Professional Documents
Culture Documents
(Cognitive Linguistics Research, 19.1) Susanne Niemeier, Martin Pütz - Theory and Language Acquisition-De Gruyter Mouton (2001)
(Cognitive Linguistics Research, 19.1) Susanne Niemeier, Martin Pütz - Theory and Language Acquisition-De Gruyter Mouton (2001)
(Cognitive Linguistics Research, 19.1) Susanne Niemeier, Martin Pütz - Theory and Language Acquisition-De Gruyter Mouton (2001)
WDE
G
Cognitive Linguistics Research
19.1
Editors
René Dirven
Ronald W. Langacker
John R. Taylor
Mouton de Gruyter
Berlin · New York
Applied Cognitive Linguistics I:
Theory and Language Acquisition
Edited by
Martin Pütz
Susanne Niemeier
René Dirven
Mouton de Gruyter
Berlin · New York 2001
Mouton de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter GmbH & Co. KG, Berlin
© Copyright 2001 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin
All rights reserved, including those of translation into foreign languages. No part of this book
may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopy, recording, or any information storage and retrieval system, without
permission in writing from the publisher.
Printing: WB-Druck, Rieden/Allgäu
Binding: Lüderitz & Bauer, Berlin
Printed in Germany
Acknowledgements
Furthermore, our sincere thanks go out to the authors, who have re-
sponded with professionalism to all the requests that have been made
of them. In this regard, we would also like to express a great debt of
gratitude to the expertise of the many scholars who acted as our refe-
rees: Angeliki Athanasiadou, Frank Boers, Willis Edmondson, Carlos
Inchaurralde, Dirk Geeraerts, Stefan Gries, Peter Grundy, Juliane
House, Bernd Kortmann, Penny Lee, Lienhard Legenhausen, Bert
Peeters, Mechthild Reh, Sally Rice, Francisco Ruiz de Mendoza,
Doris Schönefeld, Rainer Schulze, Elzbieta Tabakowska, Jef Ver-
schueren, Marjolijn Verspoor, Helmut Vollmer, Michael Wendt,
Karin Wenz.
vi Acknowledgements
Above all, we want to thank Birgit Smieja, who did a marvelous job
in designing the layout of the book and in taking care of the laser
printout.
The Editors
Duisburg, Bremen, and Landau July 2001
List of Contributors
Jenny Cook-Gumperz
Vyvyan Evans
CliffGoddard
Paul J. Hopper
Istvan Kecskes
Amy Kyratzis
Acknowledgements ν
Introduction xiii
Martin Pütz, René Dirven and Susanne Niemeier
Subject Index
Contents of volume II
Acknowledgements ν
Introduction xiii
Martin Pütz, Susanne Niemeier and René Dirven
The main title of the two volumes, i.e. Applied Cognitive Linguistics,
not only focuses on the theory of cognitive linguistics as it can be
applied to the teaching and learning of foreign languages; it also en-
compasses psycholinguistic models and theories with a focus on first
language acquisition. Thus, Volume I is concerned with the interac-
tion between language, cognition and acquisition in general (first and
second language acquisition), and Volume II addresses a series of
cognitive principles of linguistic, i.e. conceptual organization while
acquiring, learning and teaching second or foreign languages.
The two volumes address a number of important topics in the the-
ory, acquisition and pedagogy of languages seen from the perspective
of cognitive linguistics. We cannot discuss here in detail the main
principles, aims and findings of the cognitive enterprise, but would
nevertheless like to refer to the following assumptions and implica-
tions of the discipline of cognitive linguistics as set out by Rudzka-
Ostyn (1993: 1):
The first three contributions deal with an account of the use of the
English tense system seen from various perspectives. In his paper
"Cognitive linguistics, language pedagogy, and the English present
tense" Ronald W. Langacker examines the English present tense, as
well as related phenomena like the progressive and the per-
fect/imperfect contrast, from the perspective of his well-known the-
ory of cognitive grammar. Although Langacker does not specifically
focus on issues of second language acquisition as such, he refers to
several pedagogical implications of cognitive linguistic theory and
possible directions for further research. Some of these directions
follow from the usage-based nature of cognitive grammar, in which
linguistic units are seen as being abstracted from usage events. Given
the fact that regular constructions of full generality constitute only a
small proportion of conventional patterns, Langacker assumes that
complete mastery of linguistic rules does not assure any degree of
actual fluency in a language. Furthermore, Langacker underlines the
importance of the communicative, social and cultural context of any
speech encounter and refers to the notion of 'construal', i.e. "our
ability to conceive and portray the same situation in alternate ways".
In conclusion, Langacker criticizes the traditional way of looking at
tense mainly on the grounds that these approaches have been objec-
tivist in nature. They lack an awareness of 'construal' and viewing
arrangements and the subjective basis of factors like homogeneity
and bounding. These are the two major viewing arrangements that we
impose on all phenomena. In the domain of things we can impose a
homogeneous view and conceptualize phenomena as homogeneous,
unbounded substances, or else as bounded objects with an internal
heterogeneous structure. This distinction corresponds in the language
to that between mass nouns and count nouns. Similarly in the domain
of processes, we impose a homogeneous arrangement and conceptu-
alize homogeneous states or activities, or else we impose a bounded
view and conceive of processes as bounded events. This distinction
corresponds in the verb system to imperfective tense forms (progres-
sive or habitual form) or to perfective, i.e. bounded tense forms. This
major insight may stimulate many language pedagogues and psycho-
linguists to rethink their entire language pedagogy or acquisition
xvi Martin Pütz, René Dirven and Susanne Niemeier
sal semantic primitives are carved out of natural language and also
have an inherent 'conceptual syntax'. Goddard's paper "Conceptual
Primes in Early Language Development" tackles the problem that
young children in their acquisition process do not start off with the
seventy odd semantic primitives, but with more complex concepts.
The author therefore explores various hypotheses about the nature
and identities of the innate concepts which may underpin language
acquisition. Since many of the child' s early words are more complex
and cannot be explained in terms of conceptual primes, a possible
explanation might be that these words have simpler meanings than
the corresponding words in an adult's vocabulary. Goddard con-
cludes from this that the child must have a certain 'conceptual vo-
cabulary' of prime concepts prior to their emergence in language
proper. In the remaining part of the paper Goddard engages in a de-
tailed semantic analysis of the child's production vocabulary (i.e. the
non-prime words and utterances) in order to investigate the concep-
tual primes which may be 'latent' in a child's early lexicon. An inter-
esting observation refers to the fact that the production vocabulary of
primes lags several months behind the conceptual vocabulary. As
Goddard's study is based on the language development of one single
child, i.e. Goddard's son Peter, it would be interesting to see to what
extent Peter's acquisition story is representative for other children's
production vocabulary and their acquisition of primes in the concep-
tual vocabulary.
Whereas Goddard in general investigates the development of the
linguistic and conceptual vocabulary of child language, Katharina J.
Rohlfing's paper "No preposition required. The role of prepositions
for the understanding of spatial relations in language acquisition" is a
report on experiments focusing on the role of one single syntactic
category, i.e. prepositions, for the understanding of spatial relations
also in first language acquisition. Rohlfing presents the results of an
empirical study of Polish-speaking children at the age of 2. The hy-
pothesis of her study was that, in prototype situations, children un-
derstand instructions even without prepositions, which however they
need in order to understand the instructions in abstract situations. The
experiments show that infants rely not only on the appropriate loca-
xxii Martin Pütz, René Dirven and Susanne Niemeier
Notes
1. Problems of the prototype approach are also pointed out in CL, e.g. by Geera-
erts (1988). On the other hand, Hopper and Thompson (1985) themselves in-
voke prototype theory for typology. Without calling himself a cognitive linguist,
Kastovsky (1988) opposes prototype semantics as a superior alternative to
structural semantics. One example out of the hundreds of non-cognitive lin-
guists working with prototype ideas is Shaver et al. (1987).
2. Peeters (2001) devotes a whole paper to the - in his view - misnomer 'cognitive
linguistics' for the CL paradigm, which he believes should call itself 'Cognitive
Linguistics' with upper case, leaving the lower case appellation 'cognitive lin-
guistics' to its inventor, i.e. Sydney Lamb.
References
Dirven, René
1989 Cognitive linguistics and pedagogic grammar. In: Gerhard Leitner
and Gottfried Graustein (eds.), Linguistic Theorizing and Gram-
mar Writing, 56-75. Tübingen: Niemeyer.
Dirven, René and John R. Taylor
1994 English modality: A cognitive-didactic approach. In: Keith Car-
Ion, Kristin Davidse and Brygida Rudzka-Ostyn (eds.), Perspec-
tives on English. Studies in Honour of Professor Emma Vorlat,
542-556. Leuven: Peeters.
Geeraerts, Dirk
1988 Prototypicality as a prototypical notion. Communication & Cog-
nition 21: 343-355.
Hopper, Paul J. and Sandra A. Thompson
1985 The uses of the notion 'prototype' in typological studies. In: H.-J.
Seiler and G. Brettschneider (eds.), Language Invariants and
Mental Operations, 238-244. Tübingen: Gunter Narr.
Kastovsky, Dieter
1988 Structural semantics or prototype semantics? The evidence of
word-formation. In: Werner Hüllen and Rainer Schulze (eds.),
Understanding the Lexicon. Meaning, Sense and World Knowl-
edge in Lexical Semantics, 190-203. Tübingen: Niemeier.
Langacker, Ronald W.
2000 Grammar and Conceptualization. Berlin/New York: Mouton de
Gruyter.
xxiv Martin Pütz, René Dirven and Susanne Niemeier
Ronald W. Langacker
1. Introduction
Without my telling you the context, you could probably not deter-
mine the specific import of (1) or of the phrase petit cordon. Con-
versely, prior to the utterance - having been trained in French in a
classroom setting by traditional methods - I had no clue whatever
about how to describe the situation or the loop of fabric. Yet this one
utterance, immediately interpretable in context, was sufficient for me
to learn a natural way of expressing this situation in French. (Admit-
tedly, I have never since found myself in a situation where this
knowledge was useful, but if it ever happens again I am ready.)
Once more, I will not presume to make specific pedagogical rec-
ommendations. Let me simply state the obvious, namely that these
considerations argue for pedagogical approaches which emphasize
the interactive exposure to large quantities of natural speech in con-
text.
3. Construal
(3) "[L]et's ask whether each part of speech really denotes a con-
sistent kind of meaning ... Now it is true that any word that
names an object will be a noun. But on the other hand, not
every noun names an object. 'Earthquake' names, if anything,
an action, as does 'concert'; 'redness' and 'size' name proper-
ties; 'place' and 'location' pretty obviously name locations. In
fact, for just about any kind of entity we can think of, there ex-
ist nouns that name that kind of entity. So the grammatical no-
8 Ronald W. Langacker
(4) thing > animal > dog > retriever > golden retriever
Cognitive linguistics, language pedagogy, and the English present tense 9
ego
MS/IS
aunt
The mistake here lies in assuming that the bounding of a process has
to be objectively given, with an inherent endpoint observable in the
situation itself. Ultimately, what counts for linguistic purposes is
whether a process is conceptualized as some kind of bounded epi-
sode, irrespective of whether a natural endpoint is discernible. There
is in general a strong tendency to conceptualize force-dynamic occur-
rences (those requiring the expenditure of energy - cf. Talmy 1988)
as being bounded in duration, even when the process is internally
homogeneous and nothing appears to be going on. Thus (8a) de-
scribes a stable situation, but since the stability results from a balance
of opposing forces, the basic process (i.e. the dam contains the surg-
ing floodwaters) is construed as an episode of bounded duration,
hence the progressive is possible. By contrast, in (8b) the basic proc-
ess (i.e. the barrel contains water) is non-force-dynamic. It merely
describes a spatial configuration, which as such can maintain itself
14 Ronald W. Langacker
English has many perfective predicates (e.g. sleep, dream, run, walk,
sit, stand, lie, perspire, talk, chat, meditate, wear a tie) in which an
internally homogeneous activity is nonetheless construed as occur-
ring in bounded episodes. It is only their grammatical behavior - re-
sisting the simple present, occurrence in the progressive - which
alerts us to their perfectivity. This does not however imply that the
distinction is "purely grammatical", with no conceptual basis. The
grammatical classification hinges on a conceptual factor which is no
less real for being subject to construal.
Once the specific, subtly contrasting construals imposed by lexical
and grammatical elements are elucidated, innumerable structural de-
tails that otherwise seem quite arbitrary turn out instead to have an
intuitively graspable conceptual basis. The potential thus exists for
devising effective ways of teaching them. How much of this should
be explicitly taught to students? Should we burden the language
learner with technical constructs like profile and immediate scope? I
suspect not. It is not even evident that the classroom teacher should
be responsible for such technical details, which might better be
brought in at the level of overall planning and design of teaching
materials. I will leave this matter for those competent to assess it.
At the same time, these notions might find a natural place in a
language arts curriculum. Generative grammarians have often pro-
posed that linguistics ought to be more visible in school curricula at
all levels. In particular, they put it forth as a way of teaching science:
formulating hypotheses, testing them against the empirical evidence,
developing skills of argumentation, etc. While this may have some
merit, the very different vision of language embodied in cognitive
linguistics suggests another option: the concepts and descriptions of
cognitive linguistics might instead be used for inculcating an appre-
ciation of language as a means of evoking and symbolizing alterna-
Cognitive linguistics, language pedagogy, and the English present tense 15
4. Conceptual substrate
We rely on information from any or all of these strata even for the
interpretation of seemingly straightforward expressions. Let us once
again consider example (1 ), Il n'y a pas de petit cordon. This made
perfect sense given the situational context of wanting to hang up a
jacket, but without this contextual support it seems rather pointless
and hard to interpret. Apprehension of the situational context was
itself dependent on the knowledge of certain cultural practices, nota-
bly that of hanging up jackets and the usual provision of a loop of
fabric for this purpose. Moreover, certain basic properties of the
world we inhabit - such as temperature, gravity, and force dynamics
- are reflected in these cultural practices and implicitly invoked any
time we think of them.
A pivotal aspect of the conceptual substrate, quite clearly, is the
ground, comprising the speaker, the hearer, their interaction, and the
immediate circumstances. It is quite common - arguably even ca-
nonical - for the ground to remain offstage and not be mentioned. In
the unmarked situation, the ground functions as the tacit location
from which a scene is viewed and an expression's meaning is appre-
hended, as opposed to being onstage as the explicit focus of attention.
For instance, a tense marker locates a profiled process with respect to
the time of speaking (one facet of the ground), invoking it as a tem-
poral reference point, but does not directly mention it. Likewise, a
determiner - via its specification of (in)definiteness - invokes the
speaker and hearer as the individuals seeking to identify the nominal
referent, but leaves them offstage and unprofiled. The ground, then,
is the locus of conception. We do of course have ways of putting fac-
ets of the ground onstage and referring to them specifically, e.g. with
forms like I, you, here, and now. Even so, the ground's occasional
status as focused target of conception coexists with (and is subsidiary
to) its more fundamental role as the tacit locus of viewing.
Inherent in every usage event is a presupposed viewing arrange-
ment, pertaining to the relationship between the conceptualizes and
the situation being viewed. The default arrangement finds the speaker
and hearer together in a fixed location, from which they report on
actual occurrences in the world around them. There are however nu-
merous kinds of departures from this canonical circumstance. The
Cognitive linguistics, language pedagogy, and the English present tense 17
(9) The telephone poles are rushing past at ninety miles per hour.
(12) a. The brain tumor robbed him of the chance to finish his
novel. [metaphor]
b. The pen is mightier than the sword. [metonymy]
c. The living room is much more comfortable, [implicature]
In (13b), both the subject and the profiled event are fictive in nature.
