eects a pointer movement in the current tree to the # related node.
go(#)(T n; T 0 n0 ) ) T n # T 0 n0 ;
where, again, for # an external modality, the relation # consists of

the union of the appropriate #0 .
merg(#): The action merg(#) unies a node T n with a node T n0 dangling below
it, i.e. T n T n0 or T n D T n0, under the conditions that there
is such a node and the unication exists. Notice that, by the normal
form of partial trees (remark 3) this is the only kind of unication of
nodes we have to consider.
merg(#)(T n; T 0 n0 ) ) 9T n00 : T n # T n00 : (T n t T n00 ) = T 0 n0 :
By denition, all elements in the domain of merg(#) satisfy the con-

ditions above.
These basic actions can now be combined by the regular operations plus an IF
THEN ELSE construction. For ACT , a set of basic actions as described, the
set A(ACT ) is the smallest set in P P T M P P T M satisfying
1. ACT A(ACT ), and
2. for ; 0 2 A(ACT ); (x) DUF (?), where x is a sequence of all variables
occurring free in , we have ; 0 ; + 0 ; ; h(x); ; 0 i 2 A(ACT ),
where ; ; +; have their usual interpretation and
h(x); ; 0 i =
fhT n; T 0 n0 i 2 [t=x] j t 2 (D [ MV ) ; hPPT M; T ni j= [t=x]g [
fhT n; T 0 n0 i 2 0 j :9t 2 (D [ MV ) ; hPPT M; T ni j= [t=x]g:
The composite actions are standard. An0 action of the form h(x); ; 0 i is
essentially an IF (x) THEN ELSE statement. If contains variables,
then these are bound to their instantiations.
Example 2 We will now present a selection of action macros that are frequently
used in the chapters.
Addition of information A very frequent use of actions is in creation of a
node and putting some information there, either a fact or a requirement. The
standard sequence of actions is
make(#); go(#); put()
where we rst make a node at location # removed from the pointer, then we
go there and put there the formula 2 DUF (?) [ CF . Often we will also return
to our point of departure,
make(#); go(#); put(); go(# 1 ):
Conditional actions Conditional actions are IF THEN ELSE actions which

can bind variables in the process. For instance, the action
hfF o(x)g; go(h#0 i); put(F o(x)); ABORTi
maps the value of the F o feature at the current node, if there is such a value, to
the F o feature at the argument daughter; otherwise it aborts. When conditional
actions have no complex internal structure, we will also display them more
perspicuously as
IF F o(x)
THEN go(h#0 i); put(F o(x))
ELSE ABORT
Pointer movements We can also use conditional actions in movements of the
pointer. Consider
gofirst" (X ) =df hfX g; 1; go(")i ; hfX g; 1; ABORTi
and
gofirst# (X ) = hfX g; 1; go(#)i ; hfX g; 1; ABORTi:
The action gofirst" (gofirst#(X )) moves to the rst node above (a rst node
below) the starting point where X holds and stays there. So gofirst"(?T y(t))
moves to the top node of the local, minimal clause which the starting point is
contained in.
WHILE X DO = hfX g; ; 1i; hfX g; ABORT; 1i:
This conditional action executes action while X holds at the node of evalua-
tion. So WHILE DO go(") moves to the rst node above the starting point not
in the domain identied by .
If we want to undertake some action at some point not yet known and
return to the current position, we can use control features. Recall that these
features are not persistent, and this is what we want in this case. We only need
a marker to identify the current position; we do not want this marker to be
present in the eventual tree. The idea is to put a control feature C 2 CF at a
node as a requirement ?C (dened as in remark 10). Now we use, for instance,
gofirst# (X ) to move to a rst node downwards where X holds, execute some

