Language Change Variation and Universals Culicover Full Chapter

Language Change, Variation, and
Universals Culicover
Visit to download the full and correct content document:
https://ebookmass.com/product/language-change-variation-and-universals-culicover/
OUP CORRECTED PROOF – FINAL, 20/6/2021, SPi
Language Change, Variation, and Universals

Language Change,
Variation, and Universals
A Constructional Approach
P E T E R W. C U L IC OV E R
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Peter W. Culicover 2021
The moral rights of the author have been asserted
First Edition published in 2021
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2021931013
ISBN 978–0–19–886539–1
DOI: 10.1093/oso/9780198865391.001.0001
Printed and bound in Great Britain by
Clays Ltd, Elcograf S.p.A.
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
Acknowledgments ix
Preface xi
List of Abbreviations xv
PA R T I. F O U N D A T I O N S
1. Overview 3
1.1 The problem 3
1.2 Constructions 9
1.2.1 Basics 9
1.2.2 Constructions are not derivations 12
1.3 Antecedents 14
2. Constructions 16
2.1 Introduction 16
2.2 What a grammar is for 16
2.3 A framework for constructions 19
2.3.1 Representing constructions 19
2.3.2 Licensing 27
2.3.3 Linear order 28
2.4 Appendix: Formalizing constructions 31
2.4.1 Representations on tiers 31
2.4.2 Connections between tiers 36
2.4.3 Licensing via instantiation 36
3. Universals 41
3.1 Classical Universal Grammar 41
3.1.1 Core grammar 42
3.1.2 Parameters 44
3.1.3 UG and emerging grammars 46
3.2 Another conception of universals 50
3.3 On the notion ‘possible human language’ 53
3.3.1 Possible constructions 53
3.3.2 An example: Negation 56
3.3.3 Another example: The imperative 61
3.4 Against uniformity 66
4. Learning, complexity, and competition 68
4.1 Acquiring constructions 68
vi contents
4.2 Constructional innovation 75

4.3 Constructions in competition 77
4.3.1 Multiple grammars vs. multiple constructions 78
4.3.2 Defining competition 81
4.3.3 When do we actually have competition? 86
4.4 Economy 87
4.4.1 Representational complexity 88
4.4.2 Computational complexity 90
4.4.3 Interpretive complexity 96
4.5 Simulating competition 100
4.6 Summary 106
P A R T I I . VA R I A T I O N
5. Argument structure 111
5.1 Introduction 111
5.2 Argument structure constructions (ASCs) 112
5.2.1 Devices 112
5.2.2 CS features 118
5.3 Differential marking 120
5.3.1 Differential subject marking 120
5.3.2 Differential object marking 130
5.4 Modeling differential marking 133
5.4.1 Acquisition of ASCs 134
5.4.2 Simulation 139
5.5 Summary 144
6. Grammatical functions 145
6.2 The notion of ‘subject’ 146
6.3 Morphologically rich ASCs 147
6.3.1 Plains Cree argument structure 148
6.3.2 Incorporation 152
6.3.3 Complexity in ASCs 156
6.4 Split intransitive 158
6.5 The emergence of grammatical functions 160
6.6 Summary 165
7. Aᆣ constructions 166
7.1 Foundations 166
7.2 Doing Aᆣ work 169
7.2.1 Gaps and chains 169
7.2.2 Relatives 174
7.2.3 Topicalization 175
contents vii
7.3 Scope in situ 177

7.3.1 Wh-in-situ 178
7.3.2 In situ in polysynthesis 180
7.3.3 Other in situ 182
7.3.4 Cryptoconstructional in situ 183
7.4 Extensions of Aᆣ constructions 183
7.5 Toward an Aᆣ constructional typology 189
7.6 Summary 194
PA R T I I I. C H A N G E
8. Constructional change in Germanic 197
8.2 Basic clausal constructions of Modern German 198
8.2.1 Initial position in the clause 200
8.2.2 Position of the finite verb in the main clause 202
8.2.3 Position of the verb in a subordinate clause 203
8.2.4 Position of the verb in questions 203
8.3 The development of English 204
8.3.1 The position of the verb 205
8.3.2 The ‘loss’ of V2 in English 208
8.3.3 The loss of case marking 213
8.4 The development of Modern German from Old High German 215
8.5 Verb clusters 219
8.6 Conclusion 223
9. Changes outside of the CCore 225
9.1 English reflexives 225
9.1.1 Reflexivity in constructions 225
9.1.2 Variation and change in reflexive constructions 227
9.2 Auxiliary do 230
9.2.1 The emergence of do 230
9.2.2 The spread of do 234
9.3 Preposition stranding 235
9.3.1 Why p-stranding? 235
9.3.2 P-passive 237
9.3.3 Coercion 239
9.4 Conclusion 241
10. Constructional economy and analogy 242
10.1 The elements of style 244
10.2 Analogy 249
10.2.1 Maximizing economy 250
10.2.2 Routines 252
10.2.3 Pure style 257
viii contents
10.3 Beyond parameters: Capturing the style 262

10.3.1 Baker’s Polysynthesis Parameter 262
10.3.2 Greenberg’s universals 264
10.3.3 Non-Greenbergian universals 267
10.4 Summary 272
11. Recapitulation and prospects 274
References 279
Language Index 309
Author Index 311
Subject Index 316
Acknowledgments
This book has been a long time in the making, and has been profoundly
influenced by many people. As always, thanks to Ray Jackendoff and Susanne
Winkler for their friendship, collegiality, advice, and support. Jack Hawkins
read several early versions of the manuscript and generously shared his
insights—they can be seen throughout. I am very grateful to Jefferson Barlew
for the formal description of the constructional framework developed as
part of our collaboration on minimal constructions, which appears in the
Appendix to Chapter 2 and is based on Barlew & Culicover (2015). And I owe
a tremendous debt to Giuseppe Varaschin, who read the entire manuscript
in various incarnations and made countless detailed and constructive
suggestions, virtually all of which have been incorporated into the current
version.
Thanks also to Brian Joseph and Noah Diewald, from whom I learned so
much in the course of our discussions in our Cree Reading Group, to Greg
Carlson, Ashwini Deo, Adele Goldberg, Björn Köhnlein, Andrew McInnerney,
Rafaela Miliorini, Lorena Sainz-Maza Lecanda, Richard Samuels, Yourda-
nis Sedaris, Andrea Sims, Shane Steinert-Threlkeld, Elena Vaiksnoraite, and
Joshua Wampler for stimulating discussions on a range of topics, to Philip
Miller for helpful comments on the material in Chapter 3 and for his general
perspective on constructional approaches to grammar, to Afra Alishahi for
our collaboration on the simulation of change in argument structure con-
structions, to Andrzej Nowak for our collaboration on language change, and
to Marianne Mithun for helpful insight into active-stative languages. Thanks
to Zoe Edmiston, whose interest in Plains Cree stimulated my own, and to
Morten Christiansen and Nick Chater for giving me the opportunity to write
the foreword to their recent book and to think freshly about the foundations
of linguistic theory.
I am especially indebted to several anonymous reviewers, whose construc-
tive suggestions have pointed to a rethinking of this book in ways that have
led to significant improvements. Of course, any errors and deficiencies that
remain are my responsibility alone.
The Department of Linguistics and The Ohio State University awarded me
a Special Assignment in the Autumn of 2016, which made it possible for me to
x acknowledgments
make significant progress on the first draft. My thanks to Shari Speer, for her
constant encouragement and support.
I owe a deep debt to the work of Ivan Sag and Partha Niyogi, and I wish
I could thank them now—sadly, they left us far too soon, with so much left
for us to do without their guidance and insights. I am grateful as well to
the late John Davey of Oxford University Press, whose unflagging support
and encouragement over many years helped me through the publication
of five substantial volumes, each of which contributes significantly to the
foundations of the present work. And it is with great regret and sadness that
I am unable to share this work with my dear friend and colleague, the late
Michael Rochemont.
Finally, deepest thanks and much love as always to Diane, Daniel, and Jen
for always being there.
Preface
This book began with a nagging worry. By and large, the grammatical literature
assumes that grammatical functions such as subject and object are universal
and independently represented in syntax, and play an integral role in the
description of the form/meaning correspondences that comprise what we call
‘language’. But these supposed universals, like many others, pose several mys-
teries. Where do they come from? Are they part of the biological endowment
for language that is encoded in our genes? Is there a device called Universal
Grammar in our brains that incorporates such notions as subject and object, or
the equivalent? Did biological evolution select for languages whose grammars
make use of these grammatical functions?
As I looked more into the literature on grammatical functions, it became
clear that they are not universal—not all languages appear to make use of
them, and where it appears that they are used, they are not the same cross-
linguistically. Of course, we find the terms ‘subject’ and ‘object’ used all the
time to distinguish the participants in a relation such as ‘The tiger bit Sandy’ in
a given language. And it is possible to stipulate that ‘the tiger’ in some language
has the syntactic representation of what we call ‘subject’ in a language like
English. But on closer investigation, often what is being distinguished are the
phrases that denote entities with thematic roles like agent and patient.
It is often said that “A journey of a thousand miles begins with a single
step”, and each journey starts from a different place. This is one such journey.
No matter where we start from in syntactic theory, the interconnections take
us to places that we did not envision at the outset. In this case, as I thought
about grammatical functions and how arguments are distinguished cross-
linguistically, I found myself engaged in something much more far-reaching
and ambitious: the explanation of language change, variation, and typology.
Why does change proceed in certain directions, why do we get the variation
that we do, why do we get certain variants and not others? How do grammars
carry out the task of encoding the expressive functions of language, and what,
if any, are the limits on how this is done? Why are certain patterns ubiquitous,
across languages and in a single language, while others are rare or non-
existent?
xii preface
It is of course possible to formulate descriptions of change, variation, and