The trees does not refer to any actual set of objects, but is rather a
role description, designating a feature of the landscape observable at
any altitude. Nor does any tree or set of trees actually change in
length. What the sentence describes is a virtual change generated by
viewing the trees instantiating the role at different altitudes as if they
were a single, changing entity.
Finally, let me mention the frequent but seldom noted phenome-
non of resorting to type specifications as a way of describing a set of
actual occurrences that are alike in some respect. Imagine a series of
actual events in each of which a single stranger - different each time
- reaches over a fence and picks a single apple - also different each
time. If there are three such events, the entire sequence can be sum-
marized by sentence (14a). It is not essential for our purposes that the
sentence is subject to alternate interpretations. What is essential is the
possibility of using (14b) for exactly the same event sequence. There
are three different strangers, and three different apples, yet these par-
ticipants are referred to in the singular: a stranger, an apple.
(14) a. Three times, strangers reached over the fence and picked
apples.
b. Three times, a stranger reached over the fence and picked an
apple.
I suggest that (14b) describes the events at the level of their common
type characterization. The three actual events each instantiate the
event type a stranger reach over the fence and pick an apple. At the
type level, representing what the three events have in common, there
is only one stranger and one apple. These are instances of the
stranger and apple categories, but they are not actual instances or
Cognitive linguistics, language pedagogy, and the English present tense 21
The one thing that is generally agreed upon concerning the English
present tense is that it is really not a present tense, i.e. its value can-
not be that of indicating that the process in question occurs at the
time of speaking. The arguments seem straightforward. On the one
hand, events that do occur at the time of speaking generally cannot be
expressed in the present tense. This is the case with perfectives, as
we saw earlier. As descriptions of actual, bounded events occurring
22 Ronald W. Langacker
t ΚΛΛΛΛΛ
t wvWw ^ w w v
Observe that in the past tense there is no inherent limit on the length
of the immediate scope, so a perfective process of any length can
always be made to fit inside it. By contrast, in the present tense the
immediate temporal scope must be the same in duration as the speech
event. If an imperfective process endures for a span of time that in-
cludes the immediate scope (in either the past or the present), only
that portion of it subtended by the immediate scope is profiled. Re-
call that an expression's profile is necessarily confined to its immedi-
ate scope (the general locus of attention). Moreover, since an imper-
fective process is internally homogeneous and not characterized in
terms of bounding, any subpart singled out for profiling will itself
constitute a valid instance of the process type in question. (Imperfec-
tives are quite analogous to mass nouns in this respect.)
Let me start by pointing out how much this analysis accounts for,
straightforwardly and even rather elegantly. First, it accounts for im-
perfectives being able to occur in the present tense, as sketched in
Figure 5d. Since any part of an imperfective process itself counts as a
full instantiation of the process type, this will also be true for the
portion that coincides with the time of speaking. Observe that the
analysis does not imply that the stable situation described is valid
only for the brief duration of the immediate scope. For instance, an
utterance of (5b), He knows the truth, does not entail that his knowl-
24 Ronald W. Langacker
edge of the truth is limited to the time of the utterance. What is being
claimed, instead, is that the speech event defines a "window" for fo-
cused viewing. In using a present tense imperfective, the speaker is
taking a temporally coincident sample of the overall situation and
observing that - for the portion sampled - the situation is stable and
unbounded. It is possible, without contradiction, for the same overall
situation to be sampled at different times, as in (17):
vvvvv — ^ -
l
Figure 6. Tense and the progressive
The analysis also accounts for the usual infelicity of present tense
perfectives, as in (15a). The difficulty is not however a matter of
conceptual incoherence. Indeed, the conceptual configuration de-
picted in Figure 5b is perfectly coherent and non-anomalous. There is
nothing inherently contradictory about a bounded event temporally
coinciding with the speech event. And indeed, true present tense per-
fectives are sometimes permissible, as we will see. Instead, the
problem with present tense perfectives is that certain factors make the
configuration in Figure 5b hard to achieve in practice.
A perfective process is bounded, so a full instantiation of such a
process includes its boundaries. Thus, if a perfective process is to
coincide with the time of speaking, its beginning point has to coin-
cide with the initiation of the speech event, and its endpoint with its
termination. This poses both a durational problem and an epistemic
one. The durational problem is that there is no inherent connection
between the length of the event described and the length of the
speech event describing it. It takes longer to paint a fence, for exam-
ple, than it does to utter the clause He paints a fence. The epistemic
problem resides in having to observe an event in order to identify it
as a prerequisite to describing it. By the time an event is observed
and identified, it is already too late to initiate a speech event that pre-
cisely coincides with it. These problems do not arise with imperfec-
tives, given their mass-like character and the property that any por-
tion of the overall process counts as a full instantiation of the process
type. Hence an imperfective has no specific duration, and a portion
26 Ronald W. Langacker
Performative
MS
is]
, ΙΛΛΛΛΙ ^
Figure 7. Performatives
can imagine, a child is playing with toy cars and a play village and
accompanies each action she takes with a descriptive sentence. The
successive utterances in (21) then coincide with the successive acts of
pushing a toy car from one place to another. This use of the present
tense for perfective events seems perfectly natural and unproblem-
atic.
(21) Now I drive to work. Now I go to the store. Now I drive home.
More generally, the present tense is naturally used for the narration
of demonstrations. Imagine an origami class, where each clause in
(22a) accompanies the action it describes:
Note that the events described in this way have approximately the
right duration for temporally coincident description. In the context of
a sporting event, they are also quite stereotyped, so the announcer has
a good idea of what is likely to transpire at any instant. One is there-
fore able to shadow the events fairly closely, sometimes even to an-
ticipate and describe them simultaneously with their occurrence. The
goal at least is to come as close as possible to coincident description,
and the conventions of play-by-play reporting rest on either the fic-
tion that this is feasible or else the tolerance of a certain time-lag.
Undoubtedly we have to recognize flexibility and degrees of ap-
proximation in what counts as "precise coincidence". These are, after
all, matters of construal rather than objective scientific measurement.
Yet this hardly seems adequate for examples like the final sentence in
(22b). Moreover, I can easily imagine an alternative mode of narra-
tion for a demonstration, where - instead of coinciding with it - each
statement precedes the action it describes, thereby telling the listener
what to watch for. At least for cases like these, we appear to need
some other approach. The approach I suggest is to posit a distinct
viewing arrangement, one that does not specifically involve the si-
multaneous narration of actions. Although, in practice, the events in
question are correlated with actions, they are conceptualized more
30 Ronald W. Langacker
A virtual schedule pertains to the future, but its own status and loca-
tion are another matter. When a plan is in effect, the schedule itself is
stable and mentally accessible through a span of time that includes
the present. The schedule consists of virtual events, which are repre-
sentations of anticipated actual events. Moreover, the time interval
through which each virtual event is conceived as unfolding is identi-
fied with a particular time in the future, as shown by the dotted corre-
spondence lines in Figure 8. However, the events constituting the
schedule are only virtual.
Actual /
hvents I t
/
Figure 8. The scheduled future
/
/ Γ-
1 Processai—^ Process2| /
^ /Virtual
Α ,F
/ ,Λ w
w ΛΛΑ,
w
the situation coded by Process2 does not obtain. When applied to the
future, where the course of reality has not yet been determined so that
Processi is necessarily non-actual, presenting it for consideration has
the effect of suggesting that its actual occurrence (and the subsequent
occurrence of Process2) may be quite imminent.
Finally, I should mention the use of present tense in certain types
of subordinate clauses, as in (28):
(28) {If / when / until / before / after / while} you make a decision,
you should consider all your options.
Despite its present tense form, in each case the predicate {make a
decision) refers to a process envisaged as occurring in the future.
Here I basically follow the analysis proposed by Fauconnier (1997) in
terms of mental spaces. The subordinators introducing these clauses
are space builders: //establishes a hypothetical space, and the others
set up spaces defined by their temporal location. They further shift
the viewpoint to the space they establish. In other words, they incor-
porate the instruction to adopt a special viewing arrangement in
which the clausal content is apprehended from a temporal vantage
point other than the actual time of speaking. Its apprehension does of
course occur at the time of speaking, but when this is Actively identi-
fied with the time span internal to the mental space, a process con-
ceived as occurring in this space is fictively viewed as coincident
with the speech event.
6. Conclusion
For sake of discussion, let us suppose that you agree. What then fol-
lows? What conclusions can we draw for language pedagogy?
I am of course eschewing any specific pedagogical proposals. I
will however suggest that the traditional way of looking at tense,
even in linguistics, engenders confusion by obscuring its basic nature.
Standard discussions are objectivist in spirit. They ignore construal
and the subjective basis of factors like homogeneity and bounding.
They have no conception of the myriad viewing arrangements that
mediate between objective circumstances and the formulation of lin-
guistic expressions. Thus they attempt to account for tense directly in
terms of the temporal relation between the actual time of speaking
and the full duration of an envisaged actual occurrence. They do this
even when - according to the analysis presented here - the process
being viewed and temporally located is only a portion of the actual
occurrence (notably with imperfectives), or else a virtual process
connected to it in a manner specified by the viewing arrangement. It
is no wonder, then, that a cogent description remains elusive, and that
the present tense is claimed to be anything but a present tense.
The consideration of pedagogical issues can only be aided by an
accurate understanding of what is being taught. In the case of lan-
guage, unfortunately, traditional and modern understandings are usu-
ally far from adequate, even for things as fundamental as the present
tense. It is premature to suggest that cognitive linguistics is coming to
the rescue. I do however see it as a positive development, providing
new and revealing perspectives on specific problems as well as our
overall conception of language and how it relates to culture, cogni-
tion, and social interaction. In short, I think we are starting to get a
real grip on how things work. If so, it should eventually give rise to
successful pedagogical applications, which will lend it empirical
support.
38 Ronald W. Langacker
References
Austin, J. L.
1962 How to Do Things with Words. Cambridge, MA: Harvard Uni-
versity Press.
Brisard, Frank
1999 A critique of localism in and about tense theory. Ph.D. disserta-
tion, University of Antwerp.
Fauconnier, Gilles
1997 Mappings in Thought and Language. Cambridge: Cambridge
University Press.
Goldberg, Adele E.
1995 Constructions: A Construction Grammar Approach to Argument
Structure. Chicago and London: University of Chicago Press.
Goldsmith, John and Erich Woisetschlaeger
1982 The logic of the English progressive. Linguistic Inquiry 13: 79-
89.
Jackendoff, Ray
1994 Patterns in the Mind: Language and Human Nature. New York:
Basic Books.
Langacker, Ronald W.
1968 Language and its Structure: Some Fundamental Linguistic Con-
cepts. New York: Harcourt, Brace & World.
1987a Foundations of Cognitive Grammar, Vol.1, Theoretical Prereq-
uisites. Stanford: Stanford University Press.
1987b Nouns and verbs. Language 63: 53-94.
1988 A usage-based model. In: Brygida Rudzka-Ostyn (ed.), Topics in
Cognitive Linguistics, 127-161. (Current Issues in Linguistic
Theory 50.) Amsterdam/Philadelphia: John Benjamins.
1990 Concept, Image, and Symbol: The Cognitive Basis of Grammar.
(Cognitive Linguistics Research 1.) Berlin/New York: Mouton de
Gruyter.
1991 Foundations of Cognitive Grammar, Vol.2: Descriptive Applica-
tion. Stanford: Stanford University Press.
1993 Universals of construal. Proceedings of the Annual Meeting of
the Berkeley Linguistics Society 19: 447-463.
1995 Viewing in cognition and grammar. In: Philip W. Davis (ed.),
Alternative Linguistics: Descriptive and Theoretical Modes, 153-
212. (Current Issues in Linguistic Theory 102.) Amsterdam/
Philadelphia: John Benjamins.
Cognitive linguistics, language pedagogy, and the English present tense 39
1. Introduction
And so begins one of the most self consciously literary, yet most fre-
quently reissued children's books of the last 70 years. In this the
original beginning to the story, since adapted by Disney and others
into a universal tale of a "cute little bear", the literary devices of
tense/aspect provide a specific set of what Langacker (this volume)
calls the viewing arrangements that enable participants and read-
ers/hearers to shift perspectives on the story events. The following
text adds a further dimension to the tellers', participants', and
reader/hearers' perspectives, when A. A. Milne inteqects his own
character/voice: "When I first heard his name, I said, just as you are
going to say, 'But I thought he was a boy?'"
The author here enters the story briefly as a protagonist, that is an
authorial "I" telling how the tale came to be told. After this interlude
the story proper begins in the traditional story telling past tense, the
third person preterite (Benveniste 1971). "Once upon a time, a very
long time ago now, about last Friday, Winnie-the-Pooh lived in a
forest all by himself under the name of Sanders." We cite these, and
42 Jenny Cook-Gumperz and Amy Kyratzis
later examples from the Dr. Seuss books, to illustrate the kinds of
narrative experiences children participate in before they learn to read
for themselves. In attending to the often told/read tales, they are ex-
posed to linguistic constructions that can have complex semantic
implications such as signaling shifts in perspective. Through tense/
aspect contrasts, children's stories are able to provide succinct infor-
mation about the characters, the tale's context and the relationship of
the teller to the protagonists that make simply told tales dependent on
complex conceptualizations for their interpretation. As Langacker
argues: "Inherent in every usage event is a presupposed viewing ar-
rangement, pertaining to the relationship between the conceptualizers
and the situation being viewed" (16).
The progressive, as described by Langacker, for example, is dis-
tinct, not for its focus on the present time but for the viewing ar-
rangement it imposes: allowing the "zooming in" and "taking an in-
ternal view" of a bounded event (12). In story telling, and in literary
contexts, such a viewing arrangement is not appropriate; the events
must be told as though the narrator can view their endpoint. In narra-
tivizing discourse, then, to quote White's seminal paper on Narrativ-
ity: "we can say, with Benveniste, 'Truly there is no longer a 'narra-
tor'. The events are chronologically recorded as they appear on the
horizon of the story. Here no one speaks. The events seem to tell
themselves" (White, quoting Benveniste 1981). Here the events are
told as though they are unfolding at the moment of reading, and the
narrator takes a distal, uninvolved viewpoint; s/he is not in the event.
Yet sometimes, the narrator seeks to manipulate viewpoint to in-
volve the reader more, by imposing a viewing arrangement that com-
bines the affordances of the narrator's distal viewpoint which allows
story events to be viewed unfolding, with a closer-seeming viewing
arrangement that creates suspense. The simple present is a literary
form which does just this. It imposes just such a self-conscious yet
suspenseful perspective. The distal aspect of its viewpoint comes
from reporting perfective events as though the teller could view their
endpoint. As Langacker claims this imposes a seemingly "impossi-
ble" viewing arrangement for the conceptualizer unless she/he recon-
ceptualizes the event as virtual rather than actual (19). Langacker
Pretend play 43
goes on to point out that the simple present is difficult to use in de-
scribing perfective events, occurring at the moment of speaking be-
cause of what he termed the durational problem. As he puts it "be-
cause a perfective process is bounded, a full instantiation of it in-
cludes its boundaries, so if a perfective process is to coincide with the
time of speaking, its beginning point has to coincide with the initia-
tion of the event of speaking, and its endpoint with its termination"
(25). In other words, the speaker has to coordinate the event time
with the beginning of speech and wait for the termination of the re-
ported events to coincide with the time of speaking. The main use of
the present would be in event casts, which describe actions taking
place in the present time, such as sports commentaries, cooking
shows, direction giving, and even more rarely, advanced driving
tests. In event casts, such a viewpoint is possible because events are
happening quickly and each event is completed before the other be-
gins. Since such events rarely occur in everyday discourse, we tend to
assume a virtual, that is a hypothetical event is at issue when we
hear/see a simple present, as in the above beginning of the story.