action , to be followed by gofirst"(?C ), reaching the starting position (if
?C is `fresh'). Now putting C at that node removes ?C and now C itself may
disappear:
put(?C ); gofirst# (X ); DO ; gofirst" (?C ); put(C ):
Fresh names A further action we will describe will deal with the introduction
of a fresh logical name, either a constant or some variable, to be used in some
formula . The notion of freshness is here dened with respect to the linked
tree structure under development. The action freshput(a; ) generates a xed
nite number of instances of the term a and uses the rst one that is fresh in
. So, to actually contribute a fresh variable to some tree, this number has to
exceed the number of a occurrences in that tree.
freshput(a; )=
IF fhD 1 ihDiF o(a1 )g
THEN IF fhD 1 ihDiF o(a2 )g
THEN : : :
: : : IF fhD 1 ihDiF o(an )g
THEN AB
ELSE put(F o([an =a]))
:::
This macro is used in the actions projected by a proper name. For instance, the
name John will have the lexical entry
IF ?T y(e)
THEN put(T y(e)); freshput(John; F o(John))
ELSE ABORT
So distinct occurrences of the name John in one sentence will give rise to distinct
logical constants annotating tree nodes. Dierent tree nodes may end up with
instances of the same logical name, but only through a process of copying.
9.3.2 Natural Languages
A natural language will now be represented by an action set L A(ACT ) which
is closed under sequential composition, iteration, non-deterministic choice and
conditional choice. We set L= ( \f P P T M P P T M j 2 Lg). This
is again a partial order: re exivity follows because the stay put action, `1',
is present. Transitivity follows by closure under sequential composition, and
asymmetry is inherited from . The relation L represents the transitions
within the whole of PPT M that are licensed by the language L.
Every parsing task starts from the Axiom which is a one-node tree, without
tree relations, an empty valuation function, but with a requirement for an object
of type t:
f[n ; T y(t)]gn
So if a parsing task is fullled, then it must have constructed an object of type t,
a proposition, at this node. By parsing a grammatical natural language string,
the Axiom is connected to an element of F BT by a nite sequence of trees. Each
tree T ni+1 in this sequence is a development of the previous tree T ni , licensed
by the language; i.e. T ni L T ni+1 . So the construction of a full-blown binary
tree (2 F BT ) can be seen as tracing a path of the kind
Axiom = T n0 L T n1 L : : : L T nk 2 F BT:
The set L of actions, is divided into three parts:
1. A set L of L-words. For for every w 2 L the set Lex(w) A(ACT ) is a
nite set of incremental actions.
So a word may have a nite number of (complex) actions associated with
it.
2. A nite set C A(ACT ) of computational actions. These actions we will
generally call rules.
3. A set P A(ACT ) of pragmatic actions.
An action out of each of these sets determines a licensed transition. The actions
in L are generally not dened on, not triggered by, every dppt T n. That is, an
action 2 L determines a subset of P P T M , its domain given by dom() =
fT n 2 P P T M j 9T 0 n0 : (T n; T 0 n0 )g. So every action in L determines a
trigger dom() P P T M . And we will stipulate that a language has only a
nite number of triggers. That is, we stipulate that the set
fdom() P P T j 2 L [ C [ P g
is nite. So the actions associated with L split up the set of decorated partial
trees into a nite number of categories. A language can only be sensitive to a
nite spread of dierences between the decorated partial trees. Note that there
is no limit on the variety in the products of these actions. We may have an
innity of actions, giving an innity of dierent structures, but these actions
share a nite number of domains (or triggers).
The procedural semantics of the parsing process of an L string over PPT M
is now given in terms of LCP congurations. We display an LCP conguration
as a pair
(T n; w)
consisting of a pptm T n and a string w of lexical items; i.e. w 2 L. The set of
all LCP congurations will be denoted by CONF (L; C; P ). In a parse, LCP
congurations are rewritten to new congurations by lexical actions, computa-
tional rules and pragmatic actions. The basic rewrite relation of our parsing
model is the binary relation,
(T n; w) )LCP (T 0n0; w0 );
where )LCP is dened to be )L [ )C [ )P , and the components of this
union are given by
(T n; w) )L (T 0n0; w0 );
if w = w1 w0 for w1 2 L, 2 Lex(w1), and (T n; T 0n0). That is, )L-transitions
consume the rst word of the string and execute one of the actions it projects.
A transition dened by a computational rule or a pragmatic action has the
form
(T n; w) )x (T 0n0 ; w0)
if x = C0 (0 x = P ), w = w0 , and there is some rule 2 C ( 2 P ) such that
(T n; T n ).
Here no word is consumed; the action undertaken is determined by the set
C of rules or the set P of pragmatic actions.
We may restrict the sets L; C and P to subsets L0; C 0 ; P 0 and consider the
rewrite relation )L0C0P 0 (a subset of )LCP ) with the obvious restricted inter-
pretation. We will drop the empty set from the notation; so, for instance, )C
will be short for );C;. As usual, the relation )LCP will consist of the re exive
and transitive closure of )LCP .
If language L is determined by L; C and P , then we can x our relation
L =LCP of licensed development or growth by
T n L T 0 n0 () T n T 0 n0 & 9w; w0 2 L : (T n; w) )LCP (T 0 n0 ; w0 ):
(Because pointer movements are included as possible actions, we have to de-
mand that T n T 0n0 holds in the above denition). Analogously, we can
dene relations CP , C , etc. Now we can restrict the general notion of de-
velopment of tree structure as formalized in the relation between partial tree
structures to those developments that are licensed by the lexical actions L and
the computational rules C (see section 9.3.2). That is, we can evaluate logical
constants along developments licensed by a particular natural language. For
instance, an implication
hPPT M; (T n; w)i j= !LCP
will now be evaluated by
for all (T 0 n0 ; w0 ) such that (T n; w) LCP (T 0 n0 ; w0 ), if hPPT M; (T 0 n0 ; w0 )i
j= then hPPT M; (T 0 n0 ; w0 )i j= .
Moreover, we can restrict this implication to subsets of LCP ; for instance,
CP considers only developments without lexical contributions { that is, if
(T n; w) CP (T 0n0 ; w0), then w = w0 . This is the relation along which the
implication !CP is evaluated; we consider invariances over developments that
are possible without referring to the next word. Of interest is the restricted
negation :C
hPPT M; (T n; w)i j= :C ;
which states that (T n; w) has no successors that can be reached by computation
alone and where holds. On the other hand, lexical actions may still make
such a successor available.
It is now straightforward to dene a concept of well-formedness over the
sequence of transitions which a string of words must dene if it is to successfully
lead to a logical form as one possible interpretation assigned to those words.
The sequence of words must lead from the Axiom to the nal state using each
action dened by the words in the order given, computational rules or pragmatic
actions, so that a well-formed logical form results with no requirements on any
tree-node left outstanding and no words left unprocessed.
(Axiom; w) = (T n0; w) )LCP (T n1; w0 ) )LCP : : : )LCP (T nn; e) 2 F BT:
Recall that e denotes the empty string.
Now, the set of grammatical strings, given C , L and P , can be dened as the
strings that, starting from the Axiom, will reach an element of F BT exactly at
the last word.
Denition 29 (Grammatical L-strings) For L a natural language, the set
ST RU (L) of all L-structures is given by
ST RU (L) = fT n 2 P P T j 9T 0 n0 2 F BT : Axiom LCP T n LCP T 0 n0 g:
Language L is expressively adequate (EA) if
F BT ST RU (L):
If natural language L is expressively adequate, then it can project any logical
form in our representation language. Given a natural language L, the set of
L-grammatical strings, L(L; C; P ) is given by either of the two stipulations
1. fw1 : : : wn 2 L j 9T n 2 F BT : (Axiom; w1 : : : wn) )LCP (T n; e)g,
2. fw1 : : : wn 2 L j 9i 2 Lex(wi ); 1; : : : ; n+1 2 C [ P j
j hPPT M; Axiomi j= h1 ; 1 ; 2 ; : : : ; n ; n ; n+1 i(["]?^?;)g:
The representations projected by a string s 2 L(L; C; P ) are given by
Repr(s) = fT n 2 F BT j (Axiom; s) )LCP (T n; e)g:
Depending on the level of underspecication of s { that is, the use of prag-
matic actions in P { the set Repr(s) may be a singleton (All men hate chocolate),
a nite set (Every dog hates a man), or an innite one (He hates him).
Computational rules
By denition, whenever we have (T n; w) )C (T 0n0; w), then we have 8w 2 L :
(T n; w) )C (T 0n0; w), because the string is not involved in computational rules.
So the notion T n )C T 0n0 makes sense. Computational rules will be based on
principles of the form
! hi>:
Such a formula claims the existence of an action transition, given the right pre-
condition. Each principle ! hi> is associated with a relation !hi> 2 C ,
given by:
!hi> = fhT n; T 0 n0 i j hPPT M; T ni j= & (T n; T 0 n0 )g:
The progress from Axiom to F BT is non-deterministic: at every state of the
parse the word currently under consideration can generally be assigned more
than one structural role in the tree by principles in C [ P . For instance, a noun
phrase sequence initiating a natural language string may end up as subject
John, admires, Mary; it may be a fronted object, Mary, John, admires; or it
may be a topic constituent, (As for) Mary, John, admires, her. The set C must
contain actions to set up these three structures. The transitions determined by
the computational rules, i.e.
(T n; w) )C (T 0n0; w)
are non-deterministic ()C is a multi-valued function). But, as is evident, for
the parsing process to get o the ground at all, for any given tree T n the set C
should contain only a nite number of (alternative) actions which have T n in
their domain: only a nite number of developments can be computed. Moreover,
we should not be able to apply rules indenitely: after a certain number of
applications, we must have reached a tree that is no longer in the domain of any
rule in C . Thus, for a given T n, not only is fT 0n0 j 9w : (T n; w) )C (T 0 n0; w)g
nite, but also fT 0n0 j 9w : (T n; w) )C (T 0n0 ; w)g.
We will represent rules of the form ! hi> by
Tn
T 0 n0
where hPPT M; T ni j= and (T n; T 0 n0).
Pointer movements
Computational pointer movements are pointer movements not induced by lexical
information. We allow the pointer to leave a node by a computational rule only
when that node has no outstanding requirements. If this is the case, then a
computational pointer movement consists of a (complex) action composed of
go(#) actions, the regular operations and conditionalization:
?; ! hi>:
Axioms 317
We know that hPPT M; T ni j=?; implies that hPPT M; T n0i j=?; for every
T n i T n0 where i 2 f0; 1; #; g; thus movement downward from a node without
requirements is always allowed.
1. ?; ! hgo(#1 ); : : : ; go(#n )i>
moves the pointer from a node with no requirements along a path in the
tree specied by #1 : : : #n (note that along this path no test is made for
absence of requirements). Of course, this action presupposes the presence
of this path.
2. ?; ! hhfh"i>g; go(h"i); 1ii>
moves the pointer one up, if there is such a node.22
9.4 Axioms
In this section we present the axiomatization of the logic of nite ordered binary
trees, of partial trees, of requirements of action modalities and we will present
a selection of computational rules and uodate actions.
9.4.1 Finite Binary Trees
As the basis, we take any suitable axiomatization of the intuitionistic predicate
calculus, together with all instances of the schemas B1{B10, E1{E4, D1{D11
and F1{F3 below under the scope of the operator [LF]. As rules of inference we
take Modus Ponens (if and ! are provable, then so is ) and the rule of
necessitation in [#0], [#1], ["], [#], [#], ["], [LF] and G; that is, if is provable,
then so are [#0], [#1], ["], [#], [#], ["], [LF] and G. Formal proofs
are nite sequences of ws built using the axioms and rules of inference in the
usual way. If a w is provable, then we write ` and say that is a theorem.
With these generalities to hand, let us examine the details. The axioms
split naturally into four groups. First of all, there are the axioms for the basic
operators.
All the axioms of the groups B, D, E and F are under the scope of the operator
[LF].
B1 [#0 ]( ! ) ! ([#0 ] ! [#0 ] )
B2 [#1 ]( ! ) ! ([#1 ] ! [#1 ] )
B3 ["]( ! ) ! (["] ! ["] )
B4 h#0 i ! [#0]
B5 h#1 i ! [#1]
22 Notice that the righthand side of the above rule contains a conditional action h ; ; i
within an existential modality h i.
B6 h"i ! ["]
B7 ! [#0]h"i
B8 ! [#1]h"i
B9 ! ["]h#i
B10 h#1 i> $ h#0 i>
Most of this is familiar from tense/modal logics. B1{B3 are universally valid
modal principles, while B4{B6 re ect the partial functional (or `deterministic')
nature of the 0, 1 relations. B7 and B8 are familiar from tense logic: they
record the fact that both the converse of 0 and the converse of 1 are contained
in the converse of . B9 is closely related and says that the converse of is
contained in 0 [ 1 (to see this, recall that, on full binary trees (in F BT )
h#i is shorthand for h#0 i _ h#1 i). Finally, B10 takes account of the fact that
function and argument daughters come into being together.
The next group of axioms deals with the transitive closure operators and
their interactions with the basic operators.
E1 [# ] $ ( ^ [#][# ])
E2 [" ] $ ( ^ ["][" ])
E3 [# ]( ! [#]) ! ( ! [# ])
E4 [" ]( ! ["]) ! ( ! [" ])
These are familiar from temporal logic and Propositional Dynamic Logic (see
Goldblatt 1992). They re ect the fact that is the re exive transitive closure
of .
The intended meaning of the dened symbols was discussed in the previous
sections; the next group of axioms pins these down.
D1 T n(0) $ ["]?
D2 T n(a0) $ h"0 iT n(a)
D3 T n(a1) $ h"1 iT n(a)
D4 T n(aL) $ hL 1 iT n(a)
D5 h#i $ (h#0 i _ h#1 i)
D6 [#0 ] $ :h#0 i:
D7 [#1 ] $ :h#1 i:
D8 ["] $ :h"i:
D9 [#] $ :h#i:
Axioms 319
D10 [#] $ :h# i:
D11 ["] $ :h" i:
Finally we turn to the axioms that give LOFT its avour: namely, those that
re ect the fact that our intended models are all nite.
F1 h# i[#]?
F2 h" iT n(0)
F3 ! h#i( ^ [#][# ]:)
F1 and F2 are straightforward: no matter where we are in a nite tree, we are
only a nite number of backwards steps away from the root node (which is
what F2 says) and a nite number of steps away from at least one terminal
node (which is what F1 says). More interesting is F3. Roughly speaking, it says
that if holds at any node in a tree, then this node dominates a node not
dominating any other nodes. It is this axiom that will enable us to maintain
the niteness of the tree.23
9.4.2 Partial Trees
Now we turn to principles that axiomatize the trees outside F BT . We relate
the elements of F BT to the trees in P P T in general by the following principles:
LF1 hLFi>
LF2 ! [LF]
LF3 [LF]( _ :)
LF4 [LF] ^ hLFi ! hLFi( ^ )
Notice that we know already that any of the above axioms can be preceded by
[LF], G and thus by ::. This severely restricts the partial trees in P P T . As
all partial trees can be embedded in some structure where all the LOFT axioms
hold (by LF1), no circular or innite structures can arise. But we have argued
(in remark 3) that not all partial trees that can be embedded in an element of
F BT should be allowed in P P T : we want P P T to contain only trees satisfying
the normal form constraints. Thus, we add the following principles:
U1 (h#0 i> ^ h#1 i>) ! (h#i ! (h#0 i _ h#1 i))
U2 h#i> ! (h# i ! ( _ h#ih# i))
23 For more information about this axiomatization see Blackburn and Meyer-Viol 1994.
That is:
If both daughters exist, then the modality h#i is not an underspecied tree re-
lation, but an underspecied statement about a fully specied tree.
If a daughter exists, then the modality h#i can be further specied by one level;
that is, it is partly an underspecied statement (as to the daughter) and partly
a statement about an underspecied tree relation (the relation dangling below
the daughter). That is, an underspecied dominance relation attaches as low as
possible in the tree.
We will also use the standard interaction schema for universal and existential
modalities.
U3 ([#] ^ h#i ) ! h#i( ^ )
9.4.3 Requirements
Now we turn to schemas involving requirements. By denition 17 the following
four schemas hold:
R1 !s?
R2a ?; ! (> !?;)
R2b ?; ! [#]?;
R3 ? 7! 2( _?)
We will also want our models to be `saturated' in the following sense: every
persistent that is neither in F A(T n) nor in V0 ? (0 T n) (i.e. is neither a fact
nor a requirement at T n) is falsied by some T n 2 LFV? (T n); that is,
hPPT M; T 0 n0 i j=s :
The (left-to-right direction of the) following principle expresses this:
R4a G $ (_?)
R4b G?;
Now we have some interaction principles:
R5 G ^ F ! F ( ^ )
R6 F ! hLF i
R7 [LF ] ! G
Axioms 321
9.4.4 Actions
A next group of axiom schemas involves the actions in A(ACT ):
A0a h1i $
A0b [ABORT]?
A1a [put()]
A1b ! hput()i
A1c hput()i ! [put()]
A2 [make(#)]h#i>
A3a ! [go(#)]h# 1 i
A3b hgo(#)i ! h#i
A4a _ h# i ! [merg()]
for ; without question marks
A4b ? _ h# i? ! [merg()](_?)
A4c hmerg()i ! ( _ h# i)
A4d hmerg()i ! [merg()]
The next group of axioms are the standard PDL ones xing the regular opera-
tions. Our axiom for the conditional test, however, is new.
A5 [1 ; 2 ] ! [1 ][2 ]
A6 [1 + 2 ] ! [1 ] ^ [2 ]
A7 [ ] ! ^ [][ ]
A8 [ ]( ! []) ! ( ! [ ])
A9 [h; 1 ; 2 i] ! 8x( ! [1 ]) _ (s9x ^ [2 ])
where x contains all free variables in .
We also have the functionality axioms:
hmake(#)i ! [make(#)]
hgo(#)i ! [go(#)]
if # 2 fh#0i; h#1i; h"ig.
9.4.5 Partial Order

P1 2 !
P2 2 ! 22
P3 2 4 ! 42
P4 hmake(#)i ! 4
P5 hput( )i ! 4
P6 hmerg()i ! 4
Here P 1 and P 2 x respectively the re exivity and transitivity of the relation
, and P 3 states the presence of the set of maxima. P 4{P 6 guarantee that
make(); put() and merg() are -incremental actions.
9.4.6 Logical Forms