typological patterns in any reasonably explicit descriptive framework. But in
order to properly explain them, we need the right descriptive framework. I had
been working on constructional phenomena for some time, beginning in fact
with my dissertation (Culicover 1971), more recently with the publication of
my books Syntactic Nuts (Culicover 1999), Grammar and Complexity (Culi-
cover 2013c), and Explaining Syntax (Culicover 2013b), and in my collabora-
tion with Ray Jackendoff on Simpler Syntax (Culicover & Jackendoff 2005). It
seemed promising to pursue a constructional approach to these issues.
So in the end, this book is not about grammatical functions, although that
is one strand. It is about how and why grammars vary and change, and hence
why there are distinct languages, understood as overt expressions of different
grammars. To address these questions, I argue for a particular constructional
approach to the representation of grammatical knowledge, and I seek to
show how this approach helps us understand how different languages and
typological patterns might arise out of grammatical change and competition
between grammars in the natural social and cognitive environment.
The organization of the book is as follows. Part I lays out the foundations
of this exploration. It covers a statement of the problem, a reassessment of the
notion of Universal Grammar, a theory of constructions, and the conceptual
relationships between syntactic theory, grammatical variation, and grammat-
ical change. This part of the book consists of four chapters.
Chapter 1 (Overview) lays out the general problem of explaining the form
of grammars, and relates this problem to that of characterizing grammatical
complexity. Following much recent work, I take the view that reduction of
complexity is a driving explanatory force whose effects can be seen in change
and in variation. In the overview, I sketch out the general perspective that
I adopt on universals, conceptual structure, constructions, complexity and
change and variation, and how they are related.
Chapter 2 (Constructions) sets out an approach to grammatical description
in which the notions of Chapter 1 can be formally implemented. The account
is a constructional one. I outline the formalism, show how it is used to account
for grammatical phenomena, and highlight its utility in describing variation
and change.
Chapter 3 (Universals) reviews the approach to universals in contemporary
grammatical theory, which is that they are expressions of Universal Grammar
(UG), the human faculty of language. In practice UG is assumed to constrain
syntax, and thus constitutes an explicit statement of grammatical universals.
This chapter formulates a different view, which is that what is universal is
preface xiii
conceptual structure, and grammatical universals and typological patterns

arise as a consequence of pressures to formulate constructional grammars
that express conceptual structure as simply as possible. While it is possible
to characterize change and variation in any sufficiently expressive descriptive
framework, I argue that the constructional approach provides a natural frame-
work for explaining language variation and change.
Chapter 4 (Learning, complexity, and competition). This chapter explores
how envisioning a language learner as acquiring a grammar consisting of
constructions allows us to account for change and variation. I develop the idea
that change is not solely the responsibility of early language learners, but may
also occur as innovations initiated by adult speakers. One key explanatory
component is constructional complexity; another is competition between
constructions that have overlapping functions.
In Part II I look at a number of cases of variation in constructions that
deal with two core expressive functions of a language: argument structure
and Aᆣ /filler-gap constructions. I show how the constructional framework
provides a formal apparatus that is suitable both for describing the phenomena
in a given language and for accounting for the observed variation. This part of
the book consists of three chapters.
Chapter 5 (Argument structure) applies the theory to variation in systems
for expressing argument structure. A central point is that there are multiple
devices of comparable complexity that encode the thematic roles; hence it is
not necessary to assume that all languages share a uniform syntactic structure
at some abstract level of representation.
In Chapter 6 (Grammatical functions) I return to the question that triggered
this project, the source of grammatical functions (GFs). I review evidence
that not all languages require GFs, and show how to capture the relevant
correspondences between form and meaning in constructional terms.
Chapter 7 (Aᆣ constructions) applies the theory of the preceding chapters to
Aᆣ constructions, such as wh-questions and relative clauses. The main result
of this chapter is that there is a range of ways in which the conceptual ‘work’
associated with these constructions can be expressed in the correspondence
between syntax, phonology, and meaning. None of them involve ‘movement’
in the classical sense, although constructions can express links between con-
stituents not in canonical position relative to their governing heads, giving the
illusion of movement.
Part III applies the models of constructional learning and network inter-
actions developed in Chapter 3 to establish the plausibility of the account of
change sketched out in Part I. This part of the book consists of three chapters.
xiv preface
Chapter 8 (Constructional change in Germanic) tracks several of the major

changes in English and German word order and accounts for them in terms of
constructional change as formulated in Chapter 3. It argues that the changes
in Germanic are relatively simple in constructional terms, although the super-
ficial results are quite dramatic. Among the topics addressed are clause-initial
position, V2, VP-initial and VP-final verb position, the loss of V2 and case
marking in English, and verb clusters in Continental West Germanic.
Chapter 9 (Changes outside of the CCore) shows the broader applicability
of the constructional approach. I look at three well-documented developments
in English that do not fall into the category of ‘core phenomena’ as understood
in Chapter 3, reflexivity, do support, and preposition stranding. These changes
are not as central to the expressive function of language as argument struc-
ture, operator/scope, and similar phenomena. I argue that these phenomena
provide additional evidence that the constructional approach is well-suited for
providing genuine explanations for language change and variation.
In Chapter 10 (Constructional economy and analogy), I look more deeply
into what constitutes constructional economy, and why it plays a role in
shaping the form of grammars. I argue that constructional economy is the
consequence of what has been called ‘analogy’ in the traditional linguistics
literature. Specifically, I suggest that economy in constructions derives from
placing a high value on the use and reuse of the components of the processing
routines associated with constructional correspondences. I apply this general
idea to seek explanations for a range of typological patterns that I refer to
generally as ‘style’.
Chapter 11 (Recapitulation and prospects) summarizes the main results of
this book and lays out some general propositions about how to think further
about language variation and change from the perspective of constructional
grammars.
List of Abbreviations
Adj adjective
ASC Argument structure construction
AUX Auxiliary
BDT Branching Direction Theory
CCore Conceptual Core
CS Conceptual Structure
CWG Continental West Germanic
DM Distributed Morphology
DOM Differential object marking
DSM Differential subject marking
GB (theory) Government Binding theory
GF grammatical function
HPSG Head-driven Phrase Structure Grammar
IPP Infinitivus Pro Participio
IS information structure
LFG Lexical Functional Grammar
LID Lexeme Identifier
MGG Mainstream Generative Grammar
MOD Modal
ModE Modern English
ModG Modern German
N noun
Neg Negative
NP noun phrase
OE Old English
OHG Old High German
P&P Principles and Parameters Theory
PA Parallel Architecture
PG Proto-Germanic
PLD primary linguistic data
PP prepositional phrase
RG Relational Grammar
SAI subject Aux inversion
SD Standard Dutch
UG Universal Grammar
V verb
xvi list of abbreviations
VP Verb Phrase
VPR Verb projection raising
WALS World Atlas of Language Structures
WF West Flemish
ZT Zürich German
PART I
F OU N DAT ION S
1
Overview
1.1 The problem
My primary concern in this book is how human languages get to be the way
they are, why they are different from one another in certain ways and not in
others, and why they change in the ways that they do. Given that language
is a universal creation of the human mind, the central question is why there
are different languages at all. Why don’t we all speak the same language?
This chapter lays out the general foundations of inquiry into this question in
contemporary linguistic theory, and the specific assumptions that inform the
answers developed in this book.
I call this central question ‘Chomsky’s Problem’.1 Chomsky’s own answer,
hinted at in Chomsky (1965) and further developed in Chomsky (1973, 1981)
and other work, has been that in a sense we do all speak the ‘same language’.
What we produce is the external manifestation of a universal, biologically
determined, abstract faculty of the human mind, called Universal Grammar
(UG). This classical Chomskyan account, which I refer to throughout
as Mainstream Generative Grammar (MGG), has the following main
components:
(i) There is a set of very general grammatical principles, structures, and

mechanisms, UG, which define the core grammar shared by all
languages.
(ii) These principles and mechanisms are biological universals and consti-
tute I-language.
(iii) Some observable variation is due to differences in the values of core
parameters; hence I-language takes different forms as determined by
these parameters. The parameters are set by learners on the basis of
exposure to primary linguistic data (PLD).
(iv) The set of actual sentences and their meanings, produced by a group
of speakers, referred to as E-language, is itself of no theoretical
1 The parallel with Chomsky’s “Orwell’s Problem” and “Plato’s Problem” is intended.
Language Change, Variation, and Universals: A Constructional Approach. Peter W. Culicover, Oxford University Press.
© Peter W. Culicover 2021. DOI: 10.1093/oso/9780198865391.003.0001
4 overview
significance, except insofar as it counts as evidence about I-language,

UG, and the parameters of variation.
(v) Phenomena that are outside of UG are in the periphery. The periph-
ery contains idiosyncrasies, irregularity, and exceptions. It can vary
widely from language to language, although not without principled
constraints.
(vi) The principles and mechanisms of UG constitute an optimal
solution to the problem of expressing thought.
Frameworks that fall under the general perspective of (i)–(v) are Govern-
ment/Binding Theory (Chomsky 1981), Principles and Parameters Theory
(Chomsky 1981; Chomsky & Lasnik 1993) and, with (vi), the Minimalist
Program (Chomsky 1995b, 2000a).
I assume that any solution to Chomsky’s Problem must have the general
structure of (i)–(v), but only in the following sense.
(vii) It must explain why languages share so many properties, both in form
and in function.
(viii) It must attribute these properties to some universal source—
biological, cognitive, or social, or a combination of these.
(ix) It must account for the possibility of variation and for the range of
variation.
(x) It must accommodate not only regularities and generalizations, but
idiosyncrasies, irregularity, and exceptions.
(xi) It must explain why certain properties are common while others are
rare or do not occur at all.
(xii) It must explain how learners arrive at mental representations of
language—that is, grammars—that are suitably close to but not
necessarily identical to the representation(s) of members of the
linguistic community that they are learning from.
The solutions that I propose in this book are inspired by Chomsky’s Problem
and the Chomskyan program of MGG, and play off of them, but they are in
many important respects very different, both in spirit and in substance. Cru-
cially, Chomsky’s approach has been to assume point (vi), i.e. that “language
design may really be optimal in some respects, approaching a ‘perfect solution’
to minimal design specifications” (Chomsky 2000a, 93). Chomsky’s notion of
optimality is based on an abstract notion of economy, one that is not linked in
any straightforward way to the cognitive capacities and limitations of human
beings (Johnson & Lappin 1997, 1999).
1.1 the problem 5
In contrast, I assume that grammars are not computationally optimal in

some abstract sense, but reflect the outcome of the neo-Darwinian evolution of
a complex cognitive system (Ladd et al. 2008; Kinsella 2009; Kinsella & Marcus
2009). The evolution of this system is driven, at least in part, by the pressure
to reduce the complexity of the mental representation of grammatical com-
petence and associated computational complexity along specific dimensions.
I use the term economy to refer precisely to this pressure.
The approach that I argue for here has the following general structure.
(i) Regarding the core, I assume that it is grounded in conceptual struc-