It is more usual in descriptions of present actions to use progres-
sive constructions rather than the simple present. However with ha-
bituais, where there is no endpoint, the present is also used. When
progressives ("imperfectives") are used, the speaker is taking a tem-
porally coincident sample of the overall situation and observing that,
for the portion sampled, the situation is stable and unbounded. For
statements such as "I am driving to work", because a termination
viewpoint is not necessitated, and similarly for generic statements or
statements of habitual events (e.g. "I drive to work"), the endpoint
does not have to be encompassed in the time of speaking. Therefore,
to summarize Langacker's argument, there are two issues at stake in
the choice of tense-aspect. One is the focus on the event time as si-
multaneous with or prior to the speech event, that is the choice as to
whether one takes an involved or distal viewpoint; and the other is
the viewing arrangement or the choice of perspective, that is whether
the endpoint is revealed at the time of the telling.
If we accept that the simple present is unusual in ordinary, every-
day discourse, occurring only in special contexts, then children rarely
44 Jenny Cook-Gumperz and Amy Kyratzis
Oh me! Oh my!
Oh me! Oh my!
What a lot
of funny things go by.
2. Examples
The data: excerpts in this paper come from three studies: one a study
of two 3.6 year old girls who play regularly in each others' home
(Cook-Gumperz 1985, 1992). The second set come from a study of
character play with four- and seven-year-old dyads (Kyratzis 1992).
The third set of data is taken from spontaneous play situations
videotaped in a nursery school in a combined three- and four-year-old
classroom (Kyratzis and Guo 2001, Kyratzis forthcoming).
Two three and half year old girls, Lucie and Susie are playing a game
of mummies and babies, indoors in small play space, surrounded by a
collection of dolls and stuffed animals. The girls make up a loose
48 Jenny Cook-Gumperz and Amy Kyratzis
Example 1.1
Susie: We've got to go up here and cause it's a sunny day we've
got to go outside and sit down in our garfden]
Lucie: Oh... oh yes that., that can be there..that can be there.,
yes
Susie: And we sit down and have a glass of orange iuicev. (whis-
pered mummy to baby) there sit down.
(play drinking and pouring sounds)
Shh..shh..now some more., of., gumpy-shek.. it's mac-
cambe... it's maccamba
Susie: The babies don't like it
Lucie: No the babies don't like macacamba
[Two turns later]
Susie: And he sits.... And the babies sit on our laps with us
Pretend play 49
In this example the two little girls used sounds, (Shh..Shh..) and
made up words (maccacamba, gumpy shekt) to set up a rhyth-
mic/metric beat for the game. Tense contrasts, along with voice-tone
differences, mark the different perspectives represented and enacted
by the "game voices". The narrative plot is predominantly repre-
sented in the simple present tense conveying a distal perspective as if
the interactants are looking in on the stage where they are about to
perform their actions. The initial statement by Susie relies on the
model phrase "got to", but as the aspectual sense of these phrases has
something of a habitual character to them, they could be interpreted
as 'when it's sunny we always go and sit in the garden'. In Susie's
second and third utterance the simple present describes the actions as
they are about to happen, yet with the sense that these activities once
performed can be visualized as computable actions. Their potential
completeness contributes to their immediacy.
In the next example the girls get into a dispute about the conse-
quences of the action's narrative which is resolved in the last line by
Susie's addition to the plot that signals her acceptance of the story
line.
Example 1.2
Lucie: Anyway.... and you say.. "Sandra., have you got pins"., and
I'll sav 'Ves"
Susie: Have you got some pins?
Lucie: Yes
Susie: I want to hold her.. I want., it's not fair
Lucie: She's having your pin
Susie: She's mine
Lucie: ..Now you sav "Sandra., have you got pins., and I'll sav
"yes"
Susie: ..NO
Lucie: No you..And and you sav "Sandra..have you got pins to
stick in your baby and and you and I say "yes"
Susie: And um., then you come and give me two pins.
50 Jenny Cook-Gumperz and Amy Kyratzis
Example 1.3
Diane: How do you get to Belgium Susie?
Susie: First you go on a boat and then you get to Belgium and go
on a car the...then ..you get to Belgium
Lucie: and then you go on a train and...and sea and ships
Pretend play 51
Another set of examples where the two girls alternate their "mother-
to- mother" talk with mother-to-baby talk. Apart from the distinctive
voice/intonation patterning the former is more likely to be character-
ized by use of the progressive tense usually the present progressive as
the girls (as mummies) report their activities to each other, and the
mother to baby talk contains the greatest range of tense shifting and
syntactic complexities, using conditionals, imperatives with justifi-
cations, indirect imperatives e.g. "I don't want you to shout at San-
dra" reprimands in the form of questions e.g. "why did you spit water
at Sally Manga?"
Example 2.1
Lucie: Your baby is crying... Samolina...Santolina...allright...I'll
give you to your mummy...I'll give you to your mummy
now...Y our baby wants you
S: [inaudible] [out of range]
L: allright..here..she comes ..And then we are going to walk
around..should we go for a little walk?...s7za// we go for a
little walk...with Samolina..And try not to hurt her Samo-
lina...Samolina look..don't hit her ...oh let's go for a walk
you two...let's go for a walk with Sandra when she fin-
ishes There's Sandra...Sandra ..we are going for a little
walk
Susie: I know it.. where's my cup
52 Jenny Cook-Gumperz and Amy Kyratzis
gets to you"; "they say who's that"; "I say mama"; "you take a good
look and say"; "I say that's my mom"; "and he says that's my mom").
In the narrative voice, when the girls say "I", they are referring not to
themselves {in realis) but to the character of the Playmobile figure
they have designated as acting out their role. In this voice, the girls
seem to be reading off a script representation that is already con-
structed. As Langacker (this volume) states, "Tense marking on a
structural statement specifies the time at which the (already con-
structed) event representation can be consulted as a way of appre-
hending the world's structure" (Langacker this volume). In contrast,
when the girls are talking as themselves, in stage-manager voice us-
ing the first person, they include modal marking on the verb (e.g.
"I'm supposed to act myself'; "pretend I could sit the seals on"), just
as they also use commands and questions in this voice ("do it like
this", "make it like this so the seals can sit"; "wait, I'll be right
back").
The excerpts below exemplify how this dyad shifts tense and as-
pect to mark different phases of the play, using statives in the set-up,
beginning phase, then shifting to progressive and simple present to
mark background and temporally sequenced action, respectively.
[up to now, they've been doing an animal tricks show. Now they shift
to ice skating show scenario]
Example 3.1
Fran: and pretend somebody was ice-skating on the rink, there was
a mom.
wait, pretend you got off the rink to watch for a while.
[F grabs A's figure and places it in front A]
Abby: okay, I got off it. [picks up the figure]
Fran: now, now somebody's on the iceskate rink, 'cuz when
she...[grabs A's figure again]
I'll set her up to be standing.
Abby: okay, but *he is the audience, [grabs a male figure and
places it in front of F] wait. I'll be right back. [A goes to F's
54 Jenny Cook-Gumperz and Amy Kyratzis
In these lines, the girls are setting up the scenario, and they use a lot
of copulas "somebody is on the ice-skating rink" and statives "there
was a mom", "they were boyfriend and girlfriend", including one in
present tense "but *he is the audience". Statives are appropriate for
the setting-up phase of the pretend. However, after this is accom-
plished, they shift to progressive and simple present.
Example 3.2
Fran: and you're sitting where the man - where the lady zooms by,
but she slows down when she gets to you. okay? [F runs her
figure along the "ice" until it reaches F's side of the circus
ring] they say 'who's that'?
Abby: and I say- and I say- and I say [gets up and walks towards F]
Fran: 'mama!'
Abby: 'oh, he's my boyfriend', we're talking.
Fran: no, no, you ask 'who's that?', and then you take a good look
and say 'that's your mom!'
Abby: um,'who's that'?
Fran: and I say, 'that's my mom'.
Abby: and he, and he says, 'that's my mom'.
Fran: yeah, they both say that because they're twins.
In the first line, "and you're sitting where the man - where the lady
zooms by, but she slows down when she gets to you", F presents an
event sequence (lady zooms by, slows down), and uses the simple
present to denote the temporally sequenced action. Simultaneously,
she presents the background event (the audience sitting), using the
progressive, since this is background material and its endpoint is not
a focus of the story. Lower, three turns later, A uses progressive as-
pect to again denote background material - the explanation of her
character's speech ("we're talking"). Throughout the rest of the ex-
cerpt, through a long chain of story action, the two girls use simple
present to denote the temporally sequenced acts ("you ask", "and
Pretend play 55
then you take a good look" "and say", "and I say" and "he says". In
the last line, they use third person in what Langacker called the sta-
tive sense, because the endpoint of the act "say" is not a focus and the
information is merely explanatory, hence background. In example
3.2, as in the Dr. Seuss excerpt above, the simple present is used
when there is an "observer" to the event (in this case, the Mom).
Other times in their narrative, when there isn't such involved view-
point, the girls use the preterite to tell and plan their story.
Example 3.3
Fran: hello sweetie! Pretend you said, "hey that's my mama".
Abby: 'hey, that's mama/and then she...no, pretend that, I'm then,
the um, the trumpeters,
well, like,
Fran: ==the seals.
Abby: the seals come on
Fran: and I do the, um, the
Abby: and they go 'shooo--(x)
Fran: put that one on this!
Abby: no, I'm gonna do it my way.
2.4. Present tense for habitual state, progressive for ongoing action
in role-play with four-year-olds
Example 4
Aly I'm making soup// (kalatoya)// (kalatoya)// (kulakepa)//
Viv (malakapu)// pretend I'm making food// =(xx) pretend= I'm
making Chinese stuff//
Aly =I'm making food for my brother=// I'm
making food for my brother//
Viv pretend I'm making (rice)
Aly == I'm making Chinese ri- soup for my brother//
Viv yeah// and I'm making soup//
Example 5
Aly pretend I'm Chine::se// I'm a Chinese sister/ and I look
pretty//
Viv pretend this (xx)// pretend (xx) are called (rices)//
Aly and those are called chaniza// because they're- they're spe-
cial kind of tortilla// and we always need itII
3. Conclusion
Note
References
Hopper, Paul
1979 Aspect and foregrounding in discourse. In: Talmy Givón (ed.),
Discourse and Syntax. Syntax and Semantics, vol. 12, 213-241.
New York: Academic Press.
Kurland, Brenda F. and Catherine E. Snow
1997 Longitudinal measurement of growth in definitional skill. Journal
of Child Language 24: 603-625.
Kyratzis, Amy
1992 Gender differences in the use of persuasive justifications in chil-
dren's pretend play. In: Kira Hall, Mary Buchholtz and Birch
Moonwomon (eds.), Locating power : Proceedings of the Second
Berkeley Women and Language Conference 2, 326-337. Ber-
keley: Berkeley Women and Language Group.
Forthcoming Language socialization across time and context: Emotion talk in
preschool same-sex friendship groups. To appear in: Nancy Bud-
wig (ed.), Language Socialization during the Preschool Years
[Special issue]. Early Education and Development.
Kyratzis, Amy and Jiansheng Guo
2001 Preschool girls' and boys' verbal conflict strategies in the U.S.
and China: Cross-Cultural and contextual considerations. In: Amy
Kyratzis (ed.), Gender Construction in Children's Interactions: A
Cultural Perspective [Special Issue]. Research on Language and
Social Interaction.
Langacker, Ronald W.
This volume Cognitive linguistics, language pedagogy, and the English present
tense.
Milne, Alan Alexander
1926 Winnie-the-Pooh. New York: E. P. Dutton
Slobin, Dan I.
1982 Universal and particular in the study of child language. In: E.
Wanner and L.R. Gleitman (eds.), Language Acquisition: The
State of the Art, 128-170. Cambridge: Cambridge University
Press.
1986 Crosslinguistic evidence for the language-making capacity. In:
Dan I. Slobin (ed.), The Cross-Linguistic Study of Language Ac-
quisition, Vol. 2: Theoretical Issues, 1157-1256. Hillsdale, NJ:
Lawrence Erlbaum Associates.
White, Hayden
1980 The value of narrativity in the representation of reality. In: W. J.
T. Mitchell (ed.), On narrative, 1-23. Chicago: The University of
Chicago Press.
62 Jenny Cook-Gumperz and Amy Kyratzis
1. Introduction
lems with the analyses of the usage of English tense morphology. The
first involves a contradiction between the representation of the basic
sense of the past tense morpheme and several of its non-temporal
uses. Langacker (1991), whose work forms the basis of many subse-
quent analyses, represents past time as being equated with "non-
immediate reality" (ibid.: 242), and present time with "immediate
reality" (ibid.: 242). The situations or states of affairs being refer-
enced by the English tense morphemes are represented as known
reality, i.e. they "are accepted by a conceptualizer as being real"
(Langacker 1991: 242). Reality status is argued to be signaled lin-
guistically by the presence or absence of modal marking. Past time,
which largely equates with non-immediate reality, is signaled by the
past tense morpheme; present time, which largely equates with im-
mediate reality, is signaled by present tense morphology. In contrast,
future and modal forms are represented as signaling irrealis. How-
ever, English speakers regularly use the past tense morpheme to sig-
nal relatively less commitment to the reality of an event or state of
affairs; past tense is also used in certain politeness phenomena, which
are clearly not interpreted as states of affairs accepted as real. These
uses are at odds with the "known reality" representation.
The second problem concerns the explanatory power of certain
claims about metaphorical distance and tense morphology. Recog-
nizing that English tense morphemes are not solely interpreted in
terms of temporal relations, Langacker (1991) argues that there is an
epistemic opposition between "immediate and non-immediate real-
ity" (ibid.: 245-246) which is marked by the absence or presence of
past tense morphology;2 the present and past tenses contrast with
"immediate and non-immediate irreality" (ibid.: 245-246) which are
marked by various modal forms. On this basis, he labels the past
tense morpheme as a "distal marker". Significantly, his discussion in
relation to the past tense morpheme centers on temporal distance.
Expanding on Langacker's analysis, Dirven and Radden (2000:
Chapter 9) note that a sentence such as: I wanted to ask you a favor
"illustrates a metaphorical shift of the past tense...The use of past
tense in [I wanted to ask you a favor] achieves an effect of politeness;
it distances the situation in time and, as a result makes the request
68 Andrea Tyler and Vyvyan Evans
same image content links them at the conceptual level. This concep-
tual linking, then, licenses the use of tense morphology to signal
these various uses. Finally, in Section 6, we sketch some of the im-
plications of the analysis for language pedagogy.
2. The phenomena
Time-reference
(1) a. I work in advertizing (present time-reference)
b. Yesterday I went to the cinema (past time-reference)
Intimacy
(2) a. A: Jane just bought a Volvo.
B: Maureen has one.
A: John, you've got to quit talking about Maureen as if
you're still going together. You broke up three months
ago. (Riddle 1986)
Salience
(3) a. In November 1859, Charles Darwin's The Origin of Species,
one of the greatest and most controversial works in the lit-
erature of science, was published in London. The central
idea in this book is the principle of natural selection. In the
sixth edition.. .Darwin wrote: "This principle of preservation
of the survival of the fittest, I have called Natural Selection."
(Eigen and Winkler 1983: 53)
Actuality
(4) a. I wish I knew what he'll say next. (Westney 1994)
b. I wish the students liked phonetics. (Fleischman 1989)
Past tense in the if clause (the protasis) signals a negative stance vis-
à-vis the conditional situation and reality, i.e. the situation described
is non-actual. The sentence roughly paraphrases as "He is not study-
ing hard at the moment. If he changes the present circumstances, and
he studies harder, it would be possible for him to get better grades."