Up to this point we have given principles concerning nite partial and binary
trees in general,24 requirements on these trees and actions, mapping partial trees
to some incrementation. What we have not yet done is to x the intended logical
forms, that is, the legal decorations of the tree nodes and their interaction with
tree structure. We intend here only to give a avour of the possible principles
without any intention of being exhaustive.
First, principles determining the structure of our representations and the
trees embodying them:
PERSIST Lai (t) ! 2Lai (t); F o(t) ! 2F o(t)
for t 2 D [ MV ,
PROP G(["]? ! T y(t))
TERM G([#]? ! (T y(e) _ T y(e ! t) _ : : : ))
NP T y(e) ! (h#1 iT y(cn ! t) $ (h#0 i (h#0 iT y(e) ^ h#1 iT y(e ! cn))))
The idea here is to give an exhaustive listing of all possible types annotating
terminal nodes.
FoTy 9xT y(x) ! (9yF o(y)_?9yF o(y))
Ty 8xy(:Eq(x; y) ^ T y(x) ! :T y(y))
The rst principle entails that there are no Type statements without (require-
ments for) Formula statements. The last principle entails that we have no op-
erations like type lifting. Notice that ModP on(a; b) 62 DT y , thus the variables
x and y do not range over it.
24 Analogously, we also need principles regulating partial feature structures and declarative
units.
Axioms 323
ELIM(Fo) h#0 iF o(a) ^ h#1 iF o(b) ! F o(AP L(b; a))
ELIM(Ty) h#0 iT y(a) ^ h#1 iT y(b) ! T y(ModP on(b; a))
These principles relate the decorated tree structure to function application and
type reduction.
Natural languages
In a natural language three sets of actions are involved: lexical actions deter-
mined by L, a set of computational actions C , and a set P of pragmatic actions.
We will give principles involving the set C of computational rules and only hint
at some pragmatic principles.
9.4.7 Computational Rules
The computational actions in C will be given by two kinds of rules. First, there
are the rules that x the interaction between tree structure and tree decorations,
involved in the construction of a labelled term of the lambda calculus. The
second kind is involved in the rewriting of sub-terms that have been constructed.
In other words, one set of rules determines the construction of terms, the other
set of rules determines computation with these terms.
Computational actions may consider the current tree structure as a transient
object. That is, they may check for non-persistent features of trees. Of course,
the computational rules themselves must be persistent.
We will introduce operators !C , :C with subscript C , dened by:
!C df [(+C ) ]( 7! ); :C df [(+C ) ] s :
These connectives will be used to represent invariances respected by the com-
putational rules and the exclusions under such rules respectively. We have seen
an example of their use in our discussion of locality (section 3.2.3). We give
principles axiomatizing the computational rules. In the absence of such princi-
ples there is no guarantee that possible action transitions over a model PPT M
in fact exist. We will present a selection of the computational rules used in this
book.
INTR(Ty) ?T y(ModP on(b; a)) ! hput(?h#0 iT y(a)^?h#1 iT y(b))i>
COMP ?; ^ h#i> ! hgo(h#i)i>
PRED ?h#i ! hmake(#); go(#); put(?)i>
*(WH) (?T y(t) ^ s# >) ! hmake(h# i); go(h# i); put(?T y(e); ?9xT n(x))i>
*(ADV) (?T y(x) ^ s# >) ! hmake(h# i); go(h# i); put(?T y(x))i>
LINK1 F o(x) ^ T y(e) !
hmake(L); go(L); put(?T y(t)); make(h# i); go(h# i); put(?F o(x));
go(h" i)i>
LINK2 F o(x) ^ T y(e) !

hmake(L); go(L); put(?T y(t)); make(h# i); go(h# i); put(F o(x));
go(h" i)i>
LINK3 F o(x) ^ T y(e) ! hmake(L); go(L); put(?T y(t)); put(?hDiF o(x))i>
TOPIC (?T y(t) ^ s# >) !
hmake(h# i); go(h# i); put(F o(U); ?9xF o(x)); go(h" i); make(hL 1 i);
go(hL 1 i); put(?T y (e); F o(U); ?9xF o(x))i>
The ELIM rules work in tandem with the update actions (denition 26). These
actions re ect computations that are particular to a given (system of) feature(s).
For instance,
UPD1(Fo) F o(AP L(x; )) ! hput(F o([ =x]))i>
UPD1(Ty) T y(ModP on( ! ; )) ! hput(T y( ))i>
UPD2(Ty(a,b)) ?T y(a) ! hput(?T y(ModP on(b ! a; b)))i>25
UPD(L) F o(a) ^ hLi(?; ^ F o(b)) ! hput(F o(a ^ b))i>.
In an actual computation, these rules must be used cautiously, and in fact we
will have [UPD2(Ty)]for only a very limited number of types.
Consider ELIM(Fo). In a tree T n on which the antecedent h#0iF o(a) ^
h#1 iF o(b) of the rule holds, we can rewrite the consequent F o(AP L(x; ))
by -reduction (UPD(Fo)), to get F o([ =x]), and so we have a tree node where
the F o value no longer matches the consequent of the ELIM(Fo) rule, although
the antecedent still holds. But, because of the truth clause for atomic DUF (?)
formulae, and the fact that F o(AP L(x; )) F F o([ =x]), the instance of
the ELIM(Fo) rule still holds.
Notice that the COMP rule guarantees the existences of moves from a tree
node to a related one, only once it has no more requirements. This entails that
there may not be any way to move away from a node which still has requirements
outstanding.
Merge
The rule dealing with merging of two tree nodes is a dierent kind of rule from
the ones we have discussed up to now, because there is a check for consistency
involved. Now we require that, when it is consistent to merge, there should
indeed be a tree accessible through a merge action:
Recall that DU (x) = fDu1(x); Du2(x) : : : ; Duk (x)g is a xed enumeration of all
non-empty subsets of the maximal set DU (x) = fLa1(x1 ); : : : Lan(xn ); F o()g.
25 Notice that
?F o([ =x]) ! hput(?F o(AP L(x; )))i>
?F o(AP L(a; b)) ! hput(?h#0 iF o(b)^?h#1 iF o(a))i>
are not present: we may require a given type, but the process of parsing is not goal-directed
in the sense that a specic Formula value is required.
Axioms 325
Let DU (x) and DU (y) be enumerations with disjoint sets of sequences of vari-
ables x and y. Now let NOMERGE be the formula
NOMERGE = f9xy(Dui (x) ^ h# iDui (y) ^ :(Dui (x) ^ Dui (y))) j
W
Dui (x) 2 DU (x); Dui (y) 2 DU (y)g

This formula holds at a node T n if this node is connected by to a node
the atomic information of which is inconsistent with that of T n. Thus, if
NOMERGE holds at a node, then this node cannot merge with the node
dangling below it. Now, we can formulate the existence of merge transitions as
follows
MERGE h# i> ! hf# > _ NOMERGE g; 1; merg()i>
This rule leaves everything as is, if it is evaluated at a tree node which has some
0, 1 or # daughter, or it has a # daughter with information inconsistent with
that holding at the current node; otherwise it merges the current node with its
dangling daughter.
In chapter 3 we discussed a rule called Gap Merge which requires comple-
mentary information decorating the nodes to be merged, i.e. a requirement ?
at the pointed node with a corresponding fact annotating the dangling node.
Gap Merge hmerg()i> ! 9x(?T y(x) ^ h# iT y(x))
9.4.8 Update Actions

We have already seen some update actions (UPD1(Fo) { UPD(L)); here are
some dealing with the quanticational updates of chapter 7:
UPDQ1 T y(t) ^ F o([x =x]) ! hput(F o(8x( ! )))i>,
UPDQ2 T y(t) ^ F o([x =x]) ! hput(F o(9x( ^ )))i>,
UPDQ3 T y(t) ^ F o([Qx =x]) ! hput(F o(Qx( )()))i>, for Q some gener-
alized quantier.
9.4.9 Pragmatic Actions
SUBST ?9xF o(x) ^ T y(e) ! hput(F o(a))i>
SUBST ?9xF o(x) ^ T y(e) ! hhf:C h"i "1 h#0 iF o(a)g; put(F o(a)); ABORTii>
(See footnote 22 for an illucidation of the double brackets in SUBST). So,
pragmatic actions may be licensed by congurational states.
Bibliography
Abusch, D. 1994. The scope of indenites. Natural Language Semantics 2,
83{136.
Alshawi, H. 1990. Resolving quasi logical forms. Computational Linguistics 16,
133{144.
Alshawi, H. 1996. Underspecied rst-order logics. In Semantic Ambiguity and
Underspecication, ed. K. van Deemter and S. Peters, 145{158. Stanford:
CSLI Press.
Alshawi, H., and R. Crouch. 1992. Monotonic semantic interpretation. In
Proceedings of 30th Annual Meeting of the Association of Computational
Linguistics, 32{39. Newark, Delaware.
Anagnostopoulou, E. 1997. Clitic left-dislocation and contrastive left-
dislocation. In Materials on Left Dislocation, ed. E. Anagnostopoulou,
H. Riemsdijk, and F. Zwarts, 151{192. Amsterdam: John Benjamins.
Anagnostopoulou, E., H. van Riemsdijk, and F. Zwarts. 1997. Materials on Left
Dislocation. Amsterdam: John Benjamins.
Aoun, J., and E. Benmamoun. 1998. Minimality, reconstruction and PF move-
ment. Linguistic Inquiry 29, 569{598.
Aoun, J., and L. Choueiri. 1997. Resumption and last resort. MS, USC.
Aoun, J., and Y. Li. 1993. Wh-elements in situ: syntax or LF? Linguistic
Inquiry 24, 199{238.
Baker, C.L. 1970. Notes on the description of English questions: the role of an
abstract question morpheme. Foundations of Language 6, 197{219.
Baker, M. 1996. The Polysynthesis Parameter. Oxford: Oxford University
Press.
Barwise, J., and J. Perry. 1983. Situations and Attitudes. Cambridge, MA: MIT
Press.
Beaver, D. 1997. Presupposition. In Handbook of Logic and Language, ed. J. van
Benthem and A. ter Meulen, 939{1009. Amsterdam: Elsevier.
Beck, S., and S. Berman. 1996. Wh-scope marking: direct vs indirect depen-
dency. In Papers on Wh-Scope Marking, ed. U. Lutz and G. Muller, 59{83.
Stuttgart: University of Stuttgart.
Berman, S. 1991. On the Semantics and Logical Form of Wh-Clauses. Doctoral
dissertation, University of Massachusetts Amherst.
Bibliography 327
Blackburn, P., and W. Meyer-Viol. 1994. Linguistics, logic, and nite trees.
Bulletin of Interest Group of Pure and Applied Logics 2, 2{39.
Blackburn, P., W. Meyer-Viol, and M. de Rijke. 1996. A proof system for nite
trees. In Computer Science Logic, Lecture Notes in Computer Science 1092,
ed. H. Buning, 86{106. Berlin: Springer Verlag.
Borer, H. 1984. Restrictive relatives in modern Hebrew. Natural Language and
Linguistic Theory 2, 219{260.
Borsley, R. 1992. More on the dierence between English restrictive and non-
restrictive relative clauses. Journal of Linguistics 28, 139{148.
Bouma, G., and R. Malouf. 2000. Verbal gerunds as mixed categories in HPSG.
In The Nature and Function of Syntactic Categories, ed. R. Borsley, 133{
166. New York: Academic Press.
Brandner, E. 1996. Was-w construction and clausal typing. In Papers on Wh-
Scope Marking, ed. U. Lutz and G. Muller, 85{106. Stuttgart: University
of Stuttgart.
Breheny, F. 1999. Context-Dependence and Procedural Meaning: The Semantics
of Denites. Doctoral dissertation, University College London.
Breheny, R. 1996. Pro-active Focus. In UCL Working Papers in Linguistics 8,
ed. P. Backley and J. Harris, 21{60. University College London.
Bresnan, J. 1982. The Mental Representation of Grammatical Relations. Cam-
bridge, MA: MIT Press.
Bresnan, J. 1998. Morphology competes with syntax: explaining typological
variation in weak crossover eects. In Is the Best Good Enough? Opti-
mality and Competition in Syntax, ed. P. Barbosa, D. Fox, P. Hagstrom,
M. McGinnis, and D. Pesetsky, 59{92. Cambridge, MA: MIT Press.
Bresnan, J. 1999. The emergence of the unmarked pronoun. MS, University of
Stanford.
Bresnan, J., and J. Grimshaw. 1978. The syntax of free relatives in English.
Linguistic Inquiry 9, 331{391.
Bresnan, J., and J. Kanerva. 1989. Locative inversion in Chichewa: a case study
of factorization in grammar. Linguistic Inquiry 20, 1{50.
Bresnan, J., and L. Moshi. 1990. Object asymmetries in comparative Bantu
syntax. Linguistic Inquiry 21, 147{186.
Brody, M. 1994. On Lexico-Logical Form. Cambridge, MA: MIT Press.
Carston, R. 1996. Enrichment and loosening: complementary processes in de-
riving the proposition expressed. In UCL Working Papers in Linguistics 8,
ed. P. Backley and J. Harris, 61{88. London: University College London.
Chierchia, G. 1991. Anaphora and dynamic binding. Linguistics and Philosophy
15, 111{183.
Chierchia, G. 1992. Questions about quantiers. Natural Language Semantics
1, 281{334.
Chomsky, N. 1971. Deep structure, surface structure and semantic interpreta-
tion. In Semantics, ed. D. Steinberg and D. Jakobovits, 183{216. Cam-
bridge: Cambridge University Press.
328 Bibliography
Chomsky, N. 1981. Lectures on Government and Binding. Dordrecht: Foris.