ture, not in a set of formal constraints on grammars. The basic job of
language is to express thought.2 I use the term CCore (for Conceptual
Core) here to refer to the set of expressive functions that are central to
human thought and discourse. A grammar of a language encodes these
functions more or less efficiently and transparently—functions such
as argument structure, thematic structure, interrogation, imperatives,
description, binding, reference and coreference, restrictive modifica-
tion, negation and quantification, discourse structure, and so on.3 The
exact extent of the CCore, and its origins, remain open questions.
(ii) I assume that the expressive functions that languages encode are
cognitive universals. The grammatical devices for expressing them, on
the other hand, are social universals, in the sense that they reside in the
minds of speakers in social networks in virtue of their linguistic com-
petence, and are transmitted socially, not biologically, across cultures
and generations, through contact between individuals and groups of
individuals. These social universals ‘live’ in the social network, that is,
in the linguistic competence of all speakers of all languages across time
and space. They are universals in the sense that they are universally
available for the expressive functions of the CCore. However, no par-
ticular way of expressing a particular function needs to be absolutely
universal in the sense that it is active in the grammar of every possible
language, a point that I return to below.
2 Thus I agree with Chomsky (1972). The expression of thought is logically prior to the communica-
tion of thought, at least thought that corresponds to representations that are constructed combinatorily
out of primitive elements (Fitch 2011). Externalization is subsequent to the construction of thought,
and is manifested at the interface with sound and gesture. This is not to deny, however, that the
expression of thought is central to communication, and that some aspects of linguistic form may be
explained in terms of constraints imposed by the communicative task. Chomsky (2005) seems to accept
this view in his citation of ‘third factor’ explanations, which comprise communicative efficiency among
other things.
3 Chomsky’s focus in the Minimalist Program on minimal mechanisms for expressing argument
structure and extraction (external and internal binary Merge) is arguably a very restricted variant of the
approach that I am taking here, taking syntax as a proxy for a limited portion of conceptual structure,
and setting aside most conceptual and grammatical phenomena.
6 overview
(iii) Regarding variation, I assume that anything that can possibly be

expressed as a correspondence between sound and meaning is,
in principle, possible in language. Hence in principle, variation is
unbounded. However, economy, that is, the pressures on grammars to
reduce complexity, leads to a significant winnowing of the logical
possibilities, in the spirit of markedness (Chomsky 1965, chapter
1). We expect that other things being equal, the simplest ways of
expressing the CCore will be most widespread in social networks and
perhaps even completely universal in some cases.⁴
The challenge posed by Chomsky’s Problem is a fundamental one. How is

it possible to have a restrictive theory of the human language capacity that
nevertheless allows for the massive superficial variation and idiosyncrasy that
is attested in the world’s languages? Even if we entertain Chomsky’s view of a
highly restrictive UG, we must account for the variation—simply banishing it
to the periphery will not suffice as an explanation. It is important to always
keep in mind that the full range of grammatical phenomena is acquired by
learners, not just the parametric variation defined over some characterization
of a restricted UG core (Culicover 1999; Culicover & Jackendoff 2005).
Regarding this challenge for the learner, Chomsky (2013, 37) says that
either there is an infinity of options, in which case challenging and perhaps

hopeless abductive problems arise if the task is taken seriously; or there is a
finite number, and the approach falls in principle within P & P [Principles
and Parameters Theory – PWC]. That leaves open many questions as to how
parameters are set, and what role other cognitive processes might play in
setting them.
⁴ Dufter et al. (2009, 12–13) raise a number of important questions about how to explain apparent
constraints on constructional variation:
One such question is whether the factors that influence variant distribution should be an
integral part of the grammar or not. A second theoretical issue that needs to be explored
further is the nature of generalization. Can generalizations about structural properties
within a language be formulated in parallel fashion to cross-linguistic generalizations, using
the same types of inheritance networks? Shouldn’t there be a somewhat different status
accorded to more general typological principles, constraints, or parameters? Is there really
no difference between the modeling of micro-variation and macro-variation? It seems
that construction schemata can be postulated rather ad hoc, such that the limits of what
is possible in language do not follow from the theory (as it is claimed by models of the
Chomskyan tradition).
I believe that the approach to economy that I develop in this book offers a useful way to address such
questions.
1.1 the problem 7
Challenging, yes, but not hopeless. In Culicover (1999), I argued that the
idiosyncrasies that languages have—the so-called periphery—are as robust
as the more general phenomena that have classically been the province of
parameter theory, e.g. word order and movement. They are learned, and
native speakers have clear intuitions about them. At first glance, the periphery
appears to allow unpredictable variation. However, the peripheral phenomena
of a language prove to be related in systematic and often revealing ways to the
more general and regular phenomena. And the latter are by no means always
fully general and fully regular.
The classical view of variation is that it is parametric, in the sense that a
parameter has a finite number of possible values, preferably two, and each
language has a particular setting for each parameter. I argue in the course
of this book that characterizing variation in terms of parameters does not
shed much light on the nature and scope of variation. The approach that I
take is that what is essentially universal is not the inventory of formal devices
that define the syntax of a language, as in MGG, but the CCore, that set of
conceptual structure functions that a language must encode. Expressing the
CCore is the ‘work’ that a grammar does. Syntactic structure is one way of
encoding these functions and organizing them into sounds, morphological
form is another.⁵ As a consequence, the universal architecture of the CCore
is inevitably reflected cross-linguistically in the organization of syntax and
morphology, and economy restricts the range of ways in which this work is
accomplished.
This perspective is guided by a set of ideas and intuitions. Some of these are
shared with MGG and some are not.
(i) Thought and the expression of thought are universal. All humans are
born with the same cognitive apparatus for forming thoughts and the
same drive to express and communicate these thoughts in sound and
gesture.
(ii) Grammar is a distillation of thought. Through this distillation, gram-
mar becomes autonomous, at least to some extent, and its categories
only indirectly and imperfectly correspond to the categories of thought
(see (iii)). By this I mean that the categories and relations in our
conceptual representations are reflected imperfectly, and in a tighter
⁵ While some syntactic theories have sought to reduce morphological form to syntactic derivation
(see Harley & Noyer 1999 for a review), I argue in Chapter 5 that there is no empirical motivation for
such a step.
8 overview
and more restricted way, in the categories and relations that constitute
grammars.⁶ For example, our conceptual categories of time are very
complex—they cover the past, the present, and the future, and times
that precede and follow reference times relative to them and to the
time of utterance (Reichenbach 2012 [1958]). The tense systems of
languages reflect some of the basic distinctions of conceptual time, but
are for the most part simpler than the semantics, and in general do not
map directly into particular times and temporal relationships.
(iii) Syntax is autonomous. Autonomy of syntax means that although there
are correlations between syntactic structure and meaning, to a sig-
nificant extent, syntax is not reducible to meaning, that is, propo-
sitional semantics, information structure, and discourse structure.
For example, there are many distinctions in the thematic roles that
individuals may play in events and states, but few of these distinctions
are grammatically marked (see, for example, Dowty 1991, as well as
Chapter 5).
(iv) Alignment and packaging. Languages may vary in terms of which
aspects of meaning are packaged together in particular morphosyn-
tactic units and corresponding phonological forms. Simple examples
are the expressions enter and go into in English. In the case of enter, the
meaning components go′ and into′ are packaged into a single word,
while in the case of go into they correspond to distinct words. But
packaging can be quite a bit more complex than this (Jackendoff 2002;
Slobin 2004).
These intuitions lead to the characterization of a grammar as consisting of

constructions. The formal description, function, and scope of constructions
are discussed at greater length in subsequent chapters; in this chapter I provide
an informal sketch of the constructional approach to grammatical description
and explanation.
⁶ An early expression of this relationship can be found in Paul (1890, 288):
Every grammatical category is produced on the basis of a psychological one. The former is
originally nothing but the transition of the latter into outward manifestation. As soon as the
agency of the psychological category can be recognised in the use of language, it becomes a
grammatical category. Its agency, however, by no means ends with the creation of the latter.
It is itself independent of language. As it existed before the grammatical category, so it does
not cease to operate when this comes into being. In this way the original harmony between
the two may be in the course of time disturbed. The grammatical category is to some extent
a petrifaction of the psychological.
Instead of Paul’s “petrifaction” I use the term “distillation,” but the basic idea is the same.
1.2 constructions 9
1.2 Constructions
This section summarizes the essential features of the constructional approach

to grammar, and contrasts it with other syntactic theories.
1.2.1 Basics
First, the basic architecture of a constructional approach. I assume, following

Jackendoff (2002) and many others, that the minimal description of a linguistic
object such as a word or a phrase—a construct—involves a description of its
sound (phon), its grammatical structure (syn) and its conceptual structure (or
meaning)(cs). These are the tiers. Culicover & Jackendoff (2005) also assume
that the grammatical functions are represented on a distinct gf tier. cs may
comprise discourse and information structure, or they may be distinct tiers.
I leave the question open here of precisely how to incorporate these aspects of
meaning, as it appears to be a matter of notation, not substance.
Simpler Syntax adopts the view taken in Head-driven Phrase Structure
Grammar (HPSG; Pollard & Sag 1994), Lexical-Functional Grammar (LFG;
Bresnan & Kaplan 1982), and varieties of Categorial Grammar (Jacobson 1992;
Kubota & Levine 2013a,b; Morrill 1995; Oehrle et al. 1988; Pollard 2004; Steed-
man 1993; Uszkoreit 1986; Zeevat et al. 1987) that syntax is monostratal. That
is, there is a single syntactic representation that corresponds to representations
on other levels. From this it follows that there is no ‘movement’ per se. Chains
that relate constituents external to the basic clause and gaps, pronominal
copies, or affixes are products of the correspondence between phon, syn, and
cs, and are not produced by deriving one syntactic structure from another.
A construction in the sense of Culicover & Jackendoff (2005) describes
a relationship between representations on two or more tiers that license the
properties of constructs. The essential difference between construction and
construct is the difference between a description and the objects that satisfy
such descriptions; otherwise they have the same general architecture, and are
composed of more or less the same elements following the same principles.
An important difference is that a construction may contain variables, while
a construct does not. With other constructional grammarians, I assume that
the constructions, including general grammatical constructions, idioms, and
individual lexical items, reside in the extended lexicon, often called the ‘con-
structicon’ (Jurafsky 1996).
10 overview
To illustrate with a concrete example, there is a construction in English

that stipulates that under normal circumstances, the verb in a verb phrase
precedes its arguments and adjuncts. A sequence of words that consists of
a verb and strings that constitute arguments and adjuncts of the verb must
meet this condition. Such a sequence is licensed by the construction for the
English VP.
A sequence of constituents that does not meet this condition is not licensed
by this construction, and is ill-formed as a VP, unless there is another con-
struction that licenses it. So, the VP in (1a) is a well-formed VP of English,
while the VP in (1b) is not. But the VP in (1c) is well-formed, because there is
a special idiomatic construction that licenses it.
(1) a. Chris does not [VP like pepperoni pizza].