Attenuation
(5) a. I was thinking about asking you to dinner. (Fleischman
1989)
b. I was hoping we could get together next week. (Fleischman
1989)
During the course of this paper, we will argue that time-reference, the
primary meaning associated with tense forms, comes to implicate
non-temporal meanings, which in turn through entrenchment (a pro-
cess we term pragmatic strengthening) comes to be conventionally
associated with the tense morphemes. However, in order to able to
offer a motivated account for the association of non-temporal mean-
ings with the two tense morphemes, we need to establish why it
should be that time-reference should implicate intimacy, salience,
actuality and attenuation in the first place. We will argue that all five
of these response concepts are elaborated in terms of locational con-
tent pertaining to the relative physical proximity of the experiencer
because we conceptualize each of the concepts of time-reference,
intimacy, salience, actuality and attenuation in terms of proximal-
distal relations with respect to the experiencer. Hence, these distinct
concepts, while not literally being spatial concepts are all elaborated
in terms of spatial deixis, which, as we will demonstrate, is motivated
in each instance by a distinct, tight correlation in experience. It is by
virtue of being elaborated in terms of similar image content, we will
suggest, that the tense morphemes, which denote time-reference, can
come to implicate and ultimately denote non-temporal meanings.
In the remainder of this section, we examine the evidence for ex-
periential correlations which give rise to each of these distinct con-
cepts being associated with the proximal-distal dimension. We will
The relation between experience, conceptual structure and meaning 81
4.1. Time-reference
between the primary sense associated with the tense morphemes and
their non-temporal uses.
4.2. Intimacy
(8) Peggy and I have been close for many years, but lately she has
been acting a little distant.
4.3. Salience
4.4. Actuality
4.5. Attenuation
In this sentence, the degree of control over and hence ability to affect
the budget is articulated by the phrase tight grip, which literally de-
notes very close physical contact. Examples such as these are li-
censed by virtue of the experiential correlation between the ability to
affect something and physical proximity. Experience also tells us that
physical distance results in a lessening of the ability to affect a par-
ticular entity. This is reflected in sayings such as:
In this example the use of the present tense form has by B, is inter-
preted by A as a claim of (unwarranted) intimacy. John does not have
direct knowledge as to whether Maureen still owns a Volvo. He is
speaking as if they are in an on-going relationship which would give
him that knowledge. As such, in examples such as this the use of
tense provides an intimacy reading. The point is that the tight corre-
lation between intimacy and physical proximity is a corollary of the
correlation between time-reference and physical proximity. Being
temporally "located" provides immediate and verifiable experience.
As such, knowing that Maureen has a Volvo in the present represents
a stronger claim to intimacy than knowing that at some point in the
past she had one.
Due to an intimacy reading having become conventionally associ-
ated with tense markers, we suggest that A is able to interpret the
tense usage not in terms of time-reference, but rather in terms of in-
timacy, due to tense being conventionally employed in certain situa-
tions to signal relative intimacy.
A second example illustrates the use of past tense to signal lack of
emotional intimacy:
The main point of these sentences is not the precise date of publica-
tion of this book, but rather the central topic of the book. The infor-
mation presented in the first sentence establishes the frame for the
focal information which occurs in the second sentence. The informa-
tion in (3 c) provides supporting evidence for the key point, and as
such is less prominent in terms of information status. Discourse ana-
lysts have often referred to the relative status of information such as
that in sentence (3 a) and (3 c) as background and information such as
that in sentence (3b) as foreground. In this example, tense is em-
ployed to signal the relative status of the information, i.e. past tense
signals background status and present tense signals foreground status.
As before, we suggest that tense comes to signal salience for the
following reason. In certain contexts, when, for instance evaluating
the relative importance of information, some pieces of information
are implicitly more or less important than others. As time-reference
correlates with physical proximity, so too degree of salience corre-
lates with that which, as noted earlier, tends to be in foveal vision and
thus physically proximal. Accordingly, as time-reference shares
similar image content with salience, tense can in some contexts im-
plicate relative salience. We suggest that through continued use of
tense in contexts in which salience is implicated, namely pragmatic
strengthening, tense morphology has developed a conventionalized
meaning component, in which past tense denotes relatively less sali-
ence while present tense denotes greater salience.
This use of past and present tense in order to signal relative sali-
ence relates to Langacker's (this volume) arguments concerning non-
canonical uses of the present tense. He discusses vivid narrative,
play-by-play sports casting, historical present, and other discourse
uses of present tense which clearly do not conform to the typical rep-
resentation of present tense usage. It is not entirely clear how his ex-
The relation between experience, conceptual structure and meaning 93
7. Conclusion
Notes
ment, "You can use 'want' instead of 'would like' to give an instruction or
make a request... 'Wanted' is also sometimes used. It is more polite than
'want.' (ibid.: 240) No further explanation is provided. Similarly, in the section
on expressing importance, there is no discussion of tense as a signal of the rela-
tive importance of information within the discourse (ibid.: 236-237; 257); the
only remotely possible mention is in a usage note associated with the present
tense under the rubric of 'Vivid narrative' (ibid.: 257). This deals only with per-
sonal narratives coded exclusively with the present tense; present tense is repre-
sented as a device to increase audience involvement. The only example involv-
ing tense given in relation to actuality is the use of the present tense in the fol-
lowing, "Suppose we don't say a word and somebody else finds out about it."
Although native speakers also use the past tense in such structures, this use is
not mentioned This grammar also clearly exemplifies the tendency to scatter
non-temporal uses of tense, essentially representing them as arbitrary, and of-
fering no attempt at a unified, systematic account that would tie them to the
central temporal sense.
2. Specifically, Langacker argues that immediate reality is marked by the absence
of a modal and the absence of the distal morpheme; non-immediate reality is
marked by the absence of a modal and the presence of the distal morpheme. The
immediate/non-immediate contrast is argued to be "a proximal/distal contrast in
the epistemic sphere." (ibid.: 245) "Immediate reality coincides temporally with
the time of speaking, so to the extent that the notion of time is specifically in-
voked, present time is conceived as one facet of immediate reality... In precisely
analogous fashion, the predication of non-immediate reality is equivalent to one
of past time...These notions are basically epistemic, i.e. they do not refer to
time, yet they have an obvious interpretation with reference to the time-line
model: since reality subsumes the past and present (but not the future), and im-
mediate reality constitutes the present, the temporal projection of non-
immediate reality can only be the past. Presumably, then, the distal morpheme
has a prototypical value that invokes the time-line model and is reasonably con-
sidered a past-tense predication. That, however, is only one manifestation of its
basic epistemic import." (ibid.: 246).
3.Sweetser (1990) notes that experiences which have been labeled concrete tend
to be those which allow continuity across individuals. For instance, "vision
is...identical for different people - that is to say, two people who stand in the
same place are generally understood to see the same thing ... Identity across
people is a highly objective characteristic..." (ibid.: 39) In contrast, "abstract"
concepts tend to be identified with internal states. Since human beings cannot
communicate through mental telepathy, one human being cannot directly ob-
serve or know another person's mental or emotional state; we can only guess.
Internal states simply do not allow for verifiable identity across people in the
same way that entities and events in the external world do. Moreover, internal
The relation between experience, conceptual structure and meaning 101
References
Binnick, Robert
1991 Time and the Verb. New York: Oxford University Press.
Brown, Penelope and Stephen Levinson
1987 Some Universals in Language Use. New York: Cambridge Uni-
versity Press.
Celce-Murcia, Marianne and Diane Larsen-Freeman
1998 The Grammar Book. Rawley, MA: Newbury House.
Collins Cobuild English Grammar
1994 London and Glasgow: Collins Publishers.
Comrie, Bernard
1985 Tense. Cambridge: Cambridge University Press.
Crick, Francis
1994 The Astonishing Hypothesis. New York: Simon Schuster.
Crick, Francis and Christof Koch
1990 Towards a neurobiological theory of consciousness. Seminars in
the Neurosciences 2:263-275.
1998 Consciousness and neuroscience. Cerebral Cortex 8: 97-107.
Cutrer, Michele
1994 Time and Tense in Narrative and in Everyday Language. USCD
Dept. of Cognitive Science: Technical Report 9501.
102 Andrea Tyler and Vyvyan Evans
Jespersen, Otto
1924 The Philosophy of Grammar. London: Allen and Unwin.
Lakoff, George
1987 Women, Fire and Dangerous Things. Chicago: Chicago Univer-
sity Press.
1993 The contemporary theory of metaphor. In: Andrew Ortony (ed.),
Metaphor and Thought, 202-251. Cambridge: Cambridge Uni-
versity Press.
Lakoff, George and Mark Johnson
1980 Metaphors We Live by. Chicago: Chicago University Press.
1999 Philosophy in the Flesh. New York: Basic Books.
Lakoff, Robin
1973 The logic of politeness; or minding your p's and q's. CLS 9, 292-
305.
Langacker, Ronald
1987 Foundations of Cognitive Grammar, vol. 1. Stanford: Stanford
University Press.
1991 Foundations of Cognitive Grammar, vol. 2. Stanford: Stanford
University Press
This volume Cognitive linguistics, language pedagogy and the English present
tense.
Lock, Graham
1996 Functional English Grammar: An Introduction For Second Lan-
guage Learners. Cambridge: Cambridge University Press.
Mandler, Jean
1992 How to build a baby: II. Conceptual primitives. Psychological
Review 99(4): 587-604.
Moore, Kevin Ezra
2000 Spatial experience and temporal metaphors in Wolof. Ph.D thesis.
Linguistics dept., University of California at Berkeley.
Ortony, Andrew
1988 Are emotion metaphors conceptual or lexical? Cognition and
Emotion 2: 95-103.
Pöpel, Ernst
1994 Temporal mechanisms in perception. In: O. Sporns and G. To-
noni (eds.), International Review of Neurobiology 37, 185-202.
San Diego, CA.: Academic Press.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik
1973 A Grammar of Contemporary English. New York: Harcourt
Brace Jovanovich Inc.
104 Andrea Tyler and Vyvyan Evans
Rauh, Gisa
1983 Tenses as deictic categories: An analysis of English and German
tenses. In: G. Rauh (ed.), Essays on Deixis, 229-275. Tübingen:
Gunter Narr.
Reichenbach, Hans
1947 Elements of Symbolic Logic. New York: The Free Press.
Riddle, Elizabeth
1986 The meaning and discourse function of the Past Tense in English.
TESOL Quarterly 20: 267-286.
Stryker, Michael
1991 Seeing the whole picture. Current Biology 1 (4): 252-253.
Swan, Michael and Bernard Smith
1987 Learner English. Cambridge: Cambridge University Press.
Sweetser, Eve
1990 From Etymology to Pragmatics: Metaphorical and Cultural As-
pects of Semantic Structure. Cambridge: Cambridge University
Press.
Talmy, Leonard
1988 Force dynamics in language and cognition. Cognitive Science 2:
49-100.
Tannen, Deborah
1993 What's in a frame? Surface evidence for underlying expectation.
In: Deborah Tannen (ed.), Framing in Discourse, 14-56. Oxford:
Oxford University Press.
Ter Meulen, Alice
1995 Representing Time in Natural Language: The Dynamic Interpre-
tation of Tense and Aspect. Cambridge, MA: MIT Press.
Traugott, Elizabeth Closs
1975 Spatial expressions of tense and temporal sequencing: A contri-
bution to the study of semantic fields. Semiotica 15(3): 207-230.
1978 On the expression of spatio-temporal relations in language. In: J.
Greenberg (ed.), Universals of Human Language. Vol. 3: Word
Structure, 369-400. Stanford: Stanford University Press.
1989 On the rise of epistemic meanings in English: An example of
subjectification in semantic change. Language 65: 31-55.
Tyler, Andrea and Vyvyan Evans
Forthc. a Reconsidering prepositional polysemy networks. Language.
Forthc. b Spatial Scenes. New York and Cambridge: Cambridge University
Press.
The relation between experience, conceptual structure and meaning 105
Westney, Paul
1994 Rules and pedagogical grammar. In: T. Odlin (ed.), Perspectives
on Pedagogical Grammar, 72-96. Cambridge: Cambridge Uni-
versity Press.
Section 2
Facets of prototypes in
grammatical constructions
Grammatical constructions and their discourse
origins: prototype or family resemblance?
Paul J. Hopper
1. Introduction1
Some of the uncertainty, the tension between what people say and
write and what is "canonical", comes to the surface when the wh-
clause is in the perfect and the resumptive verb is the -en participle
rather than the base form or infinitive, as in ?What he has done is
spoilt the whole thing. Revealingly, the authors suggest that this us-
age is a reduced form of two discrete utterances: "of doubtful accept-
ability...may be an ellipted form of: What he's done is this: he's
spoilt the whole thing".
In the by now considerable literature on pseudoclefts, two themes
stand out. The first is that the function of the pseudocleft sentence is
to background the material in the wh- clause so as to throw certain
constituents of the predicate into contrastive relief. Thus in What my
car needs is a new battery, only the direct object a new battery is
new; the subject my car and the verb needs are both old or "given".
The constituent-focus theory runs through both sentence-level and
discourse studies of pseudoclefts. This distribution is claimed, for
example by Prince (1978), to correspond to contexts in which (in this
hypothetical instance) my car and needs are in some way recoverable
("given", "presupposed" in her terminology) from the previous dis-
course, either directly or inferentially. Similarly, in What John did to
his new suit was ruin it, the agent and the grammatical object are
already known and only the verb is asserted. Such treatments moti-
vate pseudoclefts by reference to the ambient and preceding context,
rather than by reference to upcoming segments. They derive from a
linguistic theory that holds that linguistic forms code current mean-
ings. I will here argue that in spoken discourse pseudoclefts are an-
Grammatical constructions and their discourse origins 111
Fragments such as what I'm fascinated is are said (Collins 1994: 44)
to be "system-deviant". System-deviance, in Collins' view, arises
from processing factors: "There are, broadly speaking, two categories
of incomplete pseudo-clefts and clefts: those whose completion is
prevented by a variety of contextual or processing factors, and ellipti-
cal constructions in which omitted material is textually recoverable."
In either case, the assumption is that the forms are exceptions to the
normal realization of the pseudocleft. Speakers lose their drift or do
not bother to state material that is obvious from the context. The lan-
guage system (the speaker's competence) provides for canonical
pseudoclefts, but speakers mutilate them in attempting to produce
(i.e. perform) them.
In this paper I will suggest that when natural spoken discourse is
considered, it appears that pseudoclefts do not function primarily, or
perhaps at all, to highlight any single identifiable sentence constituent
such as a verb or a noun phrase. If such highlighting occurs, it is a
by-product - an epiphenomenon - of a more rhetorical use of the
pseudocleft: pseudoclefts function in natural discourse to delay an
assertion for any of a number of pragmatic reasons. Furthermore, the
"system-deviant" fragments are just as capable of carrying out this
112 Paul J. Hopper
2. Pseudocleft fragments
(2) Anna: But have you discussed the fact that you don't believe
[...] that this person can actually help you with this
person
Mike: Yeah I have Anna. And what they're s what my psy-
chologist has told me is that they will make arrange-
ments for me to see a psychiatrist if they f th if I still
feel this
or a complete predicate:
Grammatical constructions and their discourse origins 113
(4) the kind of newspapers that they want erm are are are really in a
sense not er any longer er those down-market tabloids. And so
it's it's b it's a struggle for in what I suppose er [= a?] popula-
tion of about ten million I'm not quite certain myself exactly
what it is but it's a huge population the two er main daily news
er popular daily newspapers can barely sell eight hundred.
And at the other extreme, the continuation may be pursued over a gap
of several subsequent clauses, as in this example:
(5) I mean maybe this is because of the school. I only have them on
a Monday and then I have them next week Monday. Now what
happens is we tried in the past different groups. You know dif-
ferent forms. So you ask somebody from each of these about
three or four different forms Collect the homework for me on
Tuesday...
(6) ...wood stain almost black. It's really beautiful and it's got all
the cornices and the ceiling rose and high skirtings all original.