Chomsky, N. 1986. Knowledge of Language: Its Nature, Origin and Use. New
York: Praeger Publishers.
Chomsky, N. 1995. An Essay on Minimalism. Cambridge, MA: MIT Press.
Cinque, G. 1990. Types of A' Dependencies. Cambridge, MA: MIT Press.
Cinque, G. 1995. Italian Syntax and Universal Grammar. Cambridge: Cam-
bridge University Press.
Cinque, G. 1998. Adverbs and Functional Heads. Cambridge, MA: MIT Press.
Clark, H. 1977. Bridging. In Thinking: Readings in Cognitive Science, ed.
P. Wason and P. Johnson-Laird, 411{420. Cambridge: Cambridge Univer-
sity Press.
Clark, H., and S. Haviland. 1977. Comprehension and the given{new contract.
In Discourse Comprehension and Production, ed. R. Feedle, 1{40. Norwood,
N.J.: Ablex.
Cole, P. 1987. The structure of internally headed relative clauses. Natural
Language and Linguistic Theory 5, 277{302.
Cooper, R. 1995. The role of situations in generalized quantiers. In Handbook of
Contemporary Semantic Theory, ed. S. Lappin, 65{86. Oxford: Blackwell.
Copestake, A., D. Flickinger, and I. Sag. 1999. Minimal recursion semantics:
an introduction. MS, Stanford University.
Corblin, F. 1996. Quantication et anaphore discursive: la reference aux
complementaires. Langages 123, 51{74.
Cormack, A., and R. Kempson. 1991. On specicity. In Meaning and Truth, ed.
J. Gareld and M. Kiteley, 546{582. New York: Paragon House.
Dalrymple, M. 1993. The Syntax of Anaphoric Binding. Stanford: CLSI Press.
Davidson, D. 1967. Truth and meaning. Synthese 17, 304{323.
Dayal, V. 1994. Scope marking as indirect Wh dependency. Natural Language
Semantics 2, 137{170.
Dayal, V. 1996. Locality in WH Quantication. Dordrecht: Kluwer.
Deemter, K. van. 1996. Towards a logic of ambiguous expressions. In Semantic
Ambiguity and Underspecication, ed. K. van Deemter and S. Peters, 203{
238. Stanford: CSLI Press.
Dekker, P. 1996. Coreference and representationalism. In Reference and
Anaphoric Relations, ed. K. von Heusinger and U. Egli, 287{310. Dor-
drecht: Kluwer.
Demirdache, H. 1991. Resumptive Chains in Restrictive Relatives, Appositives,
and Dislocation Structures. Doctoral dissertation, MIT.
Demirdache, H. 1997. Dislocation, resumption and weakest crossover. In Mate-
rials on Left Dislocation, ed. E. Anagnostopoulou, H. van Riemsdijk, and
F. Zwarts, 193{232. Amsterdam: John Benjamins.
Dennett, D. 1987. The Intentional Stance. Cambridge, MA: MIT Press.
Dennett, D. 1993. Learning and labeling. Mind and Language 8, 540{593.
Diesing, M. 1992. Indenites. Cambridge, MA.: MIT Press.
Bibliography 329
Dobrovie-Sorin, C. 1990. Clitic-doubling, WH movement and quantication in
Romanian. Linguistic Inquiry 21, 351{398.
Dowty, D. 1987. Type raising, functional composition and nonconstituent con-
junction. In Categorial Grammars and Natural Language Structures, ed.
R. Oehrle, E. Bach, and D. Wheeler, 153{198. Dordrecht: Reidel.
Dowty, D. 1993. Variable-free syntax, variable-binding syntax, the natural
deduction Lambek calculus and the crossover constraint. West Coast Con-
ference of Formal Linguistics 11, 161{176.
DuÆeld, N. 1995. Particles and Projections in Irish Syntax. Dordrecht: Kluwer.
Edwards, M., R. Kempson, and W. Meyer-Viol. forthcoming. Resumptive pro-
nouns in English and Arabic. In Proceedings of 4th Afroasiatic Conference,
ed. R. Hayward, J. Ouhalla, and D. Perrett. Amsterdam. John Benjamins.
Enc, M. 1986. Towards a referential analysis of temporal expressions. Linguistics
and Philosophy 9, 405{425.
Enc, M. 1991. The semantics of specicity. Linguistic Inquiry 22, 1{25.
Engdahl, E. 1986. Constituent Questions. Dordrecht: Reidel.
Erteschek-Shir, N. 1997. The Grammar of Focus. Cambridge: Cambridge Uni-
versity Press.
Escobar, L. 1997. Clitic left dislocation and other relatives. In Materials on
Left Dislocation, ed. E. Anagnostopoulou, H. van Riemsdijk, and F. Zwarts,
233{273. Amsterdam: John Benjamins.
Espinal, T. 1991. The representation of disjunct constituents. Language 67,
726{762.
Evans, G. 1980. On pronouns. Linguistic Inquiry 11, 337{362.
Fabb, N. 1990. The dierence between English restrictive and nonrestrictive
relative clauses. Journal of Linguistics 26, 58{77.
Fanselow, G. 1997. Partial Movement. MS, University of Potsdam.
Fanselow, G., and A. Mahajan. 1996. Partial movement and successive cyclicity.
In Papers on Wh-Scope Marking, ed. U. Lutz and G. Muller, 131{161.
Farkas, D. 1981. Quantier scope and syntactic islands. In Proceedings of
Chicago Linguistic Society 17, ed. R. Hendrick et al., 59{66. Chicago:
University of Chicago Press.
Farkas, D. 1997. Evaluation indices and scope. In Ways of Scope Taking, ed.
A. Szabolcsi, 183{216. Dordrecht: Reidel.
Fernando, T. 1994. What is a DRS? In Proceedings of the International Work-
shop on Computational Semantics, ed. H. Bunt, R. Muskens, and G. Ren-
tier, 61{70. Tilburg. ITK.
Finger, M., and D. Gabbay. 1992. Adding a temporal dimension to a logical
system. Journal of Logic, Language and Information 1, 203{233.
Fodor, J. A. 1981. Representations. Cambridge, MA: MIT Press.
Fodor, J. A. 1983. Modularity of Mind. Cambridge, MA: MIT Press.
Fodor, J. A. 1998. Concepts: Where Cognitive Science Went Wrong. Oxford:
Clarendon Press.
330 Bibliography
Fodor, J. A., and J. Lepore. 1992. Holism: A Shopper's Guide. Oxford: Black-
well.
Fodor, J. D., and I. Sag. 1982. Referential and quanticational indenites.
Linguistics and Philosophy 5, 355{398.
Gabbay, D. 1996. Labelled Deductive Systems. Oxford: Oxford University Press.
Gabbay, D., and I. Hodkinson. 1996. Temporal logic in the context of databases.
In Logic and Reality, ed. J. Copeland, 69{87. Oxford: Oxford University
Press.
Garrod, S., and A. Sanford. 1981. Bridging inferences and the extended domain
of reference. In Attention and Performance IX, ed. J. Long and A. Boddeley,
331{346. Hillsdale: Lawrence Erlbaum.
Gartner, H. M. 1997. Are there V2 relative clauses in German? MS, University
of Potsdam.
Ginzburg, J. 1992. Questions, Queries, and Facts: A Semantics and Pragmatics
for Interrogatives. Doctoral dissertation, University of Edinburgh.
Ginzburg, J. 1995. Resolving questions I. Linguistics and Philosophy 18, 459{
527.
Ginzburg, J., and I. Sag. 1999. English Interrogative Constructions. MS, Stan-
ford University and University of Jerusalem.
Givon, T. 1976. Topic, pronoun and grammatical agreement. In Subject and
Topic, ed. C. Li and S. Thompson, 149{188. New York: Academic Press.
Goldblatt, R. 1992. Logics of Time and Computation. Stanford: CLSI Press.
Gorrell, P. 1995. Syntax and Parsing. Cambridge: Cambridge University Press.
Groenendijk, J., and M. Stokhof. 1989. Type-shifting rules and the semantics of
interrogatives. In Properties, Types and Meaning: Vol.2: Semantic Issues,
ed. G. Chierchia, R. Turner, and B. Partee, 21{68. Dordrecht: Kluwer.
Groenendijk, J., and M. Stokhof. 1991. Dynamic predicate logic. Linguistics
Groenendijk, J., and M. Stokhof. 1997. Questions. In Handbook of Logic and
Language, ed. J. van Benthem and A. ter Meulen, 1055{1125. Amsterdam:
Elsevier.
Grosu, A. 1996. Missing-P free relative constructions. Linguistic Inquiry 27,
257{293.
Hale, K. 1983. Warlpiri and the grammar of noncongurational languages.
Natural Language and Linguistic Theory 1, 5{47.
Hausser, R. 1989. Computation of Language. Berlin: Springer.
Hausser, R. 1999. Foundations of Computational Linguistics. Berlin: Springer.
Heim, I. 1982. The Syntactic and Semantic Properties of Denite and Indef-
inite Noun Phrases. Doctoral dissertation, University of Massachusetts,
Amherst.
Heim, I. 1990. E-type pronouns and donkey anaphora. Linguistics and Philos-
ophy 13, 137{177.
Hepple, M. 1990. The Grammar and Processing of Order and Dependency: A
Categorial Approach. Doctoral dissertation, University of Edinburgh.
Bibliography 331
Heusinger, K. von. 1997. Denite descriptions and choice functions. In Logic,
Language and Computation, ed. S. Akama, 61{92. Dordrecht: Kluwer.
Heycock, C. 1995. Asymmetries in reconstruction. Linguistic Inquiry 26, 547{
570.
Higginbotham, J. 1987. The autonomy of syntax and semantics. In Knowl-
edge Representations and Language Understanding, ed. J. Gareld, 119{
131. Cambridge, MA: MIT Press.
Higginbotham, J. 1996. The semantics of questions. In The Contemporary
Handbook of Semantic Theory, ed. S. Lappin, 361{385. Oxford: Blackwell.
Higginbotham, J., and R. May. 1981. Questions, quantiers and crossing.
Linguistic Review 1, 41{79.
Hilbert, D., and P. Bernays. 1939. Grundlagen der Mathematik II. Berlin: Julius
Springer.
Hirschbuhler, P. 1981. The ambiguity of iterated multiple questions. Linguistic
Inquiry 135{146.
Hohle, T. 1996. The w-w{construction: appositive or scope indicating? In
Papers on Wh-Scope Marking, ed. U. Lutz and T. Muller, 37{58. Stuttgart:
University of Stuttgart.
Hoji, H. 1986. Scope interpretation in Japanese and its theoretical implications.
In Proceedings of West Coast Conference in Linguistics, ed. M. Dalrymple
et al., 87{101.
Horvath, J. 1997. The status of `wh-expletives' and the partial wh-movement
construction of Hungarian. Natural Language and Linguistic Theory 15,
509{572.
Hoshi, K. 1995. Structural and Interpretive Aspects of Head-Internal and Head-
External Relative Clauses. Doctoral dissertation, University of Rochester,
New York.
Huang, J. 1982. Logical Relations in Chinese and the Theory of Grammar.
Doctoral dissertation, MIT.
Inoue, H., and J. D. Fodor. 1995. Information-paced parsing of Japanese. In
Japanese Sentence Processing, ed. R. Masuka and N. Nagai, 9{63. New
York: Lawrence Erlbaum.
Jackendo, R. 1992. Languages of the Mind: Essays on Mental Representation.
Cambridge, MA: MIT Press.
Jacobson, P. 1996. The syntax-semantics interface in categorial grammar. In
The Handbook of Contemporary Semantic Theory, ed. S. Lappin, 89{116.
Oxford: Blackwell.
Jiang, Y. 1995. Logical Dependency in Quantication. Doctoral dissertation,
University College London.
Johnson, D., and S. Lappin. 1997. A critique of the minimalist program.
Johnson, D., and S. Lappin. 1999. Local Constraints vs Economy. Stanford:
CSLI Press.
332 Bibliography
Joshi, A., L. Levy, and M. Takahasihi. 1975. Tree adjunct grammars. Journal
of Computer and System Sciences 10, 136{163.
Kadmon, N. 1990. On uniqueness. Linguistics and Philosophy 13, 273{324.
Kamp, H. 1984. A theory of meaning and semantic representation. In Truth,
Interpretation, and Information, ed. T. Janssen and M. Stokhof, 1{34. Dor-
drecht: Foris.
Kamp, H. 1996. Discourse Representation Theory and Dynamic Semantics:
representational and nonrepresentational accounts of anaphora. MS, IMS,
Stuttgart.
Kamp, H., and U. Reyle. 1993. From Discourse to Logic. Dordrecht: Kluwer.
Kaplan, R., and A. Zaenen. 1989. Long-distance dependencies, constituent
structure, and functional uncertainty. In Alternative Conceptions of Phrase
Structure, ed. M. Baltin and A. Kroch, 17{42. Chicago: University of
Chicago Press.
Karttunen, L. 1977. Syntax and semantics of questions. Linguistics and Phi-
losophy 1, 3{44.
Kayne, R. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press.
Keenan, E. 1985. Relative clauses. In Language Typology and Syntactic De-
scription: Volume II - Complex Constructions, ed. T. Shopen, 141{170.
Cambridge: Cambridge University Press.
Kempson, R. 1995. Ellipsis in a labelled deduction system. Bulletin of the
Interest Group in Pure and Applied Logics 3, 489{526.
Kempson, R. 1996. Semantics, pragmatics and natural-language interpretation.
In Handbook of Contemporary Semantic Theory, ed. S. Lappin, 561{598.
Oxford: Blackwell.
Kempson, R., and A. Cormack. 1981. Quantication and ambiguity. Linguistics
Kempson, R., and D. Gabbay. 1998. Crossover: a unied view. Journal of
Linguistics 34, 73{124.
Kempson, R., W. Meyer-Viol, and D. Gabbay. 1998. VP Ellipsis: towards a dy-
namic, structural account. In Fragments: Studies in Ellipsis and Gapping,
ed. S. Lappin and S. Benmamoun, 227{290. Oxford: Oxford University
Press.
Kibble, R. 1997. Anaphora Resolution in Dynamic Semantics. Doctoral disser-
tation, University of Edinburgh.
Kim, I. K. 1997. Agreement Phrases in DP. In University College London Work-
ing Papers in Linguistics, ed. P. Backley and J. Harris, 281{302. London:
Kiss, K. 1991. Congurationality in Hungarian. Dordrecht: Reidel.
Kuroda, S. Y. 1999. Shubu naizai kankei-setu (head internal relative clauses).
In Kotoba no Kaku to Shuuen, ed. S.Y. Kuroda and M. Nakamura, 17{103.
Tokyo: Kuroshio Shuppan.
Lappin, S., and N. Francez. 1994. E-type pronouns, I-sums, and donkey
anaphora. Linguistics and Philosophy 17, 391{428.
Bibliography 333
Lappin, S., and T. Reinhart. 1988. Presuppositional eects of strong determin-
ers: a processing account. Linguistics 26, 1021{1037.
Lasnik, H., and T. Stowell. 1991. Weakest crossover. Linguistic Inquiry 22,
687{720.
Leisenring, A.C. 1969. Mathematical Logic and Hilbert's Epsilon-Symbol. Lon-
don: MacDonald Technical and Scientic.
Lemmon, E.J. 1965. Beginning Logic. London: Nelson.
Mahajan, A. 1990. The A/A-bar Distinction and Movement Theory. Doctoral
dissertation, MIT.
Maling, J., and A. Zaenen. 1981. A phrase-structure account of Scandinavian
extraction phenomena. In The Nature of Syntactic Representation, ed.
P. Jacobson and G. Pullum, 229{282. Dordrecht: Reidel.
Manzini, R., and S. Savoia. 1998. The syntax of object clitics si in Italian
dialects. MS, University of Florence.
Manzini, R., and K. Wexler. 1987. Binding theory, parameters, and learnability.
Linguistic Inquiry 18, 413{444.
Maracz, L., and P. Muysken. 1988. Congurationality: The Typology of Asym-
metries. Dordrecht: Foris.
Marcus, M. 1980. A Theory of Syntactic Recognition for Natural Language.
Cambridge, MA: MIT Press.
Marcus, M. 1987. Deterministic parsing and description theory. In Linguistic
Theory and Computer Applications, ed. P. Whitelock, M. Wood, H. Somers,
R. Johnson, and P. Bennett, 69{112. New York: Academic Press.
Marcus, M., D. Hindle, and M. Fleck. 1983. D-Theory: Talking about talking
about trees. In Proceedings of 21st Annual Meeting of the Association of
Computational Linguistics, 129{136. Cambridge, MA.
Marten, L. 1999. Syntactic and Semantic Underspecication in the Verb Phrase.
Doctoral dissertation, School of Oriental and African Studies, University of
London.
May, R. 1985. Logical Form. Cambridge, MA: MIT Press.
McCawley, J. 1968. Concerning the base component of a transformational
grammar. Foundations of Language 4, 243{269.
McCloskey, J. 1988. Resumptive pronouns, A' binding and levels of representa-
tion in Irish. In The Syntax of the Modern Celtic Languages, ed. R. Hen-
dricks, 199{248. New York: Academic Press.
McDaniel, D. 1989. Partial movement and multiple wh-movement. Natural
Meyer-Viol, W. 1995. Instantial Logic. Doctoral dissertation, University of
Utrecht.
Milward, D. 1994. Dynamic Dependency Grammar. Linguistics and Philosophy
17, 561{606.
Milward, D. 1995. Dynamics and situations. In FRACAS Deliverable D15, ed.
R. Cooper et al., 194{216. Edinburgh: University of Edinburgh.
334 Bibliography
Monachesi, P. 1998. Decomposing Italian clitics. In Romance in HPSG, ed.