b. ∗ Chris does not [VP pepperoni pizza like].
c. One swallow does not [VP a summer make].
Thus, the function of a grammar is to state what properties a construct must

or may have in order to be well-formed in the language. A formalization of this
relationship between construction and construct is given in the Appendix to
Chapter 2. The grammar of a language is the set of constructions that state
the licensing conditions that together define the set of well-formed constructs
in this language.
An important consequence of the constructional approach is that it is
possible for a grammar to contain two constructions that license alternative
forms of the same syntactic structure, or alternative ways of encoding a
particular CS function syntactically. For example, for a VP that consists of
V and NP, a grammar may have one construction that licenses [VP V–NP],
and one that licenses [VP NP–V]. There are languages, including older forms
of English, which show such variation. The flexibility that the constructional
approach affords is what makes it so useful for our understanding of variation
and change.⁷
The descriptions on each tier are representations of the familiar sort—
phonological forms, syntactic structures, and semantic representations. For an
individual word, the construction that defines it is a lexical entry that specifies
its phonological form, its syntactic category, and morphological features, and
its meaning (Jackendoff 2002; Jackendoff & Audring 2020).
I assume for concreteness that for the most part the elements of phon are
determined by the language-specific paradigm function (in the sense of
⁷ It is, of course, possible to recruit other devices to represent such variation, e.g. multiple grammars,
or multiple settings for the same parameter (e.g. Yang 2002). There are no empirical differences, as far
as I can tell, and in the end the choice of representation must be based on what architecture is most
natural.
1.2 constructions 11
Stump 2001). For English I represent this function as ΦEnglish (or just Φ when
it is clear), and so on. For example, the representation of the bare form of the
verb kick is, roughly, the correspondence shown in (2).
(2) ⎡phon ΦEnglish (kick)=/kɪk/ ⎤

⎢syn [V kick] ⎥
⎢ ′
⎥
⎣cs λy.λx.kick (agent:x, patient:y)⎦
Crucially, when the verb has particular morphosyntactic properties in a

sentence, the corresponding forms specified by Φ appear in phon. For exam-
ple, the verb kick has the third person singular present tense inflection in Chris
kicks Fido; the paradigm function ΦEnglish specifies the form as in (3).
(3) ΦEnglish (kick,3,sg,pres) = /kɪks/
For regular inflections, like the English third person singular present, the
paradigm function may be formulated in terms of a regular correspondence,
adding the appropriate realization of -s to the form of the bare verb. For
irregular cases, the form must be stipulated, as in (4).
(4) ΦEnglish (be,3,sg,past) = /wəz/
The precise formal characterization of Φ or an equivalent alternative is not

a central concern here—what we require is simply some way of expressing
the fact that the phonological form in phon reflects the morphosyntactic
properties in syn. Inflectional and paradigmatic complexities make it impos-
sible to always state one-to-one correspondences between features of syn and
phonetic strings in phon. Assuming an appropriate Φ for each language allows
us to set aside the problem of exactly what is in phon in the formal statement
of constructional correspondences.
The specification of the syntactic representation as [V kick] identifies the
lexical properties, in this case, the lexical identity, of the word, which is
one of the terms that the paradigm function applies to. As noted by Sag
(2012), the phonological form and the meaning together are not sufficient to
distinguish the lexical item, since it may lack an interpretation in an idiom,
e.g. kick the bucket.⁸ And the use of λ and the thematic roles agent, patient
in the semantic representation roughly indicates the argument structure that
corresponds to this verb. This is the basic picture, to be refined as we proceed.
⁸ In this respect I depart from the practice in Culicover & Jackendoff (2005) and elsewhere.
12 overview
Finally, as in Relational Grammar (Perlmutter 1983; Perlmutter & Rosen

1984), the constructional framework as described here assumes that grammat-
ical functions have a role in grammatical descriptions. However, unlike RG
(but following Simpler Syntax), I do not assume that the GFs such as subject
and object are universal primitives, nor do I assume the grammar is formulated
entirely in terms of GFs. It is formulated in terms of constructions, which
may make reference to GFs. Relations between constructions may then also
be expressed in terms of the GFs.
I argue in Chapter 5 that in certain cases the GFs are categories of
correspondences that emerge in the course of language change as general-
izations of features of thematic relations. In others, they are descendants of
grammatical devices for marking information structure. They appear to play
a useful computational role, and thus constitute a part of the resources that
constitute the collective knowledge of language in the social network.
1.2.2 Constructions are not derivations
In order to appreciate the constructional approach, it is useful to contrast it

with MGG.⁹ Some of the main differences are the following:
(i) MGG licenses linear ordering by moving constituents from canonical

positions to designated positions. In many constructional approaches,
including the one in this book, the grammar states a correspondence
between a particular syntactic structure and a linear ordering which
gives the appearance of movement when constituents are adjacent in
syn (e.g. they are sisters) but their phonological forms are not adjacent
in phon, or when there is more than one possible ordering for the
phonological forms corresponding to a particular syntactic structure.
(ii) MGG assumes a strictly compositional semantics; a constructional
grammar does not (although it must account for compositionality
where it occurs)—cf. Jackendoff ’s (1997) notion of ‘enriched composi-
tion’.
(iii) MGG assumes a uniform syntactic structure for the same meaning in
a single language and cross-linguistically; a constructional grammar
does not.1⁰
⁹ I am using the term ‘constructional’ in a generic sense, in order to maintain a distinction between
the approach that I develop here and the specifics of Construction Grammar (Fillmore 1988;
Goldberg 1995; Kay 2002; Michaelis 2012). Goldberg (2013) uses the term ‘constructionist’.
1⁰ For an extended critique of the application of uniformity in MGG, see Culicover & Jackendoff
(2005, chapters 2 & 3).
1.2 constructions 13
(iv) MGG is rigid in its licensing conditions; a constructional grammar is

flexible.
(v) In MGG, the form of an inflected or otherwise complex word is
derived through movement and adjunction; the resulting constituent
is spelled out according to information stored in a list of forms.11 On
the constructional approach, the form of these words is defined by the
paradigm function Φ or the equivalent.
An extended discussion of each of these points would take us too far afield,
so I elaborate only the first point, which is central to much of what is taken up
elsewhere in this book. Consider the example of adjective-noun ordering in
French and Italian. The canonical order is Adj-N, but the order N-Adj is also
possible, as shown by the French examples in (5).
(5) French
a. une belle maison
a beautiful house
‘a beautiful house’
b. une voiture rouge
a car red
‘a red car’
Taking the underlying order as Adj-N, the order N-Adj is derived in MGG
by raising the N to a preadjectival position (Cinque 1994, 2005). To the extent
that the two orders correspond to a semantic difference, the raising of N can
be triggered by a feature on Adj that corresponds to this semantic difference
and must be ‘checked off ’ by adjoining N to it in the observed particular
configuration. Or the meaning difference can be associated with invisible
elements in the syntactic representation with the appropriate interpretation
that differentially trigger movements to different superficial positions.
In a sense the alternation here is ‘parametric’, because for a language to have
a particular order it must have a particular feature value or invisible element,
while a language with the other order has a different feature value or element
(or lacks an element). But the use of features or invisible elements to trigger
the different orders and the corresponding meaning differences is actually
constructional: the observed constructional differences are implemented in
terms of features, abstract elements, and movement. I have called this approach
11 For an extreme version of this view, see recent work in Distributed Morphology (e.g. Halle &
Marantz 1993, 1994; Embick & Noyer 2007; Harley & Noyer 1999). There is no lexicon per se in
DM. Rather, the sound and meaning information that we associate with lexemes is distributed over
a Vocabulary and an Encyclopedia, while the categorial information is in the syntax.
14 overview
‘cryptoconstructionalism’ (Culicover 2017), and will argue as we proceed that

an explicit constructional approach captures the phenomena more transpar-
ently and in most cases more simply.12
The constructional alternative in the case of adjective-noun order is to
formulate the two orders in the description of distinct constructions, and
to associate the meaning difference directly with the two orders licensed by
the constructions; the correspondence is mediated through the phonological
correspondence with the syntactic categories. The argument for one approach
over the other is not one of empirical coverage: the two approaches cover the
same phenomena. The marking of one adjective with a feature that triggers
movement, but not another, is in fact the implementation in a derivational
framework of a constructional fact. What is constructional is that the linear
ordering is sensitive to lexical properties, perhaps arbitrary lexical properties.13
Many, if not most derivational analyses of linear ordering in MGG are
in fact cryptoconstructional. That is, they use the derivational apparatus
of feature checking and movement to encode exactly the information that
can be encoded directly in a construction. In other words, they are actually
constructional analyses in which the linear order and phonological form of
elements are mediated by derivations, and not expressed directly. For this
reason, I do not dwell too much on the specifics of such analyses in this book,
except in a few cases where it is useful to emphasize this point.
Chapter 2 elaborates the constructional approach further and compares it
with the alternatives.
1.3 Antecedents
Since the scope of the present study is substantial, it touches on many issues
and topics that have been addressed in many syntactic theories. And, since
space is limited, I am not going to be able to refer in detail to most theoretical
antecedents and alternatives to my own proposals. In lieu of that, I offer the
following general observations, which are not intended to be exhaustive.
All contemporary theories are right about something, although not every-
thing, and not the same things. There is considerable insight in work within
Government Binding (GB) theory, Principles and Parameters theory, the
Minimalist Program, HPSG, LFG, Relational Grammar, Role and Reference
Grammar, and Combinatory Categorial Grammar. All of these have an illu-
12 For a remark that appears to accord with this sentiment, see Chomsky et al. (2019, 251).
13 For one approach, see Bouchard (2009).
1.3 antecedents 15
minating perspective on some (but often different) fundamental property or

properties of language. I have taken much from all of them, in the spirit of
Simpler Syntax, which stresses an eclectic approach to understanding how
language works.
• MGG is right about the existence of a universal faculty that underlies