Anyway then you go in on this it's like a parquet floor and then
114 Paul J. Hopper
what they've done is the cellar it's got a cellar and you go down
the steps to the cellar but there's like a proper two proper
rooms so on your left you've got a sort of cellar with a quarry
tile.
eral "delaying" function, we may discern from the corpus at least the
following five:
(8) ...it need to kill something like a thousand times its own size.
There's η there don't seem to be a r- real need. And in defence
I mean what snakes or what animals try <pause> like what
most animals try to do is if they tha have got a poisonous prop-
erty is another animal attacks them they give them er a dose of
venom which will not kill them it will just deter them next
time.
(10) Robin: .. .you feel <pause> I mean <pause> the campaign has
been going on in Scotland for longer than in in this
country do you think it's erm a a victory for what I
suppose you ΊI see it as a victory for the people over
government I suppose would you?
Walter: Yeah I do. I mean actually I mean it didn't work in
Scotland from the start and that was like
(11) to to clean up Mexico City air. And obviously this then became
a pretty major objective for himself in terms of his own ambi-
Grammatical constructions and their discourse origins 117
Jerry: when I go into the office I'll nut the first person I see
(14) I mean I think Peggy and Sheila will be interested to hear what
the software actually does here because this is relevant obvi-
ously to them. Erm (...) I mean basically what it does is it di-
vides the the word are defined into various groups.
(15) Frank: ... she can ring Charles and speak to him.
Lisa: Yes. Erm <pause> try yes try and get her to ring you
know <pause> t over the next day or so. Well I sup-
pose what'll happen is she'll ring and you'll answer
probably and it'll be good if I answer it won't I ooh-er
(16) Yeah but you see y y <sighs> <pause> you've got as much of a
problem in expecting one of them to be [handy] when some-
thing happens as you have with a police officer. What we need
<pause> actually is more money spent on the police isn't it?
(17) John: ...what kind you know? I've got a few brochures
today er d is there any kind of do you could you tell
me what it must need to have you know like er in
the way of disc drives and things?
Imogene: Well erm what you need is a erm there's a small
disc drive you don't need a five and a half inch disc
drive.
Grammatical constructions and their discourse origins 119
(18) If we're saving then what we want to know is how much more
could we spe buy how much more goods and services could we
acquire at the end of the year er by saving now
(19) 'Cos like I wanted really what I wanted was I'd never seen
anyone all day so what I suppose I wanted was someone to talk
to
(21) Actually, what you might do is leave this part and say, these
were the parameters originally to the committee. (CSPAE)
(22) and then what they 've done is the cellar it's got a cellar and you
go down the steps to the cellar but there's like a proper two
proper rooms so on your left you've got a sort of cellar with a
quarry tile
120 PaulJ. Hopper
(23) a rather modern play and it seems to avoid the kind of language
<pause> traditionally expected from drama pointed well-
formed eloquent witty erm and it's what I suppose is one of the
reasons why we see this play as a sort of er modern or experi-
Grammatical constructions and their discourse origins 121
The speaker "Alan"4 has a faltering delivery and is clearly fishing for
words. The two hesitation markers in a short stretch, and even the
sniff, and the resort to ready-made expressions all betray this. What's
happened is is not the only cliché in Alan's utterances (facets of the
curriculum, coming on board [twice], got to grips with, as well as we
can, at this moment in time). To the reader of the transcript he is
practically unintelligible, but of course the listener, who had access to
the entire conversation to that point, may have had no such difficulty.
The frequent association of pseudocleft pieces with disfluencies
such as hesitation markers and clichés points to a role in processing
spoken discourse on line. Pseudocleft pieces seem to occur with
some regularity after a pause and also at points in the conversation
where one speaker has not taken an expected turn and the other in-
terlocutor has to step into the breach:
ing function that has been noted in written or edited spoken texts
derives from its less specific discourse focusing function. But more
important, as has been argued here, seems to be the delaying func-
tion, a rhetorical effect of impressing the listener with the "social"
significance of something about to be said, and making the listener
aware that what follows is part of a considered argument worthy of
attention and not a casual comment. From this delaying function, in
one way or another, derive the various other functions that make the
pseudocleft such a widely used and prominent construction in Eng-
lish.
I would surmise that this general picture is true not only of pseu-
doclefts but of all of the more complex grammatical constructions.
Grammatical constructions are normativized rationalizations of
families of smaller and more fragmented quasi-lexical parts. They
result from the grammaticalization of these fragments.
A further observation is that the discourse motivations for the use
of pseudocleft fragments are essentially temporal rather than struc-
tural. These motivations reside not in the static distribution of pre-
suppositions across a completed construction, but in the timing and
delivery of utterances: anticipating upcoming material, delaying an
utterance for assessed rhetorical and cognitive effectiveness, finding
strategies for warding off interruption. The presuppositional structure
that is such a striking feature of written pseudoclefts is, one would
surmise, a by-product of the transfer of the pseudocleft to a
monologic atemporal medium and its expansion there as an extended
grammatical construction.
3. Conclusions
The native speaker has acquired a grammar on the basis of very restricted
and degenerate evidence. (Chomsky 1972:23)
Many children acquire first and second languages quite successfully even
though no special care is taken to teach it to them and no special attention is
given to their progress. It also seems apparent that much of the actual speech
observed consists of fragments and deviant expressions of a variety of sorts.
(Chomsky 1965:201).
Corpus studies suggest instead that these "degenerate" data are the
true substance of natural spoken language, and that what our descrip-
tive and prescriptive grammars give us are normativized assemblies
of these fragments that tend to impress themselves on us as mental
prototypes because of their greater social prestige - their associations
with schooling, with literacy, and with complex discourse character-
ized by long periods and uninterrupted turns. This observation has
significant consequences for both linguistic theory and applied lin-
guistics. One moral to be drawn is that foreign language teachers and
applied linguists in general should attend to the role of shorter, "in-
complete" utterances, even ones that violate rules of canonical
Grammatical constructions and their discourse origins 127
Notes
1.1 am grateful to Rosalind Moon and Jeremy Clear for their help in introducing me
to the Collins-Birmingham University International Linguistic Database (Co-
build, <http://www.cobuild.collins.co.uk/>) during a visiting fellowship at the
Center for Advanced Research in English at Birmingham University in the spring
of 1999; to Malcolm Coulthard for securing this appointment for me and for his
hospitality; to the Faculty Development Fund at Carnegie Mellon University for
financial support; and to David Kaufer, Head of the English Department at
Carnegie Mellon, for arranging academic leave for me during the period of the
fellowship.
2. Citations are from two sources. Those taken from the UK Spoken subcorpus of
Cobuild are unattributed. A small number of additional examples have been
taken from the Corpus of Spoken American Professional English, by Michael
Barlow, available from Athelstan <http://www.athel.com/cspa.html>. These are
tagged with the initials CSPAE. As is customary in corpus studies, citations are
lifted from the original corpus without modifications of spelling or punctuation.
3.For a recent account of the importance of subjectivity in language, see Scheib-
man 2000.
4. Speakers in Cobuild are identified only as "male voice" or "female voice". Ficti-
tious names have been supplied in the present article.
5. For a summary and important analysis of the mutual relevance of turn-taking to
cognition, see the Introduction to Ochs, Schegloff, and Thompson (eds.), 1996.
6.Constructional schémas are roughly equivalent to prototypes. See Taylor 1998:
177-179 and footnote 1 (p. 199) for discussion.
7. On the Cognitive Linguistics view of schémas as abstractions that are immanent
to linguistic expressions, cf. Langacker (1998): "It is reasonably supposed that
schémas are immanent in their instantiating expressions, and emerge as cognitive
entities by reinforcement of the structural properties they share at a certain level
of abstraction" (page 13; italics as in the original). In Langacker's view, grammar
resides in "schematized representations of sound-meaning pairings, abstracted
from (and immanent in) the specific symbolic configurations observable in com-
plex expressions" (page 2).
128 Paul J. Hopper
References
Bybee, Joan L.
1998 The emergent lexicon. Papers of the 34th Annual Meeting. Chi-
cago Linguistic Society.
Forthcoming Phonology and Language Use. Cambridge: Cambridge Univer-
sity Press.
Bybee, Joan L. and Paul Hopper
Forthcoming Introduction. In: Joan L. Bybee and Paul Hopper (eds.), Fre-
quency and the Emergence of Linguistic Structure. Amsterdam:
John Benjamins.
Chomsky, Noam
1965 Aspects of the Theory of Syntax. Cambridge: The MIT Press.
1972 Language and Mind. New York: Harcourt Brace Jovanovich.
Collins, Peter C.
1994 Cleft and Pseudocleft Constructions in English. London: Rout-
ledge.
Davis, Hayley
1992 Drawing the morphological line. In: George Wolf (ed.), New De-
partures in Linguistics, 90-115. New York: Garland.
Ford, Cecilia E., and Sandra A. Thompson
1996 Interactional units in conversation: syntactic, intonational, and
pragmatic resources for the management of turns. In: E. Ochs,
E. A. Schegloff, and S. A. Thompson (eds.), 134-85.
Hopper, Paul J.
1996 When grammar and discourse clash. Essays on Language Func-
tion and Language Type dedicated to T. Givon, 231-246. Edited
by Joan Bybee, J. Haiman, and S. A. Thompson. Amsterdam:
John Benjamins.
Langacker, Ronald
1998 Conceptualization, symbolization, and grammar. In: M. Toma-
sello (ed.), 1-40.
Lord, Albert Β.
1960 The Singer of Tales. (Harvard Studies in Comparative Literature,
24) Cambridge: Harvard University Press.
Mühlhäusler, Peter
1983 Stinkiepoos, cuddles, and related matters. Australian Journal of
Linguistics 3: 75-91.
Ochs, Elinor, Emanuel A. Schegloff, and Sandra A. Thompson (eds.)
1996 Interaction and Grammar. (Studies in Interactional Sociolinguis-
tics, 13) Cambridge: Cambridge University Press.
Grammatical constructions and their discourse origins 129
Ong, Walter
1982 Orality and Literacy: The Technologizing of the Word. London:
Methuen & Co. Ltd.
Prince, Ellen
1978 A comparison of WH- and IT clefts in discourse. Language 54:
883-906.
Quirk, Randolph, S. Greenbaum, G. Leech, and J. Svartvik
1985 A Comprehensive Grammar of the English Language. London:
Longman.
Ross, Haj
To appear The frozenness of pseudoclefts: towards an inequality-based
syntax. Papers of the Annual Meeting of the Chicago Linguistics
Society.
Schegloff, Emanuel A.
1996 Turn organization: one intersection of grammar and interaction.
In: E. Ochs, E. A. Schegloff, and S. A. Thompson (eds.), 52-133.
Scheibman, Joanne
2000 Structural Patterns of Subjectivity in American English Conver-
sation. Dissertation (Ph.D.): Department of Linguistics, Univer-
sity of New Mexico.
Shirai, Yasuhiro
1990 Putting PUT to use: Prototype and metaphorical extension. Issues
in Applied Linguistics 1: 78-97.
Taylor, John R.
1998 Syntactic constructions as prototype categories. In: M. Tomasello
(ed.), 177-202.
Tomasello, Michael (ed.)
1998 The New Psychology of Language. Mahwah, NJ: Lawrence Erl-
baum.
Transitivity parameter and prominence
typology: a cross-linguistic study*
The fact that these unagent-like noun phrases can appear as syntactic
subjects in English does not pose serious problems in cognitive
grammar since they are considered mere deviations from the central
semantic specification of prototypical transitivity (Taylor 1995:
214).1 Compared to English sentences like Mary killed a spider or
Tom hit a ball, the semantic relation between subject and each event
involved in (1) is less intense. The subjects in (1) do not execute the
processes denoted by the verbs. On the basis of Dutch, French and
English data, Voorst (1996) demonstrates that there is cross-linguistic
variation in the typical transitive constructions even with verbs such
as break and buy as far as the level of intensity is concerned. Thus,
Transitivity parameter and prominence typology 133
the active voice (see also Shannon 1987). However, Shibatani does
not explain why languages exist in which prototypical transitive con-
structions are not passivizable, when he states (1985: 831) that "even
in transitive sentences, passives often fail to apply if the subject is not
an agent." That is, he does not account for why not all prototypical
transitive clauses with agentive subjects find passive counterparts in
a language such as Korean. Our approach is based upon the func-
tional motivation that the richer a morphological case-marking sys-
tem a language has, the more degenerate the system of passivization
becomes.
The two transitive constructions in (9) and (10) tell us that inanimate
subjects are avoided in controlling the state of affairs in Korean dis-
course, regardless of the argument's thematic status. This animacy
constraint in the Korean diathesis might explain why the middle
voice coupled with reflexivization is not available in the Korean and
Japanese systems.3
One might argue that this kind of reasoning is not quite valid, given
the fact that compared to English alone, Dutch and German transitive
constructions show a relatively transparent semantic encoding system
of subject selection analogous to Korean, while German is nonethe-
less identified as a language with reflexive construction. This seem-
ingly plausible counter-argument is untenable, since we subscribe to
the view that the reflexivization as an important feature of subject-
prominence applies even to the German transitive construction in
(11) with passive reading. That is, in German, even inanimate patient
subjects trigger reflexivization in the middle reflexive construction.
The corresponding Korean construction (lib) employs a pseudo-
passive with a passive morpheme. Reflexivization in Korean is re-
stricted only to animate entities (cf. also Seong 1999b).
When it comes to passivization with respect to transitivity, there is
also a misunderstanding over the identification of the transitive ob-
ject. Hopper and Thompson (1980: 259) suggest that "the special
markings on definite objects, found in many languages, are better
interpreted functionally as signals of the high transitivity of the
140 Sang Hwan Seong
In both constructions in (12a) and (12b) we note that the dative noun
phrase precedes the accusative patient noun phrase as the unmarked
order.4 In German, this is only possible if the two combinations are in
free variation. This tells us that the more referential animate object
appears before the inanimate object, as far as the precedence rule is
concerned. However, this cannot be taken to mean that the dative
animate object is indeed the true transitive object of the verb, since
the dative marked argument cannot be promoted to the nominative
case in German or Korean passives (e.g. *Der Anwalt wurde das
Geld gegeben; Das Geld wurde dem Anwalt gegeben). Thus, the
identification of true transitive objects cannot be established unitarily
by means of syntactic tests such as passivization (cf. *Er wurde ein
Preis verliehen 'He was awarded a prize.'). This is in turn in conflict
Transitivity parameter and prominence typology 141
The two sentences in (13) have the same meaning (D. Shaffer, p.c.).
The only difference may be a slight difference in focus. In the first,
the focus is on letter: I wrote a letter (not something else) to Santa
Claus. In the second, the focus is on Santa Claus: I wrote Santa Claus
(not someone else) a letter. Since in (13b) the human recipient argu-
ment appears in the "object position" adjacent to the verb, this dative
object would be interpreted as the true object according to the logic
of Hopper and Thompson. Since the comparable nominal as an in-
herent case is not subject to promotion to the nominative argument in
German and Korean passivization, the attempt to establish the true
object in terms of animacy in ditransitive constructions does not suc-
ceed cross-linguistically. Thus, definiteness and animacy criteria as-
sociated with a high degree of transitivity cannot be a true diagnostic
for selecting a true transitive object, at least with respect to passivi-
zation in German and Korean (cf. G. *Er wurde geholfen 'He was
helped.'). In view of the criticism of the traditional transitivity theory,
the transitivity parameter should be constrained such that languages
with high semantic transparency and an overt case-marking system
proportionally correlate with less syntacticization.
The traditional analyses of passives centered around the formal
aspects of the constructions per se or the lexical properties of the
verbs involved (Haider 1985 for German and O'Grady 1991,
Klaiman 1988 for Korean). Contrary to the previous studies, the pre-
sent author attempts to account for the motivation behind the passive
system by setting the discourse category "topic" apart from the
grammatical category "subject".