S. Balari and L. Dini, 305{359. Stanford: CSLI Press.
Montague, R. 1974a. English as a formal language. In Formal Philosophy:
Selected Papers of Richard Montague, ed. R. Thomason, 108{122. New
Haven: Yale University Press.
Montague, R. 1974b. The proper treatment of quantication in ordinary En-
glish. In Formal Philosophy: Selected Papers of Richard Montague, ed.
R. Thomason, 247{270. New Haven: Yale University Press.
Montalbetti, A. 1984. After Binding: On the Interpretation of Pronouns. Doc-
toral dissertation, MIT.
Moortgat, M. 1997. Categorial type logics. In Handbook of Logic and Language,
ed. J. van Benthem and A. ter Meulen, 5{92. Amsterdam: Elsevier.
Morrill, G. 1994. Type-Logical Grammar. Dordrecht: Kluwer.
Murasugi, K. 1996. Head-internal relative clauses as adjunct pure complex NPs.
In Synchronic and Diachronic Approaches to Language, ed. Chiba S. et al.,
425{437. Tokyo: Liber Press.
Neale, S. 1990. Descriptions. Cambridge, MA: MIT Press.
Oehrle, R. 1995. Term-labeled categorial type systems. Linguistics and Philos-
ophy 17, 633{678.
Otsuka, M. 1999. Head-internal relatives from an LDSNL perspective. Master's
thesis, School of Oriental and African Studies, University of London.
Ouhalla, J. 1991. Functional Categories and Parametric Variation. London:
Routledge.
Ouhalla, J. 1996. Remarks on the binding properties of wh-pronouns. Linguistic
Inquiry 27, 676{707.
Pafel, J. 1996. On scope in German wh-sentences including w...w-constructions.
Pafel, J. 1999. Interrogative quantiers within scope. Linguistics and Philosophy
22, 255{310.
Partee, B. 1973. Some structural analogies between tenses and pronouns in
English. Journal of Philosophy 70, 601{609.
Partee, B. 1976. Montague Grammar. New York: Academic Press.
Perrett, D. 2000. The Dynamics of Tense Construal in Hadiyya. Doctoral
dissertation, School of Oriental and African Studies, University of London.
Pesetsky, D. 1987. Wh-in-situ: movement and unselective binding. In The
Representation of (In)deniteness, ed. E. Reuland and A. ter Meulen, 98{
129. Cambridge, MA: MIT Press.
Pinkal, M. 1995. Logic and Lexicon. Dordrecht: Reidel.
Poesio, M. 1996. Semantic ambiguity and perceived ambiguity. In Semantic
Ambiguity and Underspecication, ed. K. van Deemter and S. Peters, 159{
202. Stanford: CSLI Press.
Pollard, C., and I. Sag. 1994. Head-driven Phrase-Structure Grammar. Chicago:
University of Chicago Press.
Bibliography 335
Pollard, C., and E. Yoo. 1998. A unied theory of scope for quantiers and
Wh-phrases. Journal of Linguistics 34, 415{444.
Postal, P. 1972. Crossover Phenomena. New York: Holt, Rinehart, and Win-
ston.
Postal, P. 1993. Remarks on weak crossover eects. Linguistic Inquiry 24,
539{556.
Prawitz, D. 1965. Natural Deduction: A Proof-Theoretical Study. Stockholm:
Almqvist and Wiksell.
Progovac, L. 1991. Polarity in Serbo-Croatian: anaphoric NPIs and pronominal
PPIs. Linguistic Inquiry 22, 567{592.
Radford, A. 1977. Italian Syntax: Transformational and Relational Grammar.
Cambridge: Cambridge University Press.
Reinhart, T. 1997. Wh-in-situ in the framework of the minimalist program.
Natural Language Semantics 6, 29{56.
Reinhart, T. forthcoming. Interface Strategies. Cambridge, MA: MIT Press.
Reinhart, T., and E. Reuland. 1993. Re exivity. Linguistic Inquiry 24, 657{720.
Reis, M. 1996. On `was'-parentheticals and `was{w'-constructions in German.
Reis, M., and I. Rosengren. 1992. What do wh-imperatives tell us about wh-
movement. Natural Language and Linguistic Theory 10, 79{118.
Reyle, U. 1993. Dealing with ambiguities by underspecication: construction,
representation and deduction. Journal of Semantics 10, 123{179.
Reyle, U. 1996. Co-indexing labeled DRSs to represent and reason with ambi-
guities. In Semantic Ambiguity and Underspecication, ed. K. van Deemter
and S. Peters, 239{269. Stanford: CSLI Press.
Rizzi, L. 1990. Relativized Minimality. Cambridge, MA: MIT Press.
Roberts, C. 1989. Modal subordination and anaphora in discourse. Linguistics
Rochemont, M., and P. Culicover. 1990. English Focus Constructions and the
Theory of Grammar. Cambridge: Cambridge University Press.
Rogers, J., and K. Vijay-Shanker. 1996. Towards a formal understanding of
the determinism hypothesis in D-Theory. In Recent Advances in Parsing
Technology, ed. H. Bunt and M. Tomita, 59{78. Dordrecht: Kluwer.
Rooth, M. 1992. A theory of focus interpretation. Natural Language Semantics
2, 75{116.
Rooth, M. 1996. Focus. In The Handbook of Contemporary Semantic Theory,
ed. S. Lappin, 271{299. Oxford: Blackwell.
Ross, H. 1967. Constraints on Variables in Syntax. Doctoral dissertation, MIT.
Russo, A. 1996. Modal Logics as Labelled Deductive Systems. Doctoral disser-
tation, Imperial College, University of London.
Sar, K. 1986. Relative clauses in a theory of binding and levels. Linguistic
Inquiry 17, 663{83.
336 Bibliography
Sar, K. 1996. Derivation, representation and resumption. Linguistic Inquiry