language. But this faculty is not just about syntax—it is about CS and how
phonological and morphological forms are recruited to systematically
express aspects of CS. What is universal is the architecture of correspon-
dences that constructions express. Formal devices that do an effective
job of expressing the obligatory cs-phon correspondences are likely to
persevere in competition between languages and eventually will find their
way into many grammars. While they are not universals in the sense that
they must be found in every language, they are ubiquitous.
• Within MGG, GB theory is right about modularity. But I take the modules
to be different. In a constructional theory, there are independent princi-
ples and rules of combination for phonology, syntax, morphology, and
conceptual structure.
• Within MGG, Principles and Parameters theory is right in a certain sense
about variation. But the parameters of variation are not about UG, and
are not universal. Some are defined at the constructional level, in terms of
the logical possibilities for characterizing the cs-phon correspondences.
I discuss how to understand parameters in constructional terms at greater
length in Chapter 10.
• The Minimalist Program is right about the explanatory relevance of
economy. But the measure is not about some abstract computation, but
about the actual cs-syn-phon correspondences that a grammar must
express.
• HPSG is right about monostratality and the projection of structure from
heads.
• LFG is right about correspondences between levels of representation and
the basic grammatical architecture.
• Relational grammar is right that grammatical functions have a role in
grammar, as discussed in section 1.2.1.
• Combinatorial categorial grammar is right about the architecture of
correspondence from cs to phon. The relationship is mediated by con-
structions, however, so that the correspondence may be more or less
faithful and economical. One consequence is the possibility of variation
in generality, transparency, and complexity.
2
Constructions
2.1 Introduction
In order to be able to talk in precise terms about language variation and change
we must be able to characterize languages in terms of grammatical variation
and change. Hence we must be able to characterize grammars. As already
discussed in Chapter 1, what defines a language is not a set of sentences, but
a mental representation that embodies a grammar. This book argues that a
useful formulation of this mental representation from the perspective both of
description and explanation is in terms of constructions.
In this chapter I outline the constructional theory that I assume in order to
frame the scenario of language variation and change, with the goal of providing
a solution to Chomsky’s Problem, i.e. why there are different languages at all?
Why don’t we all speak the same language?
In section 2.2 I sketch out the constructional approach to knowledge of a
language. In 2.3 I elaborate the constructional formalism and explain how
constructions characterize the well-formedness of linguistic expressions. I also
provide a number of examples of English constructions to illustrate how the
system works. A more formal development of these ideas is given in the
Appendix to this chapter.
2.2 What a grammar is for
In this section I summarize how a constructional approach characterizes what

it is that speakers know when they know a language.
The standard assumption in linguistic theory is that the job of a grammar is
to characterize the well-formed expressions of a language. That is, a grammar
is a generative grammar in the usual sense. A well-formed expression is
an expression whose form and other properties conform to the conditions
imposed on it by the grammar. To know a language is to have a mental
representation that captures the well-formedness conditions of expressions in
that language.
Language Change, Variation, and Universals: A Constructional Approach. Peter W. Culicover, Oxford University Press.
© Peter W. Culicover 2021. DOI: 10.1093/oso/9780198865391.003.0002
2.2 what a grammar is for 17
To take a simple example, Sandy snores is well-formed in English because the

NP Sandy precedes the verb snores and, since Sandy is a third person singular
noun phrase, the verb snores shows the third person singular present tense
form. That is, subject and verb ‘agree’ in person and number. Some ill-formed
expressions that are not licensed by the grammar are ∗ Sandy snore, ∗ snores
Sandy, and ∗ snore Sandy, which violate the requirement of agreement, of linear
order, or both. The grammar of English must in some way state explicitly the
ordering and agreement requirements that only Sandy snores satisfies.
There is another sense in which to understand well-formedness with respect
to Sandy snores, however. Suppose that I said that Sandy snores means “The
square root of two is irrational.” While the form of the sentence would not be
in violation of the traditional rules of English grammatical form, and the cor-
responding meaning is coherent, there would be a failure of correspondence
between the form and the meaning. The word Sandy does not correspond to
the meaning ‘the square root of two’, and the word snores does not correspond
to the meaning ‘is irrational’.
In conventional approaches to the relationship between form and meaning,
there cannot be a failure of correspondence because the meaning of a string of
words is determined by a computation that takes as its input the individual
words and their meanings and the syntactic structure in which they are
arranged. So, assuming that the meaning of Sandy is s and the meaning of
snores is snore′ , the syntactic structure tells us to apply the meaning of the verb
to the meaning of the NP to get (roughly) snore′ (s).
This approach to semantic interpretation is called ‘compositional’ or ‘Freg-
ean’ (Partee et al. 1990). Non-Fregean approaches, such as the one sketched
out here, argue that there are aspects of interpretation that cannot be localized
in the meanings of the individual words but must be associated with the
structure in which they appear (Goldberg 1995, 2003). A classic example is the
double object construction, V-NP1 -NP2 , which conveys the notion ‘transfer
possession of NP2 to NP1 by means of V-ing’, whether or not the verb has
transfer of possession as part of its lexical meaning.1
The examples in (1) illustrate. Give literally means ‘transfer possession’, so
(1a) is consistent with the ‘transfer possession’ meaning associated with the
construction. But in (1b) head is a verb that means simply ‘hit with one’s head’.
1 It is of course possible to complicate the syntax in order to represent all aspects of meaning
in a cryptocompositional format, by positing phonologically empty elements that have the non-
compositional components of the meaning. A constructional approach applies Occam’s Razor to arrive
at a naive, simple syntactic representation in the spirit of Culicover & Jackendoff (2005), one in which
such ad hoc meaning-bearing empty elements are ruled out in principle.
18 constructions
‘Transfer possession’ in this case is part of the interpretation in virtue of the

construction itself, not the meaning of the verb.
(1) a. Sandy gave Chris the money.

b. Sandy headed Chris the ball. (= ‘Sandy transferred possession of the
ball to Chris by hitting it with his/her head.’)
On a constructional approach, then, well-formedness applies not just to

the organization and morphological form of a string of words, but to a
phonological string—in this case, a string of words—with a corresponding
interpretation. A representation consisting (minimally) of a string of words
and a meaning is a construct. In order for a construct to be well-formed,
it must be licensed. That is, its form must satisfy the phonological, morpho-
logical, and syntactic conditions and its meaning must satisfy the semantic
conditions imposed by the constructions that constitute the grammar, as in
the Sandy snores, etc. examples discussed above. Such a correspondence can
be very general, as in the case we have just been considering, or it can license
an idiomatic expresion or a single word.2
An interesting and important property of the constructional approach that
is elaborated in Chapter 3 is that constructions are ‘elastic’: the licensing
conditions for a construction can be expanded or shrunk to cover larger or
smaller sets of elements or categories to account for variation or to characterize
change. So dialects and even individual grammars may differ just in terms of
whether or not a particular lexical item participates in a given construction (cf.
give NP1 NP2 vs. ∗ donate NP1 NP2 ).3
Furthermore, it is possible to have multiple constructions in a grammar
that overlap or conflict. Under such circumstances, native speaker competence
is what we would expect if there were ‘multiple grammars’ (Kroch 1989)
competing with one another in the social network. But it is not necessary
to appeal to multiple grammars if single grammars can incorporate multiple
2 Note that many constructionalists prefer to reserve the term ‘construction’ for syntactically
complex correspondences that involve non-compositional meaning.
3 The ditransitive form with donate is generally claimed by generative grammarians to be impossible,
but is arguably not ruled out on principled grounds. In fact, the following appears quite innocently in
the 1901 Transactions of the Annual Convocation of the Royal Arch Masons, Grand Chapter (Mich.).
(i) If not, can we donate him the amount of the fee, so that he can pay for his degrees, and if so
should it be done previous to the petition?
And here is an example where the indirect object is a full NP.
(ii) The library now is looking for some organization to donate the library a subscription to the Den-
ton Record-Chronicle, so that a bound file may be kept of it …(https://www.newspapers.com/
newspage/11158344/)
For discussion of the elasticity of constructions, see Goldberg (2019).
2.3 a framework for constructions 19
constructions with licensing conditions of varying generality. It is sufficient

to assume that there are multiple constructions that are competing, not the
grammars (Henry 2008, 274). For elaboration of this point, see Chapter 4.
2.3 A framework for constructions
Let us now consider in more detail what a theory of constructions looks like.
A theory of constructions is a theory of licensing of correspondences, not a
theory of syntactic derivations or of constraints (Michaelis 2012). The goal is
to explain what determines the well-formedness of an arbitrary construct in
a language, that is, a form/meaning correspondence representing a token of
the language, in terms of the constructions that comprise the grammar of that
language.
2.3.1 Representing constructions
Following the Parallel Architecture of Jackendoff (2002), I take a construction

to be a correspondence between elements of four ‘tiers’, phon, syn, gf (for
‘grammatical functions’), and cs (for conceptual structure). Each tier has
its own primitives, rules of combination, and constraints. The well-formed
expressions of a language are just those that exemplify the correspondences
between the tiers that are stipulated by the constructions of the grammar.
A construction is a generalization over such correspondences that states what
correlated properties of representations on the tiers determine the well-formed
expressions of the language. A construct that satisfies the conditions of a
construction is said to be licensed by that construction.
Crucially, phon is responsible for all aspects of overt form, such as the
pronunciation of individual words, their morphological inflections, and their
temporal ordering.⁴ syn is responsible for the hierarchical organization of the
words and phrases. I assume that linear order is represented only in phon, and
not in syn.⁵ This is a point that calls for some elaboration, and I return to it in
more detail in section 2.3.3.
⁴ A complication that I set aside here is the proper treatment of prosody.

⁵ This approach to constructions assumes that linear order is a phonological property, while mor-
phosyntactic category, constituency, and hierarchical structure are represented in syn. The separation
of phonological and syntactic properties echoes one originally made by Curry (1963). It was not
assumed in early MGG, but is seen in a number of other approaches, such as HPSG and LFG. For
additional discussion of this and related points, see section 2.3.3.
20 constructions
To illustrate the kinds of constructions that can be defined in these terms,

I formulate a series of examples of increasing complexity. I start with individual
lexemes, look at several idioms, and work up to phrasal constructions.
Individual lexemes are minimal constructions. Each lexeme specifies a
correspondence between a string of sounds, morphosyntactic properties, and
a meaning, and this correspondence must be represented explicitly in the
lexicon of the language.
For example, the lexical entry of the verb kick is (2).
(2) kick
⎡phon ΦEnglish (kick)=/kɪk/1 ⎤
⎢syn [V kick]1 ⎥
⎢ ′
⎥
⎣cs [λy.λx.kick 1 (agent:x,theme:y)]⎦
I assume for convenience that the interpretation consists of a primitive

concept, here represented by the word marked with a prime, following com-
mon practice in formal semantics. Thus, I do not address directly the lexical
semantics of the individual words—that is a major topic in its own right, quite
apart from the current focus.
The interpretation includes as well thematic relations such as agent and
patient that distinguish the arguments. The decomposition of the thematic
relations into more basic thematic features, along the lines of Dowty (1991), is
discussed at some length in Chapter 5.
There are two parts to the correspondence for kick that we must consider,
the phon-syn correspondence and the syn-cs correspondence. As discussed
briefly in Chapter 1, the entry in (2) is approximate, because it does not take
into account the fact that the lexeme kick may take different forms depending
on its inflection: kick, kicks, kicked, kicking. For irregular verbs, the set of forms
is larger, e.g. go, goes, went, gone, going. This information is captured by the
paradigm function Φ, applied to the lexical item and its inflectional features.
In order to simplify the exposition and develop the main intuitions of
our approach to constructions, it will be useful to adopt several notational
conventions. As shown in (2), the corresponding parts of a construction on
different tiers are coindexed. The terms in a construction may be constants, as
in (2), or they may be variables, when categories are involved.
For the phonological constants, I use the orthographic form in phon instead
of the paradigm function or its phonetic value. Thus, (2) would be written
as (3).
(3) kick
⎡phon kick1 ⎤
⎢syn [V kick]1 ⎥
⎢ ⎥
⎣cs [λy.λx.kick′ 1 (agent:x,theme:y)]⎦
For constructions that state correspondences between categories in syn and