142 Sang Hwan Seong
Korean is well known for its subject and topic-prominence (Li and
Thompson 1976). For the discussion of "subject and topic" the pres-
ent author adopts Chafe's (1976: 50) functional definition of "topic"
as something which "limit[s] the application of the main predication
to a certain restricted domain," or as something which "sets a spatial,
temporal, or individual framework within which the main predication
holds." In previous research (cf. Hawkins 1986, Müller-Gotama
1994), it was pointed out that a grammatically prescribed fixed word
order language like English shows more ambiguous surface syntax
with respect to the pragmatic functions compared to German and the
Slavic languages. In these approaches, however, the traditional no-
tions of subject and object still play a central role in interpreting the
core cases of grammatical organization. The present author maintains
that for a more adequate description of the typological regularities,
we need to set the notion of topic apart from the category of subject
(Seong 1999a, 1999c). Gundel (1988) in her elegant study also illus-
trates what kinds of strategies are available to mark "topic-comment
structures" across languages, following the tradition of Li and
Thompson (1976). However, she does not show how the relative de-
gree of topic-comment structure can vary among the languages in-
vestigated. In this section, we want to further investigate how the
word orders of English, German, and Korean are pragmatically
regulated in controlled dialogue situations. Let us compare the fol-
lowing question and answer pairs.
(15) German
Q: Was ist mit der Lampe passiert?
A1 : Der Hund hat sie umgeworfen.
Lit. 'The dog has it knocked over.'
Transitivity parameter and prominence typology 143
(16) Korean
Q: chundung-i etteke doinke-ya?
lamp-NOM how became-Q
[definite] [wh-rheme:focus]
Lit. 'What has become of the lamp?'
Al: (guguss-un) gae-ka nume-tturyu-ss-tta.
[it -TOP] dog-NOM fall-cause-pret-Decl.
[TOPIC] [FOCUS] [VERB]
A2: ??guguss-un gae-eyeuyhae nume-ci-ess-ta.
[it - TOP] dog-by fall-inchoative-pret-Decl
Lit. 'It was knocked over by the dog.'
(18) so sah er, das die Messer, die er geschliffen het, der Rück was
als die Schneid...
'Then he saw that the knives that he had sharpened, the back
was like the blade...' (taken from Shannon 1999)
Topic Comment
Figure 1. Diagram of topic-comment connection
gram shows the functional boundary of topic and comment. The di-
rection of the arrow in this diagram also captures the possibility that
the category of topic functions as the core argument of the predicate
as in SP languages. Schematic characterization of the English system
would be the pattern in which the two categories "subject" and
"topic" have converged with each other.
Gapping between unlike constituents with differing grammatical
relations also belongs to the general features of TP languages. Again,
we find a striking typological similarity between Korean and early
Germanic.
The Korean construction (19a) suggests that the initial constituent the
pasta can be construed as both the object in the first clause and the
omitted subject in the second clause. In ENHG example (19b) as
well, the "dative-marked noun phrase" dem Herren 'the Lord' in the
first clause functions as the controller of the missing "nominative
nominal" er 'he' in the second clause. The highlighted constituent in
Middle Dutch example (19c) represents the syntactic item omitted in
the conjoined clause. This suggests that in Korean, ENHG and Mid-
dle Dutch the zero noun phrase-anaphor is not syntactically restricted
148 Sang Hwan Seong
The fact that English grammar entails these opaque grammatical re-
lations tells us that the English transitive construction has undergone
enormous extension due to case syncretism in its history. Languages
differ in the way features of actions are encoded by lexical items.
However, it is no coincidence that German and Korean show a more
transparent overall encoding system for the relationship of verb and
object.
Split intransitive constructions in German and Korean also fit well
into our system of prototype parameters. The experiencer argument
of a mental state is readily encoded as a transitive subject in English
(e.g. The king likes pears). In German and Korean, however, the ex-
istence of the dative case is held to be responsible for encoding most
typically the more topical or semantically salient "experiencer argu-
ment" of a mental state in a valency network of lexical predicate.
< •
to the first subject. On the other hand, level 2 students seem to over-
generalize this principle so that we often find the following ungram-
matical sentence (24b) in which a subject noun with topic marker
should be mentioned sentence-initially for the utterance to be gram-
matical. In the sample text we find the first sentence (24a) with three
subjects (parents, older sister and older brother) which seem to func-
tion ambiguously as the discourse subjects of the sentence in (24b).
This pattern suggests that German learners of Korean may first over-
produce subjects (cf. level 1) and then gradually learn to drop them as
their proficiency improves. However, it is clear that one cannot freely
drop a subject in a context in which a bare noun with topic marker
pumonim-un 'parents-topic' as the subject should be specified as the
old information as evidenced in (24b).
Among the different types of TP features of Korean, topic con-
structions with the topic marker (n)un seem to be relatively harder to
acquire. Traditionally, it has been observed that the nominative
marker is usually construed with a neutral description, whereas the
topic marker triggers both a thematic and a contrastive reading
(Seong 1999a: 362). The two separate readings become stronger
when the topic particle is used twice in parallel constructions for
conveying the information of equal status. The following example
(25a) directly follows the sentence (24b) in the same sample text
taken from one of level 2 participants.
sikmulhaJg'a-ipnita
botanist-be
'My mother is a housewife. My father is a botanist at the
waterworks.'
b. emeni-nun jubu-i-mye apuji-nun
mother-TOP housewife-be-conjunction father-TOP
kepsuso-eyse sikmulhakja-ipnita
waterworks-at botanist-be
'My mother is a housewife and my father is a botanist at the
waterworks.'
While writing about her family, the German female student produced
(25a) instead of (25b). Since the parents of this student are previously
identified information in the preceding context, the two nouns
'mother' and 'father' in (25a) should be marked with topic marker
nun respectively. The expected grammatically correct version of
(25a) would be (25b) in which the first NP marked with nun triggers
a thematic reading, whereas the second NP with nun is associated
with contrastive reading 'on the other hand'. Thus, it is evident that
German students tend to overproduce nominative subject markers
and that the acquisition of Korean by German students is guided by
typological differences. We notice that the typological transfer of
subject prominent features takes place continuously in the acquisition
of Korean and that the pragmatically sensitive grammatical structures
of Korean are cognitively hard to acquire for German native speak-
ers.
As regards the double nominative constructions as in (17) and
(18), not a single instance is attested in the current database for level
1 and level 2 participants. It seems safe at this point to argue that
compared to the other TP features such as null subject and the use of
topic marker, the acquisition of double subject construction takes
place later, although this construction is accessible to the learner as
an early input.
In this section, we have observed that when German students learn
a TP language such as Korean, they rely on known linguistic knowl-
edge before identifying the TP features of Korean. For a SP language
Transitivity parameter and prominence typology 157
6. Conclusion
In this paper we have developed the idea that the transitivity pa-
rameter should be constrained such that pragmatically determined
word order languages with high semantic transparency proportionally
correlate with less grammaticization. The effects of this theoretical
position are also reflected in the L2 data, such that the kinds of L2
competence that need to be acquired going from languages at each
end of the typological continuum proposed in section 4 are either
predominantly constructional or predominantly discoursal. It is evi-
158 Sang Hwan Seong
Notes
* I thank René Dirven and the other two anonymous reviewers for their valuable
comments and suggestions on this paper. I am also indebted to John Attfield
and Noah Isenberg for carefully reading the manuscript. All responsibility for
errors remains of course my own.
1.As an anonymous reviewer points out, example la (intransitive verb) is of a
different class from lb and lc (transitive verbs). However, this difference is not
so relevant to this discussion because we are dealing with non-agentive subjects
in English.
2. Some examples from the old Germanic dialects are in order (Valentin 1999):
(1) jah warP in jainaim dagam... Gothic, Mk 1, 9 ('and it happened in
those days')
(2) uuerda thin uuilleo obar thesa uuerold alia. Old Saxon, Heliand 1604
('your will be done all over the world')
(3) a. that thi kind giboran...scoldi uuerdan. Old Saxon, Heliand 123
('that the child should be born')
b. nu ist Krist giboran. OS. Heliand 399 ('Christ is now born', 'Christ
has now been born')
(4) thaer hali-ern wearth tha geopenod and tha lac waeron in gebrohte.
Old English ('the sacred house was opened and the offerings were
brought in')
Transitivity parameter and prominence typology 159
3. Kuno (1973) illustrates the Japanese equivalent of the similar reflexive con-
struction in (11). However, he does not explain why such a grammatical differ-
ence exists between English and Japanese.
4. There is evidence that the serialization of dative argument + accusative argu-
ment in ditransitive constructions should count as basic and unmarked order in
German. Look at the following observation of Lenerz (1977: 44) which in-
volves the focus pattern of dative and accusative objects:
When the accusative patient precedes the dative recipient, the accusative argu-
ment has to be unfocused or unrhematised for the sentence to be grammatical.
However, no such restriction applies to the reverse order "dative + accusative"
serialization. Thus, the latter order should count as "unmarked" in the ditransi-
tive constructions. This is also confirmed by Eisenberg (1994: 422).
5. Interestingly enough, we find this regularity in the diachronic development of
German and Dutch. Shannon (1999) demonstrates, based on the following in-
stances that the ordering of pronominal object and nominal subject in the mid-
dle field of Middle Dutch (MD) was an unmarked case and that Modern Dutch
has largely given up this structure in favor of the opposite ordering.
Shannon also adds that German has preserved this pragmatically determined
word order considerably, but has seen a substantial decrease.
6. An identical grammatical phenomenon is confirmed in Japanese (Y. Kasai
[p.c.]). For details, see Seong (2000).
References
Abraham, Werner
1995 Diathesis: The middle, particularly in West-Germanic. In: Werner
Abraham, Talmy Givón, and Sandra Thompson (eds.), Discourse
Grammar and Typology, Papers in Honor of John W. M. Ver-
haar, 3-47. Amsterdam: John Benjamins.
160 Sang Hwan Seong
Burridge, Kate
1993 Syntactic Change in Germanic. Amsterdam: John Benjamins.
Chafe, Wallace
1976 Givenness, contrastiveness, definiteness, subjects, topics, and
point of view. In: Charles Li (ed.), Subject and Topic, 25-55.
New York: Academic Press.
Drosdowski, Günther (ed.)
1995 Duden. Grammatik der deutschen Gegenwartssprache.
Mannheim: Bibliographisches Institut.
Eisenberg, Peter
1994 Grundriss der deutschen Grammatik. Stuttgart: J. B. Metzler.
Faarlund, Jan T.
1992 The subject as a thematic category in the history of Scandinavian.
Folia Linguistica. XXVI. 1(2): 151-68.
Fillmore, Charles
1968 The case for case. In: Emmon Bach and Robert T. Harms (eds.),
Universals in Linguistic Theory, 1-88. New York: Holt, Rinehart
& Winston.
Firbas, Jan
1992 Functional Sentence Perspective in Written and Spoken Commu-
nication. Cambridge: Cambridge University Press.
Foley, William A. and Robert D. Van Valin
1984 Functional Syntax and Universal Grammar. Cambridge: Cam-
bridge University Press.
Givón, Talmy
1979 On Understanding Grammar. New York: Academic Press.
Goodluck, Helen
1991 Language Acquisition. Oxford and Cambridge: Blackwell.
Gundel, Jeanette
1988 Universals of topic-comment structure. In: Michael Hammond,
Edith Moravcsik, and Jessica Wirth (eds.), Studies in Syntactic
Typology, 209-242. Amsterdam: John Benjamins.
Haider, Hubert
1985 The case of German. In: Jindrich Toman (ed.), Studies in German
Grammar, 65-101. Dordrecht: Foris.
Hawkins, John A.
1986 A Comparative Typology of English and German". Unifying the
Contrasts. Austin: University of Texas Press.
Helbig, Gerhard and Joachim Buscha
1994 Deutsche Grammatik: Ein Handbuch för den Ausländer-
unterricht. Leipzig: Langenscheidt.
Transitivity parameter and prominence typology 161
Sydney M. Lamb
Like other papers in this collection, this one is concerned with learn-
ing, in particular, with the learning of syntax. It addresses the ques-
tion of how children, or adults learning a second language, learn to
handle what appear to be the syntactic categories needed for using a
language. And in order to talk about this question there is a very im-
portant prerequisite: We need to understand just what it is that is be-
ing learned.
According to a traditional view of the learning of syntax, the child,
or the adult second-language learner, must somehow acquire syntac-
tic rules. In one manner of speaking it is assumed that a speaker has
"internalized" such rules. The assumption is that since syntactic rules
are useful in descriptions of sentences, they must be present within
the system that produces them. This mode of thinking comes from an
unstated assumption, that patterns observable in linguistic data repre-
sent knowledge in the minds of those who produce such data. Is this
assumption supported by any evidence? I have a hard time finding
any basis for it. It is somewhat like supposing that since we can de-
vise equations for describing the movements of planets around the
sun, those planets must have internalized such equations. If we can
find other sources for the patterns found, there is no reason to adopt
this assumption (Lamb 1999: 227-247).
168 Sydney M. Lamb
Relatively short:
horse sense strictly speaking painfully obvious
no laughing matter a people person a no-brainer
not written in stone the bottom line a dumb question
as clear as mud a New York minute right then and there
Relatively longer:
round up the usual suspects it ain't over till it's over
if it ain't broken don't fix it you know what I mean
you can bet your bottom dollar the truth of the matter is
between a rock and a hard place been there, done that
2. Constructions
Now we are ready for our basic question, that concerning the nature
of categories. We have to ask just what information must a child (or
second-language learner) acquire in order to handle the syntactic
categories of the language. Does such knowledge consist, for exam-
ple, of a listing of the membership of the simple categories (like
noun, verb) together with a set of rules for generating the more com-
plex ones. If not, what? It is important to recognize that the notion of
category comes from analytical linguistics, an essentially non-
cognitive endeavor. As with other tools of analytical linguistics, we
are not obliged to suppose that they are internalized in the mental
systems of speakers.
Learning syntax - a neurocognitive approach 175
In keeping with what has been observed so far, we could rephrase the
question in terms of variable constituents rather than categories: How
does a language learner learn how to use variable constituents? But
for now, let us continue with the more traditional notion of syntactic
categories.
The first observation, an important one, is that as syntactic catego-
ries have been traditionally employed, they don't work. They are just
rough approximations - not wholly useless, just not cognitively plau-
sible. Approaching them with prototypicality notions helps, but ulti-
mately, the closer we look at any syntactic category, the more it
breaks down, until we get down to the individual lexeme.
Consider, for example, the category of prepositions. If they do
constitute a syntactic category, they behave alike syntactically. If they
do not, how can the category specify their combinations? So let us
take a look: We have in love but not *on love, yet we have on vaca-
tion but not *in vacation. The following are examples of the use of
basic prepositions with various objects for describing states that one
can find oneself in or in which one can do things. We have:
Preposition Object
*at, *by, ?in, *out of, ?under, with assurance
*at, *by, ?in, *out of, *under, with candor
?at, by, ?in, *out of, *under, *with chance
?at, *by, in, out of, *under, ?with danger
*at, *by, in, out of, *under, ?with desperation
*at, *by, in, out of, ?under, ?with doubt
*at, *by, in, ?out of, *under, ?with fear
*at, *by, in, ?out of, *under, with insecurity
at, *by, *in, *out of, *under, *with leisure
*at, *by, in, out of, *under, with love
*at, *by, in, out of, *under, *with pain
at, *by, in, out of, *under, ?with play
?at, *by, ?in, *out of, *under, with pleasure
*at, *by, *in, *out of, under, *with pressure
at, *by, ?in, *out of, *under, *with rest
*at, by, in, *out of, *under, *with thought
*at, *by, *in, *out of, *under, *with vacation
176 Sydney M. Lamb
Notice that the last noun listed, vacation, does not occur with any of
the prepositions considered here, but does occur with on, and that on
does not occur with any of the other nouns in the list. By the way, the
judgements presented are mine and they may differ in some details
from those of other native speakers. For a neurocognitive approach,
that is no problem, since the object of investigation in neurocognitive
linguistics is the neurocognitive system of the individual speaker, not
some disembodied "language"; and we recognize that the system of
every speaker differs from that of every other.