27, 313{340.
Sag, I. 1997. English relative clauses. Journal of Linguistics 33, 431{483.
Sag, I., and J. D. Fodor. 1994. Extraction without traces. MS, Stanford.
Shlonsky, U. 1993. Resumptive pronouns as a last resort. Linguistic Inquiry 23,
443{368.
Silvers, S. 1989. Re-representation. Dordrecht: Kluwer.
Simpson, A. 1995. Wh movement and the theory of feature-checking. Doctoral
dissertation, School of Oriental and African Studies, University of London.
Simpson, A. 1999. On the status of modifying de and the structure of the
Chinese DP. In On the Formal Way to Chinese Languages, ed. S-W. Tan
and L. Chen-Sheng, 569{597. Stanford: CSLI Press.
Speas, M. 1990. Phrase Structure in Natural Language. Dordrecht: Kluwer.
Sperber, D., and D. Wilson. 1986, 1995. Relevance: Cognition and Communi-
cation. Oxford: Blackwell.
Sportiche, D. 1992. Clitic Constructions. MS, University of California, Los
Angeles.
Stalnaker, R. 1974. Pragmatic presuppositions. In Semantics and Philosophy,
ed. C. Munitz and F. Unger, 197{214. New York: New York University
Press.
Stalnaker, R. 1984. Inquiry. Cambridge, MA: MIT Press.
Steedman, M. 1996. Surface Structure and Interpretation. Cambridge, MA:
MIT Press.
Suner, M. 1998. Resumptive restrictive relatives: a cross-linguistic perspective.
Language 74, 335{364.
Swinburne, D. 1999. LINK and the Dynamics of Utterance Interpretation. Doc-
toral dissertation, School of Oriental and African Studies, University of
London.
Szabolcsi, A. 1983. The possessor that ran away from home. The Linguistic
Review 3.
Szabolcsi, A. 1997. Strategies for scope taking. In Ways of Scope Taking, ed.
A. Szabolcsi, 109{154. Dordrecht: Kluwer.
van der Does, J. 1996. An E-type logic. In Semantic Ambiguity and Under-
specication, ed. K. van Deemter and S. Peters, 555{570. Stanford: CSLI
Press.
van Eijck, J., and H. Kamp. 1997. Representing discourse in context. In Hand-
book of Logic and Language, ed. J. van Benthem and A. ter Meulen, 179{
239. Amsterdam: Elsevier.
Vat, J. 1997. Left dislocation, connectedness and reconstruction. In Materials on
Left Dislocation, ed. E. Anagnostopoulou, H. van Riemsdijk, and F. Zwarts,
67{92. Amsterdam: John Benjamins.
Wahba, W. 1991. LF movement in Iraqi Arabic. In Logical Structure and Lin-
guistic Structure, ed. J. Huang and R. May, 253{276. Dordrecht: Kluwer.
Bibliography 337
Watanabe, A. 1992. Subjacency and s-structure movement of WH in situ.
Journal of East Asian Linguistics 1, 255{291.
Williams, E. 1994. Thematic Structure in Syntax. Cambridge, MA: MIT Press.
Winter, Y. 1997. Choice functions and the scopal semantics of indenites.
Yiang, J. 1995. Logical Dependency and Quantication. Doctoral dissertation,
Zribi-Hertz, A., and L. Mbolatianavalona. 1999. Towards a modular theory of
linguistic deciency: evidence from Malagasy personal pronouns. Natural
General Index
In this index boldface indi- ADR, see tree node predicate
cates a primary occurrence of agreement, 93, 126, 255
the item in question. Alshawi, 25
ambiguity, 9, 11, 26, 68, 133, 255,
abduction, 57 260
Abusch, 226 indenite, 226
acceptability, 133, 216, 217, 257, 262, lexical, 27, 224
263; see unacceptability scope, 227, 307
accommodation, 14, 250 structural, 27
action, 77{80, 269, 273, 307{317 wh question, 151, 153, 155
set ACT , 269, 312, 313 Anagnostopoulou, 127, 130, 219
axiomatization, 321 anaphora, 8, 224, 261
basic bridging, 10
Abort, 309 cross-sentential, 10
go, 90, 310 E-type, 10, 249{251
identity, 309 indexical, 10
make, 90, 309 pronominal, see pronoun
merge, 310 semantic account, 15
put, 90, 309 annotation, 156, 169, 303, 309; see
computational, 77, 80{89, 214, success function
252, 254, 257, 258, 307, 315{ antecedent, 214, 216, 217, 224, 228,
317 253
axiomatization, 323 Aoun, 16, 194, 223
pointer movement, 316 Arabic, 79, 95, 106, 124, 128, 193,
conditional, 90, 311 194, 203, 208, 212, 256
freshput, 92, 237, 238, 312 classical, 16, 204, 220
gofirst# , 311 Egyptian, 106, 121, 130, 131
gofirst" , 91, 311 Iraqi, 159, 160, 162, 163, 167
incremental, 269, 308 Lebanese, 194
lexical, 59, 69, 77, 89{98, 113, standard, 220
145, 214, 235, 237, 254, 257, AT OM , 268, 270, 293
264, 307, 315 AT OMF , 285
complex, 90 attribute value matrix, 279, 284, 287,
condition, 90 293
modality, 308 Axiom, 69, 76, 307, 312, 316
requirement, 317 axiomatization, 317
pointer movement, 311
pragmatic, 16, 62, 78, 95, 98, Baker, 146, 148, 188
135, 214, 224, 252, 254, 255, Barwise, 5, 9, 34, 40
257, 307 Beaver, 14
axiomatization, 325 Beck, 165, 179
Adjunction, 64, 81, 85, 144 Benmamoun, 16, 194
*Adjunction, 81, 85, 127, 172 Berman, 165, 179, 223
General Index 339
Bernays, 35 context dependence, 8, 12, 14, 63,
-reduction, 63, 271, 272, 280, 323, 95, 260
324 control feature, 289
binder, 232, 245 Cooper, 5, 12, 13, 34
Blackburn, 36, 272 Copestake, 25, 231
Borer, 205 Cormack, 226, 228
Borsley, 108 crossover, 190, 222, 257, 258
Bouma, 150 extended strong, 199
Brandner, 166 secondary strong, 193, 199
Breheny, 12, 212, 216, 223 strong, 192, 193, 195, 206, 214
Bresnan, 99, 101, 120, 195, 213, 218, weak, 192, 193, 195, 199, 206,
220, 222 213, 218
Brody, 166, 194 in questions, 213{216
weakest, 192, 193
Carston, 216 Crouch, 25
case, 69, 134, 165 Culicover, 212
accusative, 93, 136, 300
nominative, 176 Dalrymple, 101, 148
categorial grammar, 33, 41, 109, 195, daughter, 39, 110
225, 258 argument, 39, 272, 275, 295
Chierchia, 10, 13 function, 39, 272, 275, 295
Chinese, 106, 139, 142, 158, 221, Dayal, 163, 165
228 declarative, 268
choice function, 152 structure, 273{307
Chomsky, 2, 99, 108, 148, 166, 192, declarative unit, 233, 268, 281, 284,
210, 212, 254, 258 285, 286, 302
Choueiri, 16, 194 DU formula, 76
Cinque, 128, 132 DU language, 38
Clark, 10 DUF language, 32, 284{297
clitic doubling, 258 DU set, 50
clitic left-dislocation, 130, 132, 133, partial, 32, 293, 303
219, 256 Dekker, 9
clitic raising, 258 Demirdache, 203, 205, 220
Cole, 140, 141 Dennett, 7
competence linguistic, 267 dependency long-distance, 63{67, 105{
Completion, 59, 63, 81, 83 144, 150{185, 190{220
compositionality, 99, 152, 261 destructor APL, 271
connective determiner, 11, 113, 114, 116, 167,
dynamic, 45 207, 240
static, 45 indenite, 224, 227, 231, 235
consistency, 305 quantifying, 223
lemma, 305 wh, 215
content development path, 269
model-theoretic, 20, 261 Diesing, 75, 224, 255
context, 12 disambiguation, 25, 26, 263
340 General Index
discontinuity, 100; see long-distance feature structure, 278, 284, 287