their representations in phon and cs I use just the indices. For example,
consider a construction in language L where the verb is inflected. Letting V
denote the verb and φ its inflectional features, we have the construction in (4).
(4) V in language L
phon ΦL (1,2)
[ ]
syn V1 [φ2 ]
To simplify, I use the index itself in phon to refer to the value of the paradigm
function applied to the item with that index; other indices may correspond to
relevant morphosyntactic properties. Thus, the content of (4) is written as (5).
(5) V in language L
phon 1(2)
[ ]
syn V1 [φ2 ]
Finally, many constructions involve complex correspondences between

phonological form and items with inflectional and syntactic features. These
items are easiest to represent as attribute-value matrices, e.g. the attribute
category has the value V, the attribute person has the value 3rd, etc.
I represent the lexical identity as an attribute whose value is the particular
lexical item; this is the lexeme identifier, or lid (a notion adapted from Sag
2012; see also Stump 2001; Spencer 2013; Harley 2014; Blevins 2016).
(6) ⎡category V ⎤
⎢lid kick⎥
⎢ ⎥
⎢person 3rd ⎥
⎢ ⎥
⎢number sg ⎥
⎢ ⎥
⎣tense pres⎦
When the full detail is not needed, I continue to use representations like
[V kicks].
22 constructions
Now let’s look at idioms. The representation of the multiword expression

by and large is given in (7). The ordering of the individual words is critical
and must be specified in the construction in phon; cf. ∗ large and by. I use ‘–’
to represent strict ordering in phon, and > to represent relative ordering. As
noted earlier, I assume that there is no ordering in syn, and that syn represents
only the hierarchical structure, the categories and the morphological features
(see section 2.3.3).
(7) Construction: by and large

⎡phon [by2 –and3 –large4 ]1 ⎤
⎢syn [ADV [P by]2 , [CONJ and]3 , [ADJ large]4 ]1 ⎥
⎢ ⎥
⎣cs mostly′ 1 ⎦
The construct ∗ large-and-by with the meaning ‘mostly’ is not licensed by this
construction, since while it has the correct syn and cs, it does not have the
correct phon.
Consider next the idiom kick the bucket. The stipulation that ‘kick’ precedes
‘the bucket’ should be cost free in the syntactic description, because this is
a perfectly normal VP in English and thus corresponds in a regular way to
the syntactic structure (see (11)). In the case of idioms with normal structure,
the linear ordering is licensed by the appropriate ordering conditions for the
corresponding syntactic categories, as are the grammatical functions. The
phonological form of the verb, as in kicks the bucket, kicked the bucket, is kicking
the bucket, etc., does not need to be specified in the construction itself, since it
follows independently from the paradigm function ΦEnglish .
The only thing that is necessary to distinguish the idiom from the literal
expression with the same form is the identification of the set of words that
constitute the idiom with the corresponding meaning. So we can represent
kick the bucket in the lexicon as in (8).The lid’s are represented by the lexemes
in italics. GF4 is the grammatical function of the bucket.
(8) kick the bucket

⎡syn [VP [V kick]1 , [NP [ART the]2 , [N bucket]3 ]4 ]5 ⎤
⎢gf GF4 ⎥
⎢ ′
⎥
⎣cs [λx.die 5 (experiencer:x)] ⎦
Next consider sell down the river, which is partially lexically specified, but
contains a variable term.
(9) sell NP down the river

⎡syn [VP [V sell]1 , NP2 , [PP [P down]5 , [NP [ART the]6 , [N river]7 ]8 ]3 ]4 ⎤
⎢gf GF2 ⎥
⎢ ⎥
⎣cs [λy.λx.betray′ 1+3 (agent:x, theme:y)(2′ )]4 ⎦
There is no phon defined for this construction, because it is built around a

VP. As before, it is incumbent on this construction to specify the hierarchical
structure on the syn tier, but not the order of elements. Other constructions,
such as those that license VP, NP, and PP, license the order of elements and the
grammatical functions. Thus, the ordering 1–2–3 is just one possible ordering
in VP (10a). Another is 1–3–2 (10b), which is licensed if 2 corresponds to a
‘heavy’ NP, as discussed in Culicover et al. (2017).
Sandy
(10) a. They sold { them } down the river.
the people who trusted them
?Sandy
b. They sold down the river { ∗ them }.
the people who trusted them
This ordering is the responsibility of a construction or set of constructions that

link the order of constituents in the VP to information structure and other
factors (Wasow 2002; see also Culicover & Winkler 2008).
An example of a fully general construction is the one that licenses VP in
English. This construction is not specified for any particular V or set of Vs.⁶
The construction says that a VP may consist of a V followed by other, possibly
null, material; it has the form in (11). The phon of (11) says that any daughter
in VP follows V.⁷ Since for each VP the interpretation is determined by the
semantics of the verb, we do not specify the CS in the formulation of this
particular construction.
(11) Construction: VP-initial V

phon [1>2]3
[ ]
syn [VP V1 , X2 ]3
⁶ Strictly speaking, the notation should distinguish between V as the category of an individual lexical
item and V as a variable over items in this category. I have chosen not to complicate the notation, leaving
it to the different contexts in which the category symbols appear to distinguish them.
⁷ A more comprehensive account might reflect the fact that this condition may be violated by
preverbal adverbs, as in Sandy will completely fail the test, if such adverbs are daughters of VP. If they
form a complex V with the lexical verb, or are attached higher to the VP, then the generalization holds
as stated.
24 constructions
Crucially, (11) alone does not fully license an actual VP with non-null sisters
of V. The complements and adjuncts must be licensed by other constructions
for expressions of the relevant categories, e.g. NP, CP, AP, PP, VP, and so on.
Consider next the GFs. Subject and object are the traditional names for
the highest ranked GF and the next highest ranked GF in the GF hierar-
chy; I assume that they are not primitive universals.⁸ The constructions in
(12)–(13), based on Culicover & Jackendoff (2005), explicitly express the cor-
respondence between the subject and object GFs and syntactic configurations.
(12) says that the NP daughter of S corresponds to the highest grammatical
function in the domain of that S, and (13) says that the NP sister of V
corresponds to the next highest function.
(12) Construction: Subject

syn [S NP1 , AUX2 , …]
[ ]
gf [GF1 (> …)]2
(13) Construction: Object

syn [VP V, NP1 ]2
[ ]
gf [GF > GF1 ]2
The Subject construction (12) does not specify the linear position of the
subject NP. That is the responsibility of other constructions, such as declara-
tive, which places it before the inflected verb, subject Aux inversion (SAI),
which places it immediately after an inflected auxiliary, and special focus
constructions, which place it after V (Culicover & Levine 2001; Culicover &
Winkler 2008).
Note that the Object construction assigns the second grammatical function
in the hierarchy to the direct object sister of V.⁹ We also need to assign a GF to
the complement of certain prepositions.
(14) Construction: Oblique object

syn [VP V, [PP P, NP1 ]]2
[ ]
gf [GF > GF1 ]2
For discussion of the history of this last construction, see section 9.3.
⁸ In fact, the highest ranked GF need not have the same properties across all languages. For
discussion, see Chapter 6.
⁹ Following Culicover & Jackendoff (2005), I assume that the structure of this VP is flat.
The well-known double object construction, exemplified by Chris gave

Sandy a book, can be formulated as in (15). The specification means:1′ in CS is
the non-compositional part of the meaning; informally, it is the action denoted
by the verb by means of which the transfer is accomplished (Goldberg 1995).
(15) Construction: Double object

phon [1–2–3]4
⎡ ⎤
⎢syn [VP V1 , NP2 , NP3 ]4 ⎥
⎢gf [GF > GF ] ⎥
⎢ 2 4 ⎥
⎣cs [λz.λy.λx.transfer′ (source:x,goal:y3 ,theme:z4 ,means:1′ )(2′ )(3′ )]4 ⎦
The application of this construction to (Chris) gave Sandy a book is shown

in (16). s and b are the cs representations of Sandy and a book, respectively.
(16) gave Sandy a book

phon [gave1 –Sandy2 –a-book3 ]4
⎡ ⎤
⎢syn [VP V[give, past]1 , NP[Sandy]2 , NP[a book]3 ]4 ⎥
⎢ ⎥
⎢gf [GF > GF2 ]4 ⎥
⎢cs [λz.λy.λx.transfer′ (source:x,goal:y,theme:z,means:give′ 1 )(s2 )(b3 )]4 ⎥
⎢ ⎥
⎣ ⇒ λx.transfer′ (source:x,goal:s2 ,theme:b3 ,means:give′ 1 ) ⎦
Note that I do not associate a GF with the direct object in the double object
construction. This is because there are no grammatical phenomena in English,
such as passive, that refer to this constituent. The subject argument of the
passive in Standard English corresponds to the object that is closest to the verb
in the active. So, when there are two objects, only the first object can become
a passive subject.1⁰
1⁰ However, in some dialects, and particularly the style of the King James Bible, the second object
may become the subject of the passive. This is most common when the first object is a pronoun.
(i) a. Therefore I prayed, and prudence was given me; I pleaded and the spirit of Wisdom came to me.
b. Do not neglect the gift you have, which was given you by prophecy when the council of elders
laid their hands on you.
c. So they took the bull which was given them, and they prepared it . . .
d. They were given this land on long-time payments, the water was supplied, every possible
assistance was given those people.
For such dialects we could assign GF2 to the direct object when the indirect object is pronominal.
Plausibly, the analysis of the VP in such cases would not treat the pronominal argument as NP2 , but as
a clitic adjoined to V, in which case the object construction in (13) would suffice; alternatively, there
might be a separate construction for pronominal direct objects, as well as full NPs as in (i.d), in which
both arguments have GFs. I leave the question open here.
26 constructions
(17) a. Chris gave Sandy a book.