Given findings like those listed above, what cognitive sense can
we make of the notion that there is a construction utilizing the cate-
gories Preposition and Noun Phrase?
Syntactic categories are based upon an old tradition, that of the
"parts of speech", which goes back to the ancient Greeks. According
to the doctrine of parts of speech, every word must belong to one or
more of these categories - and there is a catch-all category "adverb"
for the difficult cases. Now, what about tantamount! According to
the ordinary dictionary, which treats the parts-of-speech myth as re-
ality, it is classed as an adjective, as is the equally unique akimbo.
Also of clearly unique distribution, but classed as a preposition, is
between.
Let's take a brief look at verbs. From the point of view of their
syntactic distribution there are clearly many different kinds. We
might be tempted to suppose that if we subcategorize to a sufficient
degree we will get down to subcategories whose members behave
alike. We might, for example, consider just the subcategory of verbs
of perception: see, hear, listen, smell, etc. But only a little observa-
tion makes it clear that even this tiny subcategory doesn't help us to
define what can and can't occur syntactically. Apart from the fact that
we see visible objects but hear audible things, we have the different
aspectual possibilities: You see and hear punctually and comple-
tively, but you listen [to] and look at duratively. We don't have to
look very far to see that each of these verbs, and indeed every verb of
any other subcategory, has its own distribution. And the same can be
observed about members of any of the other parts of speech.
Learning syntax - a neurocognitive approach 177
And so we conclude that every lexeme has its own syntax. Now that
is a conclusion which puts the validity of the concept of syntactic
category into considerable doubt. How can there be syntactic catego-
ries, as constituting part of the information used by speakers of a lan-
guage, if they don't work? And if they are useless, how can we ex-
plain why they come up so regularly in discussions of syntax? The
answer to this question is that such categories result from applying
the taxonomic techniques of analytical linguistics. Analytical lin-
guistics is concerned with analyzing and categorizing and describing
patterns found in things that people say. It is natural in such a pursuit
to classify things that appear to share properties. In such classifica-
tion it is easy to overlook that the things categorized together do not
really behave alike. In any case, there is no justification for assuming
that concepts of analytical linguistics can be taken over directly into
an understanding of the cognitive basis of language.
An alternative to the taxonomic approach is that most syntactic in-
formation, or even all of it, is attached to individual lexical items. In
that case, most of the syntactic generalizations that can be formulated
in rules are epiphenomenal, and the actual internal information that
gives rise to them is widely distributed, among thousands of separate
items. According to such a view the acquisition of syntactic knowl-
edge is the acquisition of lexical knowledge.
If it is the case that every lexeme has its own syntax, then it follows
that the only way to learn syntax is to learn lexicon. If this is so, then
what seemed to be a process of learning syntax is really just the
learning of vocabulary, a process that occurs one lexeme at a time.
This view makes sense not only because every lexeme has its own
syntax, but also because, as a consequence, you can't know how to
use a lexeme without knowing how it connects with other lexemes.
This conclusion is strongly supported by findings of Elizabeth
Bates et al. (In press), who have examined the correlation between
development of grammatical complexity and vocabulary size in chil-
178 Sydney M. Lamb
5. Participant roles
Let's consider the case of a simple verb like eat, accepting the obser-
vation that its syntactic distribution is unique. That being the case, we
are tempted to conclude that the operative knowledge used in pro-
ducing a sentence like Mommy's eating an apple, with its ordering of
the two participants in relation to the process - the agent before and
the patient after the verb - is in large part information connected di-
rectly with the lexeme eat. Yet it seems also to be the case that the
actor-action construction and the do-smthg-to-patient construction
(i.e. the transitive construction) are also involved, even if their in-
Learning syntax - a neurocognitive approach 179
child learns the lexeme Leat. For any normal child has such struc-
tures, and therefore a concept of what eating is, from early on in its
life. To be sure, the knowledge of what eating is may well undergo
further development as the child matures and learns more about the
culture, the mythology, stories, and so forth. Eventually a person may
come to accept, or to reject, such notions as an automobile eating
gasoline or a furnace eating coal. He may well come to appreciate the
riddle What has three legs and eats marbles? (Answer: A three-
legged marble-eater). Whether accepted or rejected, it depends not
upon the "category" <EATER> as such but on EAT. If the process can
be called eating, then whatever is performing it is the <EATER>. NO
separate knowledge of that "category" is needed.
The foregoing observations definitely simplify the task of the cog-
nitive syntactician, for they eliminate the whole problem of deter-
mining what knowledge must be learned to learn such „categories".
For the answer is that no additional knowledge is needed beyond
knowledge of the process itself. To be sure, there is still plenty to
investigate: What form does the knowledge of eating have in a per-
son's brain, and how is that knowledge acquired? But note well that
the problem of answering those questions was a problem already pre-
sent for cognitive neuroscience anyway. What I am claiming is that
no further knowledge beyond that is needed for syntactic purposes.
In the preceding two paragraphs I have started to put category in
quotation marks, because it is apparent that we are no longer talking
about what the term has commonly meant in discourse about syntac-
tic categories. The difference is clearly seen in the context of the
question of what information in the cognitive system gives rise to the
appearance of categories in analytical linguistics. More commonly
that information would be seen as, in one way or another, specifying
the membership of the category. That notion of category, as involving
one-and-many, however useful it may be in analytical linguistics, is
now seen to represent an illusion from the neurocognitive point of
view. For this reason I prefer the term variable, free from such con-
notations, and so I shall use it from now on.
Learning syntax - a neurocognitive approach 183
And so, as has been observed many times in the past, we have verbs
presupposing participants while nouns do not. And we have not just
theoretical reasons based on analysis of linguistic data for such an
assertion. It is surely related to the fact that Broca's aphasies, whose
area of damage is in the frontal lobe, typically have trouble not only
with phonological production and with grammar, but also with verbs,
much more so than with nouns. And they also have trouble with
prepositions and with "function words" generally.
All lexemes other than nouns evidently presuppose some other
constituent or constituents with which they normally co-occur, just as
eat presupposes <EATER> and <EATEE>. We have for example clause
introducers like clearly, which presuppose a following clause or
<ASSERTION>. Like it are other assertion introducers, including those
described above as lexemes with variable constituents. We now see
that what was written above in citing them was incomplete and that
they should be written as follows, to include the presupposed con-
stituent <ASSERTI0N> (with Lclearly also shown):
In the first of these examples we now see that there are three variable
constituents, not just <X> and <7>. We also observe that a variable
can range over a small number of values, like
<X> and <Y> of this example, as well as over a large number of val-
ues, like <ASSERTION>, <EATER>, and <EATEE>.
And so the difference between the construction and the lexeme with
variable constituents is not that great. In fact a construction might be
considered a lexeme with more than one variable constituent. Some
constructions also include one or more fixed constituents, others do
not. Among those which do are the way-construction (they made
their way to the door) and the passive construction, which includes
the verb be asa fixed constituent. And as we have seen, there are also
variables which have a very small number of values, which can be
enumerated, for example, it doesn't take a X to Y that <ASSERTION>,
in which both X and 7 have just a few possible values.
9. Concluding observations
References
Tomasello, Michael
1998 The return of constructions. Journal of Child Language 25: 431-
442.
Tomasello, Michael and Patricia J. Brooks
1999 Early syntactic development: A construction grammar approach.
In: Martyn Barrett (ed.), The Development of Language. Hove,
UK: Psychology Press.
Conceptual primes in
early language development
Cliff Goddard
1. Introduction
Many of the examples discussed in this paper are drawn from diary
notes of the language development of my son Pete (a pseudonym),
who was born in August 1996. The notes were made by myself and
by my wife, Mee Wun Lee, commencing (in earnest) from the time
just before the child's second birthday. Our procedure was not as
rigorous or systematic as some other diary studies, but it did yield a
continuous sampling of the child's utterances over an 18-month pe-
riod. I was the main record-taker. I tried to make a point, whenever I
was with the child, of having some notepaper and a pencil with me,
and I simply jotted down any utterance which seemed either typical
of the way the child was speaking at that time, or interesting in the
sense of showing "emergent" meanings or structures. Often some
notes on the context or apparent intention of the child were also nec-
essary. At different times, observations were made all over the house
- at the breakfast table, in the bath, in the living room, in Pete's bed-
room, and in the backyard, as well as in the car, at friends' houses, on
shopping trips, at playgrounds, and so on. This is, admittedly, a rather
Conceptual primes in early language development 197
haphazard and intuitive technique but it is not different from the pro-
cedure followed by many linguists in taking field notes of an indige-
nous language in a naturalistic setting.
Diary studies are most useful when they selectively focus on spe-
cific issues of interest to the investigators (Mervis et al. 1992). My
observations were guided firstly by an interest in semantic primes,
and secondly by an interest in general syntactic development. The
fact that one is taking a day-by-day record tends to make one sensi-
tive to new developments and thus more likely to record them; but on
the other hand there were many hours each day when the child was
not being observed, so the diary records cannot be taken to indicate
the earliest occurrences of any word or structure. In most cases they
probably pick up features which have been present for an unknown
previous time. (We did not record utterances which were simply
repetitions, in whole or part, of something which had just been said
by an adult.) My wife, who is a native speaker of Cantonese, often
spoke to the child in Cantonese when they were alone together; and
especially in the early days I also used to address the child using my
own rudimentary command of this language. Not surprisingly, a good
proportion of Pete's earliest words were Cantonese, but for the pur-
poses of this paper I have usually given English equivalents.
Unlike several children who feature prominently in the child lan-
guage literature, Pete is not an "early talker" but his general language
development seems fairly typical of a child growing up in a middle-
class Western household. From an examination of the corpus I have
divided his early language development into the following rough
stages.
Prelinguistic stage
Stage I (from about 14 months): single words
Stage //(about 21 to early-26): two word combinations
Stage III (early-26 to mid-29): some multi-word sentences, usually
limited to a simple clause with an adjunct; also, only one utter-
ance at a time.
Stage IV (mid-29 to early-32): two or more related sentences in a
row; begins to have conversational exchanges.
198 CliffGoddard
The present paper focuses primarily on the very early stages, i.e. the
Prelinguistic stage, Stage I (one-word stage) and Stage II (two-word
stage).
I YOU
PERSON PEOPLE BODY
THING/WHAT PART KIND
THIS OTHER MUCH/MANY ONE
THE SAME(TOO) TWO ALL SOME
GOOD
BAD
BIG
SMALL(LITTLE)
VERY(SO)
WANT WORD(CALLED) KNOW
SEE MOVE HAPPEN SAY HEAR
DO TOUCH THINK FEEL
HAVE THERE IS DIE
MORE NOT(NO) CAN MAYBE IF
BECAUSE(COS)
NOW
BEFORE(FIRST) AFTER(LATER) A SHORT TIME A LONG TIME
(LITTLE WHILE)
WHERE/PLACE
HERE NEAR(NEXT TO) SIDE
INSIDE(LN) FAR
ABOVE(UP)
BELOW(DOWN)
LIKE
I hasten to note that the presence of a lexical exponent does not mean
that a prime is "fully acquired", in the adult sense. The child may
have active command over only a small part of the prime's syntactic
possibilities (as they exist in the adult language), with the result that
200 Cliff Goddard
water, go out, hurry up, milk, hand, foot, fly, ball, pick up, teddy, wet,
hot, don't want, see, come, nappy, light, this, and many others. From
the NSM point of view, if something is understood (in a linguistic
sense) then it is necessarily understood in terms of some conceptual
primes. In short, the child must have a certain "conceptual vocabu-
lary" of prime concepts even before the onset of intelligible words.
Theoretically, this is all well and good, but from a methodological
point of view it is problematical. How can we identify conceptual
primes which may be active in the child's mind, in the absence of
tangible surface exponents? As far as I can see, there are two possible
sources of linguistic evidence: (i) semantic analysis of the child's
production vocabulary, and (ii) semantic analysis of the child's com-
prehension capabilities. The second option presents even more meth-
odological difficulties than the first, and I have nothing to say about it
in this paper. What I will try to do is to undertake semantic analysis
of words in the child's non-prime production vocabulary, for if we
can determine the meaning of these words this would furnish direct
evidence of the conceptual vocabulary of the child at that age. In this
I am taking a lead from Tien (1999), who, as far as I know, was the
first to argue that conceptual primes may be "latent" in a child's early
lexicon, in the sense of being hidden or implicit in the meanings of
other, non-prime words.
The task may seem like a daunting one, but it is not altogether dif-
ferent from that which faces a field linguist who undertakes semantic
analysis of an unknown adult language from an unfamiliar culture. In
either case we have to begin with close naturalistic observation of
usage: documenting the range of contexts in which a certain expres-
sion is used - and not used - and comparing the usage of alternative
expressions which can be found in different contexts. Then we ex-
periment to discover the most economical semantic explications
which match the attested range of usage. Obviously one must guard
against the assumption that the child's meaning for a particular form
corresponds to the adult meaning, i.e. against "adultocentrism", the
child language analogue of ethnocentrism.
In this paper I deal only with a fairly small number of early child
words and utterances, and the fine details of the analyses in many
202 CliffGoddard
cases remain open to question. Even so, the exercise supports some
highly specific proposals about which primes emerge first at a con-
ceptual level and about the time lag between conceptual acquisition
and lexicalization.
WANT THIS
DO THIS
SEE THIS
mama, papa
bath!
nati 'hot/cold' (from Cantonese 'hot')
oh-oh!
bird! (Cantonese)
duck! (Cantonese)
broom-broom! 'car'
What can we infer about the meanings of these words, when Pete was
18-20 months of age? Obviously we cannot attribute to the child
anything like the semantic complexity of the comparable words in the
adult lexicon. Nevertheless it is necessary to attribute some meanings
to them, presumably the simplest conceivable meanings which are
compatible with their range of use in his speech.
The words mama and papa surely involve - minimally - the ele-
ment SOMEONE(PERSON), presumably in combination with THIS. What
else? One possibility would be to interpret mama and papa as, so to
speak, proto-names; i.e. to attribute a semantic structure along the
following lines (the use of inverted commas around mama in the ex-
plication is intended to refer to the sound of the word only).
mama =
this person
this person is called "mama" (i.e. word for this person is "mama")
papa =
this person
this person is called "papa" (i.e. word for this person is "papa")
206 CliffGoddard
ictic elements like THIS, HERE, and NOW. Also implied is the loca-
tional (WHERE) relationship.
bath! =
I see something
something is happening here
I (can) be in water (this 'stuff) now
nat! ('hot/cold') =
I don't want to touch this
when I touch this, I feel something bad (or: I feel-bad)
In his one-word period, Pete uttered the word oh-oh! when he saw
that 'something bad' had just happened to something, typically that
something had fallen over, been broken, dropped or spilled. Needless
to say, this usage was heavily modeled for him by his parents, but
Pete's oh-oh! embodied a simpler semantic structure than adult us-
age.7
oh-oh! =
something bad happened to something now
208 CliffGoddard
Coming now to bird, duck, and car, we see what first appear to be
clear examples of "nominal" terms, in the sense of words which are
clearly and only used about things. In their very earliest uses, how-
ever, such words are used essentially to make an observation about
immediate experience: the illocutionary frame of the one-word utter-
ance is Ί see (Consistent with this, Halliday [1975] placed
Nigel's early one-word utterances of this type under the "Personal"
(expressive) function.) As parents too, we tend to use words like
these to young children precisely to point out things: 'Look baby -
(a) bird!'.