dependency partial, 280
Discourse Representation Theory, 5, F e-morphism, 279
9, 12, 15, 25, 99, 225 Fernando, 12
Dobrovie-Sorin, 125, 128, 132, 214 Finger, 34, 284
DOM predicate, 235 focus, 15, 212
dominance, 39, 49, 109, 113, 272, Fodor, J.A., 7, 8, 266
275, 320 Fodor, J.D., 67, 69, 150, 224, 226
immediate, 272, 275 formal language, 1, 2, 8, 99, 260
Dowty, 33, 195, 258 Formula, 27, 33{36, 113, 216, 285
DRT, see Discourse Representation formula
Theory labelled, 271, 278
D-Tree Grammar, 25 Francez, 10
DU descriptions, 49 function spine, see locality
DuÆeld, 126 functional uncertainty, 101
Dutch, 106, 208, 209, 220
Dynamic Predicate Logic, DPL, 10 Gabbay, 27, 31, 34, 99, 198, 221,
Dynamic Semantics, 9 228, 284
Gap Merge, 212
Elimination, 63, 81, 82 genitive, 140, 144{148, 196, 199, 206
ellipsis, 301 German, 79, 106, 154, 157, 163, 164,
Engdahl, 157 170{181, 229, 258
English, 79, 106{120, 124, 157, 160, Ginzburg, 151, 153, 157, 158, 223
196, 203, 211, 212, 218, 224, Givon, 258
230, 254, 264, 300 goal, 77
Epsilon Calculus, 35 goal-directed, 270, 302
epsilon term, 232, 235, 246, 248, 249 goal-directedness, 18, 44, 138, 253,
Erteschek-Shir, 212 258, 259, 265
Escobar, 128 Goldblatt, 318
Espinal, 108 Gorrell, 67
Evans, 10 grammar formalism, 2, 68, 75, 99,
expletive, 125, 255, 256 260, 262, 262
indenite, 125 grammaticality, 315; see well-formedness
tense, 256 Grimshaw, 120
wh, 163, 256, 258 Groenendijk, 9, 10, 12, 151, 153,
German, 172, 174{181 154
Hungarian, 181{183 Grosu, 120
Iraqi Arabic, 184{185 growth, 313, 314
f-structure, see LFG Hadiyya, 256
F-structure, 278{281; see feature struc- Hale, 67, 75, 255
ture Hausser, 101, 259
extension, 282 Hebrew, 126, 129, 131, 133, 193, 203{
Fabb, 108 208
Fanselow, 166, 179, 181 Heim, 10, 12, 13
Farkas, 226, 227 Hepple, 195, 222
General Index 341
Heusinger, 249 Kuroda, 140
Higginbotham, 2, 151, 153
Hilbert, 35, 245 Label, 27, 285
Hindi, 106 label tree node, 296
Hohle, 176, 181 Labelled Deductive Systems, 27
Hoji, 229 labelled formula, 271
Horvath, 165 lambda calculus, 33, 38, 231, 234,
Hoshi, 140, 141, 143 271
HPSG, 2, 99, 100, 109, 168, 217, typed, 295
225, 281 lambda term, 231, 271
Huang, 159, 160 unreduced, 20
Hungarian, 165, 167, 181{183 language L(L; C; P ), 78
Lappin, 10, 154, 168, 185, 213, 215
implication, 45 Lasnik, 192, 201
incremental, 307 left-dislocation, 63{67, 111, 130, 131,
incrementality, 79 194, 210, 212, 214, 217, 218;
Inoue, 67, 69 see long-distance dependency
input string, 272, 273 Leisenring, 245
Introduction, 57, 80, 82, 264 Lemmon, 15
Irish, 126 Lepore, 7
island constraints, 15, 120, 126, 128, LFG, 99, 100, 213, 254
129, 194 Li, 223
Italian, 128{130, 132, 133, 219 LINK, 104, 105, 110, 109{148, 203,
255
Jackendo, 3 LINK Introduction, 111{130, 196, 205,
Jacobson, 150 211, 256, 264
Japanese, 67{75, 79, 93, 95, 106, Arabic, 123
107, 133, 134, 138, 140, 229, English, 111, 113, 117
250, 264, 300 Japanese, 135, 139, 141
Jiang, 228 locality, 124, 128, 133, 177, 197, 198,
Johnson, 154, 168, 185, 213 203, 214
Kadmon, 10, 12, 13, 250 LOFT, 36, 51, 64, 66, 86, 265, 305
Kamp, 5, 9{12, 34, 40, 99 functionality axioms of, 305
Kaplan, 100, 254 Logic of Finite Trees, see LOFT
Karttunen, 151 logical form, 4, 8, 56, 198, 234, 258,
Kayne, 109, 143, 148, 256 259, 269, 290, 291, 295, 307,
Keenan, 140, 143 315
Kempson, 9, 11, 198, 221, 226{228 axiomatization, 322
Kibble, 10 normal, 270
Kiss, 67, 75, 255 partial, 307
PT+ , 271
Kleene star operator, see dependency set of, 269
long-distance logical name fresh, 312
knowledge, linguistic, 259, 266, 267 long-distance dependency, 6, 24, 84{
Korean, 140 88, 100, 164, 224, 227, 254,
Kripke model, 284 258
342 General Index
Mahajan, 163, 179, 181 Montague, 223, 225, 260

Maling, 15 Montalbetti, 210
Manzini, 100, 148 Morrill, 2, 33, 109, 150, 225
Marcus, 5, 25, 75 Murasugi, 140
Marten, 33, 96, 173 Muysken, 255
May, 151
Mbolatianavalona, 12 name
McCawley, 3 arbitrary, 35, 224, 245
McCloskey, 126 proper, 69, 92, 231
McDaniel, 154, 166 natural deduction, 35, 245
Merge, 81, 86, 84{88, 139, 198, 199, natural language, 1{3, 6{9, 16, 312,
201, 208, 209, 211, 214, 216, 323
217, 220, 221, 300, 309, 324 Neale, 10, 12, 223
Basic, 86 negation, 45
Gap Merge, 87, 208, 209, 220, classical, 289, 290
325 restricted, 315
meta-variable, 33, 34, 44, 61, 78, strong, 289
136, 137, 282 node
denite, 237 address, 296
indenite, 227, 228, 231, 240 description, 50{51, 303, 304, 306
pronominal, 198, 201, 217, 218, head, 205
221 root, 52, 63, 69, 105, 130, 134,
WH, 150, 156, 158, 187, 214{ 137, 196, 219, 231, 234, 235,
216 237, 296, 296
Meyer-Viol, 36, 245, 246, 249, 272 terminal, 76, 218
Milward, 12, 13, 19, 33, 99 top, 52, 56, 122, 136, 235, 239,
Minimal Recursion Semantics, 25, 243, 254, 270, 276, 296
231 unxed, 64, 68, 111, 131, 134,
Minimalism, 99, 165, 254, 258, 260, 137, 156, 157, 174, 197{199,
264 201, 205, 216, 218, 219, 221,
modality, 109, 284 254, 258
action, 308 node construction, 309
dominance, 320 nominal, 115
external, 48{49, 113, 291 construct state, 206, 207
internal, 48{49, 113, 291 non-determinism, 316
LINK, 110 non-persistence, 289, 292, 294, 308
long-distance, 307, 320 noun, 136
tree, 37, 320 common, 115, 116, 201, 207, 231,
underspecied, 296 235, 237{240, 245
model operator
dynamic, 44 epsilon, 35, 115
static, 44 external, 206
Modus Ponens, 57, 82, 271 internal, 206
labelled, 30, 63 static, 41
Mohawk, 146
Monachesi, 219
General Index 343
variable binding, 150, 154, 156, resumptive, 15, 106, 121, 122,
195, 224, 235 124, 126, 129, 131{133, 193,
Ouhalla, 162, 266 194, 201, 203{205, 208, 211{
214, 216, 218{220, 256, 258,
Pafel, 154, 155, 188 297
parallelism, 301; see structural sim- Propositional Dynamic Logic, see PDL
ilarity model
parser, 3, 32, 262, 307
parsing, 5, 6, 25, 75, 254, 258, 272, quantication, 115, 141, 257
312 mixed, 227, 232
and linguistic knowledge, 266 quantier, 153
uncertainty, 26, 68 existential, 224
parsing model, 270 generalized, 223, 225, 235
Partee, 6, 34 indenite, 224
partial movement, 163{169, 185 non-indenite, 225
PDL model, 268 universal, 235
PT , 269 quantier binding, 233, 242
PDL states, 270 quantier evaluation, 241
performance, 267 quantier storage, 225, 240, 257
Perrett, 34, 256 question
Perry, 5, 9, 34 functional, 154
persistence, 289 wh, 216
Pesetsky, 120, 223
pied-piping, 108, 113, 128, 166, 203, Radford, 219
208 Reinhart, 152, 215, 222, 225, 226,
Pinkal, 25, 26 228
Poesio, 26 Reis, 168, 176, 181
pointer, 51, 56, 79, 134, 137, 269 relative clause, 1{149
movement, 82, 88, 309, 311 complementizer-less, 119
Pollard, 2, 120, 148, 150, 154, 225 head-nal, 133{139, 256
P OSS , 144{148 head-internal, 139{142, 250, 256,
Postal, 190, 192, 195, 258 258
Prawitz, 15, 245 headless, 120
predicate logic, 247 non-restrictive, 107, 108, 140,
extension, 247 196
Prediction, 57, 65, 81, 83, 264 restrictive, 107, 115, 119, 134,
procedural, 268 201
structure, 307{317 relevance, 8, 19, 26, 212, 216, 263
pronoun, 9{16, 121, 255 representationalist, 7{8, 18, 221, 253,
bound variable, 243 255, 261
clitic, 100, 122, 125, 128, 207, requirement, 42{49, 55, 76, 82, 125,
214, 218{220, 258 254, 270{273, 303, 309
expletive, 125, 126, 128, 230 axiomatization, 320
lexical action, 61, 96 formula, 124, 126, 128, 136, 258
relative, 112, 116, 119, 125, 128, relative clause, 196, 197, 204,
132, 194, 197, 209 205, 211, 219
344 General Index
relative-clause, 256 semantics