b. Sandy was given a book.
c. ∗ A book was given Sandy.
(18) a. Chris gave a book to Sandy.
b. A book was given to Sandy.
c. ∗ Sandy was given a book to.
(19) a. Chris talked to Sandy about Lee.
b. Sandy was talked to about Lee.
c. ∗ Lee was talked to Sandy about.
(20) a. Chris talked about Lee to Sandy.
b. Lee was talked about to Sandy.
c. ∗ Sandy was talked about Lee to.
This type of correspondence between active and passive is central to

accounting for phenomena that have been described in terms of meaning-
preserving transformations in classical MGG. In a constructional framework,
we can state two constructions as schemas, and express the relationship
between them by indexing the parts of the schemas that they share. When two
constructions share CS representations and differ only in syntactic structure
and corresponding phonological form, they have the appearance of classical
transformations. But the correspondences may be less exact, and it is in
principle possible for them to have different representations on the cs tier.
For example, the construction that licenses yes-no questions in English
is related to the construction that licenses declaratives. There are two dif-
ferences between the two constructions. First, the inflected Vaux is in initial
position in the question, and in second position, after the subject, in the
declarative. Second, the question has an interrogative interpretation, while
the declarative does not. So we can represent the relationship between the
two constructions as in (21), where the superscripts x,y mark the terms that
correspond across the constructions. The schemas are sister schemas in the
sense of Jackendoff & Audring (2020).11
(21) Correspondence: declarative ⇔ yes-no question

phon [1–2–…]3 ⇔ phon [2–1–…]3
⎡ y
⎤ ⎡ y
⎤
⎢syn [S NPx1 , V[finite]2 , …]3 ⎥ ⎢syn [S NPx1 , Vaux [finite]2 , …]3 ⎥
⎢ ⎥ ⎢ ⎥
⎣cs 3′ ⎦ ⎣cs Q(3′ ) ⎦
11 Ray Jackendoff (p.c.) suggests that such constructional correspondences are realizations of the
original notion of ‘transformation’ due to Harris (1951, 1957).
This relational formulation says explicitly that the question that corresponds
to a declarative is one in which the order of subject and Aux is inverted.12
Similarly, the passive is formulated in (19) as a relation between active and
passive. This relational statement says that the argument that corresponds to
the second ranked GF in the active corresponds to the first ranked GF when
the verb is a passive participle.
(22) Relation: active ⇔ passive

⎡ y
NP1 ⎤
⎢syn [S …, [VP V , {[ P, NP ]} ]]2 ⎥
⎢ PP 1 ⎥
⎢ x ⎥
⎣gf [GF > GF ]
1 2 ⎦
⇔ syn […, [VP Vy [passive]]]2
[ ]
gf [GFx ]2 ]
This formulation departs from the treatment of passive in Culicover &

Jackendoff (2005), and reflects the Relational Grammar intuition that passive
is a construction that promotes a 2-argument to a 1-argument, by suppressing
the higher argument. Since we do not have to explicitly represent the position
of the corresponding GFs in either active or passive in the statement of this
correspondence between constructions, we are able to account for passives
where there is no overt subject, e.g. in cases of control (23a) and reduced
relatives (23b). In these constructions, the highest GF does not correspond
to anything in syn.
(23) a. Chris expected to be [VP[passive] selected].

b. the student [VP[passive] selected as our representative]
See Culicover & Jackendoff (2005, 194ff) for discussion of such correspondences.13
2.3.2 Licensing
The examples of the preceding section illustrate two key aspects of licensing:
(i) Each element of each tier of a construct must satisfy some condition of
12 For the distribution of do-support in this and other constructions, see the discussion in section 9.2.
The description of more complex constructions such as wh-questions, topicalization, and Germanic V2
is taken up in Chapters 7 and 8.
13 This relational treatment of passive also avoids a problem with ‘double passives’ noted by Müller
(2013, 925–7) in connection with the treatment in Culicover & Jackendoff (2005).
28 constructions
some construction, and (ii) all aspects of the construct must be licensed. For
instance, in the construct Chris kicked Fido, the phon tier must represent
the appropriate forms for each of the words in the appropriate order, and
the meaning must be one in which the relation kick′ holds of the entities c,
the CS representation of Chris, and f, the CS representation of Fido. Similarly,
Sandy kicked the bucket is licensed by the idiom construction and the more
general constructions if die′ is predicated of s, and if the proper linear ordering
is observed.1⁴
While the idea is intuitively simple, the implementation of a definition
of licensing that satisfies this description is non-trivial. The most important
parts of the definition are given in sections 2.4.1–2.4.3 in the Appendix to this
chapter.
2.3.3 Linear order
In section 2.3.1 I noted that in the current framework, linear order is repre-
sented only in phon, while syn is where hierarchical structure is represented.
This does not mean that syntactic structure has nothing to do with linear order.
In fact, licensing of linear order is largely dependent on syntactic structure—
more so in some languages than others, of course—and it would be impossible
to state generalizations about linear order in a language without reference to
representations in syn.
The nature of this relationship is a complex one with a long history. Docu-
menting it adequately could easily occupy a monograph on its own. In order
not to depart too much from the current narrative, I devote just this brief
section to the question and leave a fuller discussion to another venue.
To begin, it is a truism that linear order is associated with particular phrase
structure configurations in many languages. For instance, in English the verb
is initial in the VP.
So we could say, following a long tradition, that the syntactic structure for
VP in English represents the initial position of V, and this corresponds redun-
dantly to the linear order of the corresponding elements in phon. The syntactic
representation would essentially be the familiar [VP V XP]. Alternatively, we
could say that the syntactic structure does not represent linear order. The order
is defined through the correspondence with phon, which does represent linear
1⁴ There is a possibility that the ability of an idiom to appear in various structures, e.g. passive,
correlates with the transparency of the correspondence between syn and cs, as suggested by Nunberg
et al. (1994); see also Sag (2012). Similar considerations may account for the degree of productivity in
derivational morphology; see Jackendoff & Audring (2020).
Another random document with
no related content on Scribd:
The Project Gutenberg eBook of Inconstancy
This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.
Title: Inconstancy
Author: Roger D. Aycock
Illustrator: Leo Summers
Release date: November 22, 2023 [eBook #72203]
Language: English
Original publication: New York, NY: Ziff-Davis Publishing Company,

1961
Credits: Greg Weeks, Mary Meehan and the Online Distributed

Proofreading Team at http://www.pgdp.net
*** START OF THE PROJECT GUTENBERG EBOOK

INCONSTANCY ***
INCONSTANCY
By ROGER DEE
Illustrated by SUMMERS
The trouble with a Martian-Terran romance

is that it has to buck things like tradition.
Up on Mars, when they sing "If you were the only
girl in the world," they really mean it.
[Transcriber's Note: This etext was produced from

Amazing Stories January 1962.
Extensive research did not uncover any evidence that
the U.S. copyright on this publication was renewed.]
His first day on Earth promised to be even worse than Mirrh Yahn y
Cona had feared when he left Yrml Orise y Yrl, his fiancee, to
become Mars' first interplanetary ambassador. The frenetic bustle of
Denver spaceport, his ominous spiriting away through screaming
hordes of spectators, left him bewildered and uneasy.
Alone in the first brief privacy of his Denver Heptagon apartment, he
ideographed a facsimile transmission to Yrml at once. "I long for you
already," he said. "And for the serenity of home. Earthpeople are as
barbarous and mercurial as their weather."
Babelous decades of taped newsreels and video serials should have
prepared him for that inconstancy, but the first-hand reality was
appalling. He would gladly have returned home at once, before
planetary conjunction's end cut him off for two interminable years, but
for the inevitable stumbling-block: Earth had sent an exchange of her
own, and Mirrh Yahn y Cona could not back down without disgracing
his planet as well as himself.
"Write often," he pleaded, in closing. "That I may take comfort in your
steadfast regard even in this simian hurlyburly."
The missive finished, he found time remaining before Ellis, of
Diplomatic, arrived to switch on the multisensory projection of his last
evening with Yrml. The projection had been cubed in a Privileged
Couples nook complete with real plants and hermetically sealed
fountain, and near its close the two of them had sung the traditional
Song of Parting from the ancient Tchulkione Serafi.
Ellis arrived all too soon, trailing an aura of Scotch, diplomatic
enthusiasm and geniality.
"No time to waste," Ellis said briskly. "Little enough of it before you
leave us, and you're going to see Earth from pole to pole. The three
of us begin this evening with a sample of Denver night life."
"Three?"
"Came early to brief you," Ellis said. "Found a guide for you. Can't run
about unescorted, you know."
He answered the door buzzer and admitted a young woman in
evening dress. Rushed from the spaceport in what amounted to
cloak-and-dagger secrecy, Mirrh Yahn y Cona had until now seen
Earthwomen only on video and at indistinguishable distance, and the
sudden appearance of this one in the flesh unnerved him completely.
The girl was small and slender, well under Mirrh Yahn y Cona's
athletic six-foot height. She was warmly and roundly vital with a
stunning abundance of life at which the two-dimensional simulacra of
recorded soap-opera could only hint.
"Miss Leila Anderson," Ellis introduced her. "Member of Diplomatic,
so it's all in the family."
She took the hand that Mirrh Yahn y Cona raised as if to defend
himself.
"I'm to see that you aren't bored to death here among strangers," she
said. "All work and no play isn't good for anyone. Especially," she said
to Ellis, "for one so handsome. I didn't dream he'd look so—"
"So Terran," Ellis finished before she could say so human. "And why
not? We're from the same original stock, separated ages before our
history begins. Martian annals run back for millennia, did you know?
Gold mine of information, settle problems our experts have puzzled
over for centuries."
"I am not truly representative of my people," Mirrh Yahn y Cona said
with some bitterness. "A special case, reared from birth for this
assignment."
The multisensory projector swung into the Tchulkione Serafi's Song of
Parting. Mirrh Yahn y Cona's resonant baritone, operatically assertive
above Yrml's reedy soprano, filled the room. He shut off the machine
abruptly, feeling a sense of desecration that the tender scene had
been bared to alien eyes.
Still he felt a puzzling premonitory twinge of guilt when the projection
collapsed. Yrml had been infinitely desirable when the sequence was
cubed; why should she now seem so sallow and angular, so suddenly
and subtly distant?
"Remarkable voice," Ellis said. "You could make a fortune with it
here."
"It was lovely," Leila Anderson said. "Could I hear the rest of it some
time?"
"No." He realized his curtness and added, "It is the Song of Parting
for lovers. Very personal."
He found that he was still holding Leila's hand, and dropped it hastily.
Ellis, who had risen high in Diplomatic for good reasons, stepped
competently into the breach.
"Night duty calls," Ellis said. "Let's be off."
A diplomatic limousine without insignia took them to a nightclub large