scope, 239, 240, 242, 250 denotational, 260
wh, 169 procedural, 313
requirement function, 295; see suc- Shlonsky, 16, 126, 204
cess function Silvers, 7
restrictor, 35, 115, 207, 232, 237, Simpson, 160, 165, 167, 179, 185
240, 245, 248 situation, 14
extension, 249 Situation Semantics, 5, 9, 12, 13
Reuland, 222 Spanish, 129
rewrite relation, 78, 314 Speas, 67, 255
Reyle, 5, 9, 99 Sperber, 8, 16, 26, 95, 109, 212, 216,
Rizzi, 5 263
Roberts, 250 Sportiche, 100
Rochemont, 212 Stalnaker, 12
Romance, 108, 203, 219 Steedman, 19, 33, 99, 150
Romanian, 125, 128, 129, 132, 133, Stokhof, 9, 10, 12, 151, 153, 154
214, 219, 220 Stowell, 192, 201
Rooth, 14, 212 stress
Ross, 193, 217 contrastive, 210, 212
Russo, 34 structural similarity, 301, 302
Substitution, 62, 78, 95, 96, 139, 197,
Sar, 108, 193, 194 198, 207, 217, 231, 237, 240,
Sag, 2, 119, 120, 148, 150, 223, 224, 257, 260, 280, 307
226 success function, 55, 292
Scanning, 58 Suner, 126, 129
scope Swedish, 15
lexical constraints, 257 Szabolcsi, 224
lexical variation, 244
nuclear, 238 T -morphism, 274{276
predicate, 239 T -structure, see tree structure
quantier, 153, 187, 224, 225, task state, 76
232, 257 tau term, 246, 248
indenite, 224{231 temporal index, 98, 233
truth denition, 233 tense, 8, 11, 98
wh expression, 152{156, 256 term-operator, 115, 232
scope action Thinning, 81, 84
indenite, 241 topic, 15, 130{132
non-indenite, 241 transitions, 307
scope relation, 36, 227, 231, 233, tree
234, 242, 244 binary, 274, 296, 317
scope sequence, 242 complete, 272, 277, 278, 290
scope statement, 36, 227, 231, 232, decorated, 274
234, 239{248, 257 nite, 290, 317
complete, 240 frontier, 270, 276, 300
indenite, 240 linked, 274, 275
requirement, 231 pointed, 278
General Index 345
complete maximal, 290 linked basic LBTR, 110
decorated partial, 272 partial, 41, 64, 253, 259
global, 110, 275 decorated, 77
local, 110, 126, 204, 275 linked, 110
minimal element, 299 T -structures, 274
partial PT, 268, 269, 272, 275, type label, 29
302, 314, 317 type normalization, 280
axiomatization, 319 type predicate
decorated, 270, 274, 308 Ty, 32
normal form, 276 type reduction, 280
pointed PPT, 269, 275, 278,
281, 282, 290 unacceptability, 203, 210, 213, 257
tree address language underspecication, 3{6, 260, 265
ADR, 53 lexical, 27, 255
Tree Adjoining Grammar: TAG, 85, semantic, 259
146 structural, 22, 27, 49, 68, 84,
tree atom, 299 85, 205, 254, 255, 258, 259,
tree description, 51{54, 79, 259, 302 272, 276, 296, 307
language, 24, 265 type, 255
partial, 21, 259 ungrammaticality, 133
value sharing, 282 uniqueness, 12, 237
tree domain update, 324
TrDom, 39
tree extension, 275; see tree growth van Deemter, 26, 40
tree growth, 138, 146, 254, 256, 259, van der Does, 10, 12, 14, 249
272, 307 van Eijck, 11, 12, 34, 99
goal-directed, 21 variable, 232
incremental, 79, 272 bound, 10, 35, 224, 234, 242
node decorations, 254 fresh, 238
node relations, 254 nominal, 115, 141, 201, 235, 237,
tree modality, 36{41 240, 245
tree model, 292, 295, 297, 299, 301 temporal, 227, 233, 235
tree morphism, 298, 301 variable binding term-operator, 35,
tree node 235, 251; see term-operator
update, 63 variable nominal, 207
tree node address, 296 verb, 78, 92, 254, 264
tree node identier, see tree node Arabic, 95
predicate auxiliary, 98
language, 303 German, 258, 265
tree node predicate, 296, 303 Japanese, 69, 95
Tn, 51 Verb Adjunction, 172
tree relation, see dominance verb-nal languages, see Japanese
TrRel, 39 Watanabe, 140, 141
tree structure well-formedness, 9, 79, 203, 212, 221,
basic, 39{44 255, 261{264, 315
feature-decorated, 273, 291
346 General Index
Wexler, 148 Wilson, 8, 16, 26, 95, 109, 212, 216,

wh expression, 119 263
wh question, 150{189, 223 Winter, 152, 228
lexical action, 94 witness, 10
wh expletive, see expletive word order, 79, 254
wh-in-situ, 119, 150, 152, 155{
162, 223, 256 Yoo, 225
wh-initial, 157, 161, 169, 256
wh-scope, see scope Zaenen, 15, 100, 254
Williams, 191 Zribi-Hertz, 12
Symbol Index
In this index only the deni- merg(ND1 ; ND2 ), 90
tional occurrences of the item MV, 38
in question are indicated.
P BT , 278
ABORT, 90, 309, 321 P P T , 275
ACT, 77, 79, 309 PPT , 77, 275
[], 308 PPT M, 292
hi, 308 hPPT ; (Z; z )i, 281
ADR, 303 P T , 268
PT , 268
B = hB; <B i, 232 PTR, 77
put(), 90
hDi, 110
DF o, 38 Rf , 279
D, 39
DLa , 38 SC (), 239
DT y , 38 ST RU (L), 315
i
du, 32, 38 SUBST (), 97

, 232, 246 , 232, 246
T n, 50, 296
F , 293, 320 T , 39
FA, 288 T , 274
F BT , 283 T n(0), 52, 318
Fe, 279 T n(a0), 318
F S , 280 T n(a1), 318
F o, 33, 322 T n(aL), 318
hF ; D; !i, 279 T+ , 269
G, 293, 320
T , 269, 272
go(#), 90, 309
T n, 275
T y, 32, 322
Indef, 235 hU i, 110
LF , 284 Va , 268, 308
L(L; C; P ), 78 VAR, 38
[LF ], 285, 288, 319 V? , 270, 291
hLF i, 285, 288, 319 V?C , 295
hLi, 110 Vp , 268
hL 1i, 110
LFV? , 292 ^, 38, 285
make(#), 90, 309
_, 38, 285
merg(#), 90, 309
7 , 285
!
!, 46, 286
348 Symbol Index
8x, 285 u, 300, 301

9x, 38, 285 tF , 301
(x), 285 uF , 301
4, 285 <, 232
s, 285 , 43
:C , 315 F , 282
:, 45, 47, 286 La , 44
2, 285 L , 312
i
, 42 T , 275
#, 48, 291 h(x); ; 0 i, 310
# , 48, 291 1, 90, 309
[#], 37 j=, 32, 39, 46, 48, 287
[#0], 37, 317 }, 51, 56, 79
[#1], 37, 317 , 39, 274
[#], 37, 318 D , 110, 274
["], 37, 317 # , 274
["], 37, 318 1 , 39, 274
h#1 i, 37, 317 F , 279
h#0 i, 37, 317 L , 110, 274
h#1 i1, 96 0 , 39, 274
h#0 1 i, 37 , 39, 274
h"0 1 i, 37 1 , 96
h#1 1 i, 37 ?, 38, 42, 270, 293
h"1 i, 37 ?;, 38, 293, 320
h# 1 i, 37 !C , 323
h" 1 i, 37 )LCP , 78, 314
h# 1 i, 37 >, 38, 285
t, 299{301 ?, 38, 285

Dynamic Syntax

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dynamic Syntax

Uploaded by

Copyright:

Available Formats

Contents

8 Re ections on Language Design 253

of building up interpretation proved fruitful in addressing phenomena which

completion of the book were supported by the Leverhulme Foundation's award

Towards a Syntactic Model

By analogy with formal languages, a natural language may be characterized

(1.1) John upset Mary.

(a) ? (b) P AST :?

John ? John Upset(Mary )

Upset Mary Upset Mary

of the goal { to derive a propositional formula as decoration for that node on

there is a systematic gap between information integral to the natural language

characterized as multiply ambiguous. The challenge, then, for a semantic ac-

third an E-type interpretation. And similarly demonstrative pronouns and tense

In others, they serve as a means to enable gap devices to be available in those

The General Framework

xy [Admire(x)(y )] Mary

Figure 2.1: Development of an unreduced lambda term

this left-to-right process directly projects model-theoretic content. Rather, we

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

Figure 2.2: Evaluation of the term of gure 2.1

xy [Admire(x)(y )] Mary

Figure 2.3: Development of an unreduced lambda term with requirements

to be replaced by an actual term. In all such cases, the interpretation given

Mary John ?(e ! t)

Mary John ?(e ! t)

xy [Admire(x)(y )] Mary

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

This string is of course the simplest possible example of a long-distance

h: : : e ! (e ! t)i : xy[Admire(x)(y)] Mary

Basic Tree Declarative Units

Figure 2.4: Basic tree structures

[h:::ei : John  ] [  T y(e ! t)]

[  T y(t); h#1 i(T y(e ! t))]

Figure 2.5: Partial tree structures

Partial Tree Structures are connected by the relation of development or exten-

[ T y(t); h#1 i(T y(e ! t))] h: : : ti : Admire(Mary)(John)

fh" iT n(a); F o(John); T y(e); [#]?g fh"1 iT n(a); ?T y(e ! t)g

Figure 2.6: Tree developments and tree descriptions

DT y = fe; t; cn; e ! t; : : : g and the Formula predicate, which takes values in

The Type Predicate T y

fF o(); T y(X )g fF o( ); T y(X ! Y )g

There are several idiosyncratic features of this representation language. The

formula S : Admire(; y; W oman(y))(; x; Man(x)) and second the statement

[h: : : ei : ((; x; Man(x)) ] [h: : : (e ! t)i : Admire(; y; W oman(y))]

[h: : : (e ! (e ! t))i : Admire] [h: : : ei : (; y; W oman(y))]

Figure 2.7: Projecting quanti ed terms, with scope-dependency marking

h: : : ei : John h: : : (e ! t)i : Admire(Mary)

2 ; h"0 i1 3 ; h#i4 ; h#i5

4 ; h" ih#0 i2 5 ; h"ih#0 i4

Figure 2.8: Interpretation of the existential tree modalities

2.2.2.1 De nition of the Language

De nition 1 (The Language DU9 )

A proposition A of the language DU is de ned by

predicate has a value. And a meta-variable just won't do as a proper value.

T ; n j= h" i(Cat(+Q) ^ T y(t)) `At, or somewhere above, n is a

2.2.4 Partial Tree Structures

[h:::ei : John  ] [  T y(e ! t)]

 structure T 0 can be constructed from T by adding nodes, tree relations

an update relation on the labels, where l La l0 if l0 `is more informative'

xy [Admire(x)(y )] Mary

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

xy [Admire(x)(y )] Mary

xy [Admire(x)(y )] Mary

John y [Admire(Mary )(y )]

xy [Admire(x)(y )] Mary

h: : : e ! (e ! t)i : xy[Admire(x)(y)] Mary

[h:::ei : John ] [ T y(e ! t)]

[ T y(t); h#1 i(T y(e ! t))]

[ T y(t); h#1 i(T y(e ! t))] h: : : ti : Admire(Mary)(John)

fF o(); T y(X )g fF o( ); T y(X ! Y )g

formula S : Admire(; y; W oman(y))(; x; Man(x)) and second the statement

[h: : : ei : ((; x; Man(x)) ] [h: : : (e ! t)i : Admire(; y; W oman(y))]

[h: : : (e ! (e ! t))i : Admire] [h: : : ei : (; y; W oman(y))]

Figure 2.7: Projecting quantied terms, with scope-dependency marking

2 ; h"0 i1 3 ; h#i4 ; h#i5

4 ; h" ih#0 i2 5 ; h"ih#0 i4

2.2.2.1 Denition of the Language

Denition 1 (The Language DU9 )

A proposition A of the language DU is dened by

[h:::ei : John ] [ T y(e ! t)]

structure T 0 can be constructed from T by adding nodes, tree relations

an update relation on the labels, where l La l0 if l0 `is more informative'

incremental change. If the label li is dened, then it can be one of two

every requirement in R is either a member of R0 (it still is required)

The models we are going to dene will be essentially dynamic, in that they

= ! and for all T 0 ; n0 : T ; n T 0 ; n0

is, for atomic , if T ; n j= and T ; n T 0n0, then also T 0 ; n0 j= . The

[h: : : ; ei : Mary ] [h: : : ; ei : John ] [h: : : i : T y(e ! t)]

[h: : : ; e ! (e ! t)i : [h: : : i : T y(e)]

fF o(H )g fF o((; x; R(x)))g f?T y(t)g f?T y(t); }g

3.2 The Parsing Process Dened

John arg C John

lated both as a parser and as a generator, denes a mapping between a string

?F o(John) on the unxed node in the newly induced LINK ed structure. In

fF o(P (the P )); f?T y(cn)g

fF o(U); T y(e)g fTFy(oe(Y (Y; Man(Y )));

fF o(U)g fF o(X (X; Man(X )))g

fF o(U)g fF o(X (X; Man(X )))g