enough, and dim enough, to promise anonymity. On the way a quick
summer shower left the streets wet and glistening and turned the
night into a many-scented freshness that was sheer fantasy to one
accustomed to the sterile air of sealed underground ways.
The rain had ended when they left the car, but the brief moment
outside, under a vast openness of night sky empty except for
dispersing clouds and speeding white moon, struck Mirrh Yahn y
Cona suddenly cold with too-familiar panic.
They had found their table before anyone spoke.
"Agoraphobia?" Ellis said, in frowning concern. "I should think you'd
be conditioned against that, with all the time they've had to prepare
you."
Leila Anderson put an impulsive hand on the Martian's.
"I'm a touch claustrophobic, so I know how it must be." She shivered.
"To be buried under all those tons and tons of—"
"Immurement is security," Mirrh Yahn y Cona said. "The ultimate
stability."
"You'll get acclimatized," Ellis said. "It takes time."
He broke off to peer through the gloom beyond the dance floor.
"Good Lord, there's Ryerson of the Post, camera and all. If he
recognizes me he'll know who Mirrh is and—"
"Yahn," Mirrh Yahn y Cona corrected automatically. "With us the
second name is impersonal. First is used only by loved ones."
"Yahn, then," Ellis said. "If Ryerson tumbles, he'll want pictures and
an interview. Yahn will be lionized before he's ready. Can't publicize
him until he knows the ropes."
"You'd better skip," Leila said. "If we all go, he'll spot us for sure."
"Right." Ellis shoved some money at Leila. "Call me at my office when
it's safe."
When Ellis had gone and their waiter had brought drinks, they faced
each other across the table, Yahn visibly on guard and Leila with the
beginning of speculation in her eyes.
"Maybe it's better like this, without protocol," she said. "Yahn, can you
—do you dance in our gravity?"
He was bitter again. "Remember my training. I am taller, stronger and
more freakishly agile than any Martian—including my fiancee—has
been for thousands of years."
Her clear look made him ashamed and he added, "With us the dance
is an art form only. Here the intent seems different."
"It is," Leila said almost grimly. "Finish your drink, Buster. You're going
to need it."
He needed several before the evening was finished. The Terran

dance in its limited variations offered small challenge; Yahn mastered
it with an ease that delighted Leila and brought tacit envy from other
couples. The cocktails may have contributed to his own mixed
reactions, lending primitive tactility to Leila's pliant response.
Neither of them, when Ryerson of the Post went away with his
camera, considered calling Ellis.
"I don't often enjoy my work so much," Leila said. "Let's not spoil the
evening with diplomacy, shall we?"
They left the Diplomatic vehicle for Ellis, rented an agency car and
drove through the charged serenity of the night into the mountains.
They talked the Moon down and the Sun up. Nothing took place that
might have shocked a reasonably tolerant duenna, but by dawn they
had reached the sort of understanding that comes spontaneously or
not at all.
"The biologists who tailored me to Terrestrial standards," Yahn said,
"did their work too well. I find myself more Terran than Martian."
The immovable obstacle, of course, was Yahn's obligation to Yrml,
who would be waiting with enduring Martian patience for his return.
Leila went into that matter later with Ellis, not so much to enlist his
dubious sympathies as to clarify the bristly problem in her own
troubled mind.
"Martians use our broadcasts as a standard of judgment," Leila said.
"And you know where that leads. The more prominent the people in
the newscasts, the higher the divorce rate. The more popular a video
serial, the greater its emotional shilly-shallying. To Martians we're the
last word in fickleness."
"I know," Ellis agreed. "Our cultural geometry was always triangular."
"Exactly. So how can Mirrh-Yahn break the news to his dry little
fiancee back home? We're accustomed to inconstancy and to
incontinence. We sing corny songs about girls who write jilting letters
to their men in service. Our opera flaunts Perkinses and Mesdames
Butterfly, and the fact that we enjoy them shocks the ascetic pants off
the Martians. Did you know that their population control quota
demands a strictly equal sex-ratio, so that there's never more than
one boy for one girl from the beginning? Mirrh-Yahn simply hasn't it in
him to leave Yrml dangling. He'd feel a renegade for the rest of his
life."
"Mirrh-Yahn," Ellis noted. "Obviously he's willing enough, if you're on
a first-name footing."
"I can't call him Yahn any longer, like a stranger. Mirrh-Yahn is a
compromise."
Ellis rummaged in his desk and brought out a personnel folder.
"Dossier on J. Frederic Thomas, our young man on Mars. Maybe we
can turn up an angle through him."
The exchange ambassador's folder was neither interesting nor
helpful. J. Frederic Thomas stood revealed as a dwarfish scholastic
type, complete with massive glasses and receding hairline.
"He looks more Martian than Terran," Leila said. "Is that deliberate?"
"Mars sent us a man specially bred to fit into our culture, didn't they?
Simple job here to turn up a Martian type. Matter of fact, J. F.'s
reports show he fits in up there like a native."
"Check with him, then," Leila said. "Though I can't imagine what help
we can expect from a wizened little stick like that."
Leila was wrong. J. Frederic Thomas—who quite predictably, being

paired off with the only unattached female on Mars as his cicerone,
had immediately found himself caught in the same thorny dilemma
that gouged his opposite number on Earth—was eager to help. The
result of Ellis' inquiry was a swift letter from Yrml Orise y Yrl to Mirrh
Yahn y Cona; a letter which Ellis turned over in duplicate, one in
Martian ideograph, the other a translation, to Leila.
It broke Yrml's engagement to Yahn for the excellent reasons that J.
Frederic Thomas was not only more Martian in physique and
deportment, but also possessed a fine reedy tenor which blended
ever so better with Yrml's soprano in the less poignant duets from the
Tchulkione Serafi.
"The man never lived," Ellis pointed out, "Martian or Terran, no matter
how relieved he might be, whose ego wouldn't need attention after a
letter beginning Dear Yahn. Shall I let it go on through the mails, or
will you—"
Leila answered him on her way out. "Don't bother," she said.
THE END
*** END OF THE PROJECT GUTENBERG EBOOK INCONSTANCY
***
Updated editions will replace the previous one—the old editions will
be renamed.
Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying copyright
royalties. Special rules, set forth in the General Terms of Use part of
this license, apply to copying and distributing Project Gutenberg™
electronic works to protect the PROJECT GUTENBERG™ concept
and trademark. Project Gutenberg is a registered trademark, and
may not be used if you charge for an eBook, except by following the
terms of the trademark license, including paying royalties for use of
the Project Gutenberg trademark. If you do not charge anything for
copies of this eBook, complying with the trademark license is very
easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.
START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK
To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.
Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund from
the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.
1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law in
the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name
associated with the work. You can easily comply with the terms of
this agreement by keeping this work in the same format with its
attached full Project Gutenberg™ License when you share it without
charge with others.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the terms
of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.
1.E. Unless you have removed all references to Project Gutenberg:
1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears, or
with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.
1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.
1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning of
this work.
1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.
1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1 with
active links or immediate access to the full terms of the Project
Gutenberg™ License.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or expense
to the user, provide a copy, a means of exporting a copy, or a means
of obtaining a copy upon request, of the work in its original “Plain
Vanilla ASCII” or other form. Any alternate format must include the
full Project Gutenberg™ License as specified in paragraph 1.E.1.
1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.
1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”
• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt that
s/he does not agree to the terms of the full Project Gutenberg™
License. You must require such a user to return or destroy all
copies of the works possessed in a physical medium and
discontinue all use of and all access to other copies of Project
Gutenberg™ works.
• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.
1.F.
1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.
1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except

for the “Right of Replacement or Refund” described in paragraph
1.F.3, the Project Gutenberg Literary Archive Foundation, the owner
of the Project Gutenberg™ trademark, and any other party
distributing a Project Gutenberg™ electronic work under this
agreement, disclaim all liability to you for damages, costs and

Language Change Variation and Universals Culicover Full Chapter

Uploaded by

Document Information

Original Title

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

Language Change Variation and Universals Culicover Full Chapter

Uploaded by

Copyright:

Available Formats

Language Change, Variation, and

Language Change, Variation, and Universals

4.2 Constructional innovation 75

7.3 Scope in situ 177

10.3 Beyond parameters: Capturing the style 262

It is of course possible to formulate descriptions of change, variation, and

conceptual structure, and grammatical universals and typological patterns

Chapter 8 (Constructional change in Germanic) tracks several of the major

xvi list of abbreviations

1.1 The problem

(i) There is a set of very general grammatical principles, structures, and

significance, except insofar as it counts as evidence about I-language,

1.1 the problem 5

In contrast, I assume that grammars are not computationally optimal in

(i) Regarding the core, I assume that it is grounded in conceptual struc-

(iii) Regarding variation, I assume that anything that can possibly be

The challenge posed by Chomsky’s Problem is a fundamental one. How is

either there is an infinity of options, in which case challenging and perhaps

1.1 the problem 7

These intuitions lead to the characterization of a grammar as consisting of

This section summarizes the essential features of the constructional approach

First, the basic architecture of a constructional approach. I assume, following

To illustrate with a concrete example, there is a construction in English

(1) a. Chris does not [VP like pepperoni pizza].

Thus, the function of a grammar is to state what properties a construct must

(2) ⎡phon ΦEnglish (kick)=/kɪk/ ⎤

Crucially, when the verb has particular morphosyntactic properties in a

(3) ΦEnglish (kick,3,sg,pres) = /kɪks/

(4) ΦEnglish (be,3,sg,past) = /wəz/

The precise formal characterization of Φ or an equivalent alternative is not

Finally, as in Relational Grammar (Perlmutter 1983; Perlmutter & Rosen

1.2.2 Constructions are not derivations

In order to appreciate the constructional approach, it is useful to contrast it

(i) MGG licenses linear ordering by moving constituents from canonical

(iv) MGG is rigid in its licensing conditions; a constructional grammar is

‘cryptoconstructionalism’ (Culicover 2017), and will argue as we proceed that

minating perspective on some (but often different) fundamental property or

• MGG is right about the existence of a universal faculty that underlies

2.2 What a grammar is for

In this section I summarize how a constructional approach characterizes what

2.2 what a grammar is for 17

To take a simple example, Sandy snores is well-formed in English because the

‘Transfer possession’ in this case is part of the interpretation in virtue of the

(1) a. Sandy gave Chris the money.

On a constructional approach, then, well-formedness applies not just to

2.3 a framework for constructions 19

constructions with licensing conditions of varying generality. It is sufficient

2.3 A framework for constructions

2.3.1 Representing constructions

Following the Parallel Architecture of Jackendoff (2002), I take a construction

⁴ A complication that I set aside here is the proper treatment of prosody.

To illustrate the kinds of constructions that can be defined in these terms,

I assume for convenience that the interpretation consists of a primitive

2.3 a framework for constructions 21

For constructions that state correspondences between categories in syn and

Finally, many constructions involve complex correspondences between

Now let’s look at idioms. The representation of the multiword expression

(7) Construction: by and large

(8) kick the bucket

2.3 a framework for constructions 23

(9) sell NP down the river

There is no phon defined for this construction, because it is built around a

This ordering is the responsibility of a construction or set of constructions that