Clearly these words can have only a very simple structure com-
pared to the enormous complexity of the adult words (cf. Wierzbicka
1985, 1996). Significantly, all three words refer to things that move.
For bird and duck, I suggest a further salient feature is, so to speak,
the locus of movement. Birds move up in the air; ducks move in the
water. As for broom-broom 'car', I suggest its salient feature is that
there is a person inside it.
bird! =
I see something
it can move up-high
this kind of thing is called "bird" (i.e. word for this thing is "bird")
duck! =
I see something
it can move in water
this kind of thing is called "duck" (i.e. word for this thing is "duck")
broom-broom! 'car' =
I see something
it can move
someone can be inside it
this kind of thing is called "broom-broom" (i.e. word for this thing is
"broom-broom")
Conceptual primes in early language development 209
Notice that these explications require the element KIND OF, even
though the word kind does not "surface" till Stage V. Otherwise, they
would be depicting the words bird, duck, and broom-broom as proper
nouns, rather than as designations for recognizable classes of things.
This is consistent with mounting psycholinguistic evidence for
genuine categorization, i.e. categorization by kinds rather than simply
by perceptual similarity, in very young children (Gopnik and
Meltzoff 1997: Chapter 6, cf. Markman 1989, Keil 1989, Mervis
1987).
Even from this small sample, then, we can see the implied pres-
ence of a dozen-and-a-half primes which do not surface as words in
their own right for some months to come: I, SOMEONE(PERSON),
SOMETHING(THING), KIND, THIS, HERE, NOW, SEE, WANT, DON'T-
WANT, BAD, HAPPEN, MOVE, TOUCH, CAN, ABOVE(UP), INSIDE, and
WORD(CALLED). TO this list we can add DO, MORE, and FEEL-GOOD,
which were indicated already in the proto-linguistic period (even
though they have not turned up in the handful of Stage I words we
have just looked at). At the one-word stage, it seems, Pete has a con-
ceptual vocabulary of at least 20 semantic primes, about one-third of
the eventual adult inventory.
In the three-month period between about 26.07 and 29.15 Pete learnt
to say a large number of new words, and increasingly to produce two-
word combinations (though he also continued to use a lot of single-
word utterances). As set out in Table 2, by the end of Stage II the
child had about 21 exponents of primes in his production vocabulary.
This set is not sufficient, however, to plausibly explicate the wow-
prime vocabulary of Stage II. Rather, semantic analysis of his non-
prime vocabulary suggests that Pete's conceptual vocabulary at this
period already included most of the primes posited to appear in Stage
III, plus several others which would not appear till subsequent stages.
First, it is worth pointing out a few fairly simple "semantic mole-
cules" (cf. Wierzbicka 1995):
210 Cliff Goddard
fu-fu (pants) =
something
a person does something with it
afterwards the bottom part of a person is inside it
saam (top) =
something
a person does something with it
afterwards the top part of a person is inside it
The terms 'top part' and 'bottom part' are obviously based on the
primes ABOVE(UP) and BELOW(DOWN). The simplest thing would be
to regard 'top' and 'bottom' as simply "adjectival" variants of ABOVE
(UP) and BELOW(DOWN), respectively.
The need for the prime PART is surely clear from the proliferation
of body-part terms, such as mouth, eyes, nose, ear, head, and foot.
However, I do not think we have to posit the prime BODY at this
stage; it seems enough to explicate hands, for example, as 'parts of a
Conceptual primes in early language development 211
eyes - mouth =
two things part of a person
they are part of a person things can 'go' inside a person there
because of these things, a person can see a person can do things (to things)
with this part
hands = head =
two things one thing
they are parts of a person it is part of a person
a person can do many things with them it is above the other parts
Another couple of Pete's early nominal words are wheel and door.
The former is really a favorite word for Pete at age two, and perhaps
would qualify as the prototypical part-term in relation to physical
objects. An interest point about both words (especially clear in the
case of door) is that they seem to require the semantic component
SIDE. 8
wheel = door =
part of something something
it is round (i.e. when you see it, it is in a place
it is the same on all sides) someone can be on one side of it
it moves after this, it moves
when it moves, the other thing moves after this someone can be on the other
side of it
212 CliffGoddard
Like many other children, in his two-word stage Pete began to use a
largish number of animal names, such as horsie, dog, pig, cow, seal,
cat, and monkey. To analyze these meanings in detail would be a fas-
cinating project, which I believe would show that, even at this early
stage, such words can involve semantic components describing the
animal's size, some salient body-part features, characteristic sound,
reference to habitat and (at least in some cases) reference to its typi-
cal food. This kind of study is beyond the scope of the present paper,
however, and for present purposes I mention these words only to
make the point that they further attest to the conceptual presence of
the notion of KINDS. I will move instead to some explications for a set
of "verbal" words. As with the nomináis, this listing is not exhaustive
but it is a broad enough sample to indicate the range of semantic
components which are needed.
Perhaps the expression come here! is not entitled to be termed
fully "verbal", since it is, at this stage, essentially an imperative for-
mula; and the same applies to help me! Even simple structures such
as the following make it plain that the prime YOU is called for. In-
deed, one could claim that any example of genuinely "addressee-
directed" speech implies YOU - for YOU is what an "addressee" is.9
The expressions fall down and bump both seem to imply the element
HAPPEN. At age two, Pete used fall down not only about himself (af-
ter he fell), but also about something he threw or dropped. Bump not
only seems to imply HAPPEN, but also TOUCH (without TOUCH, it is
hard to see how the "contact" aspect of the event could be captured).
The explications also seem more plausible with an explicit causal
component BECAUSE OF THIS (as shown).
Conceptual primes in early language development 213
fall down =
something happened to this thing/person
because of this, it is down now
bump (head) =
something happened to me
part of me (my head) touched something now
because of this, I feel something bad now
broke =
something happened to it
because of this, it is not one thing any more
214 CliffGoddard
Give and make are an interesting pair. They both seem to demand an
explicit "before-and-after" scenario: the situation as it was before is
changed as a result of someone doing something. In the case of give,
the change concerns possession (i.e. someone having something); in
the case of make the change concerns the existence of something.10
I make house =
there wasn't a house here before
after this I did something with some things
because of this, there is a house here now
(I) ate it =
something was in my mouth
I did something to it with my mouth
because of this, it is inside me now
words surely call for HEAR. Second, this was a time when Pete was
already quite interested in the characteristic sounds made by animals
and birds: "quack-quack" for ducks, "tweet-tweet" for birds, "moo-
moo" for cow, "oink-oink" for pigs, and so on. He could produce any
of these sounds at around two years of age. Surely this implies
something like (for example):
"quack-quack" =
ducks do something
when they do it, a person can hear something like this: "quack-
quack"
There are also two primes, specifically, KNOW and SAY, whose pres-
ence is implied by functional (illocutionary) facts. They do not ap-
pear in Pete's production lexicon for three or four months, but their
conceptual presence is implied by the fact that the child begins, in
Stage II, to ask simple information-seeking questions (both polar and
wh-), as shown below, and also to respond appropriately to questions
asked by adults. The illocutionary intention of questions, conveyed
by intonation, involves the components: Ι WANT TO KNOW SOME-
THING, I WANT YOU ΤΟ SAY SOMETHING. (Interestingly, all the re-
corded examples of Pete's early questions concern location. I am not
sure what to make of this: perhaps that is just the main kind of thing
he wanted to know about.)
25.26 Moon there? Roo there? (re. a photo, i.e. is that the moon
there? is that a kangaroo there?)
26.12 (CG told Pete they were going to the coast tonight, to the
sea) Pete: Seal there? CG: No, no seals. Pete: Boat there?
26.17 Mama slippers where?
26.18 CG: Pete, we're going to Jaew tse-tse's place for dinner.
Pete: Baby there?
28.06 Where plane?
primes at Stage II, despite the fact that they have not yet surfaced as
individual words. They are presented in three groups: those which are
destined to appear as words in the next stage (Stage III), i.e. within a
couple of months; and those which do not appear until subsequent
stages.
In the case of the Stage III group, what we are seeing is virtually the
entire list of primes which appear at Stage III (with two exceptions:
NEAR(NEXT TO) and FAR). This hardly seems like a surprising result
any more. We saw the same pattern in relation to Stage I; and, as
mentioned earlier, it is a pattern which is consistent with the fact that
comprehension competence runs several months ahead of production
competence. The conceptual presence of primes which only surface
lexically four, five, or six months afterwards, however, does seem to
call for some special explanation (see below).
3. Discussion
their own conceptual primes which are different from, or even in-
commensurable with, the adult system? I will conclude by saying
something about these theoretical questions.
One cannot a priori rule out the possibility that the young child's
semantic system is incommensurable with the adult system. On the
other hand, practical experience shows that there can be a high de-
gree of mutual understanding between young children and adults. A
radical incommensurability thesis in relation to young children would
also face the problem of accounting for the developmental continuity
of child and adult understanding. In my view, the issue must remain
open for the time being. One thing is certain, however. We will never
get anywhere at resolving the issue unless we are willing to knuckle
down and attempt serious semantic analyses of early child language.
If we try, and fail, using a representational system which assumes
some limited conceptual continuity with the adult system, then this
will be evidence for incommensurability - and vice versa if we suc-
ceed. Above all, we have to try; and I submit that the NSM system
offers a highly facilitative framework for doing just that.
On the second point, it is not really correct to say that the NSM
system uses adult primes to explicate child language speech (thus
imposing an adult point of view). From the present study it would
seem that the meanings of a child's early words can be adequately
paraphrased in terms of the language of the same child - as it will be
a few months down the track. That is, even when the explications
cannot be framed entirely within the child's own production vocabu-
lary (which may be the case until the child is, say, four years of age),
they can at least be framed in terms which will soon be part of that
child's production vocabulary. Another way of putting it is that al-
though the language of toddlers and young children is not yet "meta-
semantically adequate" (as, apparently, all adult languages are) it
does not stay that way for long. There is every indication that the
entire NSM metalexicon exists in the production competence of a
four-year old child. If so, the continuity/incommensurability issue
should not be pitched in terms of child vs. adult, but in terms of two-
year old vs. four-year old. In this respect, the NSM primes retain a
significant advantage over other representational systems such as
220 CliffGoddard
Acknowledgements
Mee Wun Lee has made a substantive contribution to this paper, es-
pecially in relation to the semantic analyses of Pete's early words.
Vicki Knox also made a number of helpful suggestions. I am grateful
to Anna Wierzbicka, René Dirven and Nick Enfield, who read and
made valuable comments on earlier drafts, to participants in the
LAUD Symposium held at Landau, Germany, in March 2000, and to
an anonymous reviewer.
Notes
1.A similar position was long held by Dan Slobin (1985), Melissa Bowerman
(1985), and others, but in recent years they have begun to repudiate their earlier
emphasis on an innate prelinguistic conceptual basis for language acquisition, in
favor of an emphasis on the characteristics of the adult "input" language, espe-
cially its language-specific aspects (cf. Bowerman 1996, Slobin 1997). In my
view there is no necessary conflict between these two positions, but it is not
possible to pursue the matter here.
2. The NSM bibliography is extensive and cannot be reviewed here for reasons of
space. Aside from works cited elsewhere in the chapter, representative works
include: Ameka (1990), Chappell (1986), Goddard (1996, 1997), Harkins
(1990,1996), Hasada (1994), Peeters (1993), Travis (1998), Wilkins (1986).
3. It is true that in the NSM model it can make sense to speak of "universal syn-
tax" or "universal grammar" (cf. Goddard and Wierzbicka forthcoming). For
some readers such expressions may carry a connotation of autonomous syntax,
so it is perhaps worth stressing that what is intended is literally a "conceptual
syntax". What we are trying to say is that certain combinations of primes neces-
sarily make sense and should be expressible in all languages, e.g. 'something',
'do', and 'good' can combine to form 'do something good'; 'say', 'something'
and 'someone' can combine to form 'say something to someone'. Importantly,
the potential for these combinations is inherent in and springs from the mean-
Conceptual primes in early language development 221
mama =
this person
when I see/want this person, I say "mama"
7. Adult Oh-oh! can be used in a broader range of situations; for example, I could
say Oh-oh! upon reaching into my pocket for my keys and realizing that they
are not there. Aside from conveying the idea that something bad and unforeseen
is imminent, there is also a component of "minimisation" (Goddard 1998a:
190).
222 CliffGoddard
References
Ameka, Felix
1990 The grammatical packaging of experiencers in Ewe: A study in
the semantics of syntax. Australian Journal of Linguistics 10(2):
139-181.
Conceptual primes in early language development 223
Mervis, Carolyn B.
1987 Child-basic object categories and early lexical development. In:
U. Niesser (ed.), Concepts and Conceptual Development: Eco-
logical and intellectual factors in categorization, 201-233. Cam-
bridge: Cambridge University Press.
Mervis, Carolyn B., Cynthia A. Mervis, Kathy E. Johnson and Jacquelyn Bertrand
1992 Studying early lexical development: The value of the systematic
diary method. In: Carolyn Rovee-Collier and Lewis P. Lipstitt
(eds.), Advances in Infancy Research, Vol. 7, 292-378. Nor-
wood, NJ: Ablex.
Ninio, Anat
1992 The relation of children's single word utterances to single word
utterances in the input. Journal of Child Language 19: 87-110.
Oshima-Takane, Yuriko
1999 The learning of first and second person pronouns in English. In:
R. Jackendoff, P. Bloom and K. Wynn (eds.), Language, Logic
and Concepts, 373-410. Cambridge, M.A.: MIT Press.
Peeters, Bert
1993 Commencer et se mettre à: une description axiologico-concep-
tuelle. Langue française 98: 24-47.
Plank, Frans
1989 On Humboldt on the dual. In: R. Corrigan, F. Eckman and M.
Noonan (eds.), Linguistic Categorization, 293-333. Amsterdam:
John Benjamins.
Reddy, Vasudevi
1999 Prelinguistic communication. In: Martyn Barrett (ed.), The De-
velopment of Language, 25-50. Hove, UK: Psychology Press.
Schieffelin, Bambi B. and Elinor Ochs (eds.)
1986 Language Socialization across Cultures. Cambridge: Cambridge
University Press.
Schwartz, R. and L. Leonard
1982 Do children pick and choose: an examination of phonological
selection and avoidance in early lexical acquisition. Journal of
Child Language 9: 319-36.
Scollon, Ronald
1976 Conversations with a One-Year-Old: A Case Study of the Devel-
opmental Foundation of Syntax. Honolulu: University Press of
Hawaii.
1979 A real early stage: An unzippered condensation of a dissertation
on child language. In: Elinor Ochs and Bambi B. Schieffelin
(eds.), Developmental Pragmatics, 215-228. New York: Aca-
demic Press.
Conceptual primes in early language development 227
Slobin, Dan I.
1985 Crosslinguistic evidence for the language-making capacity. In:
Dan I. Slobin (ed.), The Crosslinguistic Study of Language Ac-
quisition, Vol. 2: Theoretical issues, 1157-1256. Hillsdale, New
Jersey: Lawrence Erlbaum.
1997 The origins of grammaticizable notions: Beyond the individual
mind. In: Dan I. Slobin (ed.), The Crosslinguistic Study of Lan-
guage Acquisition, Vol. 5: Expanding the Contexts, 265-323.
Mahwah, New Jersey: Lawrence Erlbaum.
Tien, Adrian
1999 Early lexical exponents and 'related' lexical items as manifesta-
tions of conceptual/semantic primitives in child language. MA
Thesis. Australian National University.
Tomasello, Michael
1987 Learning to use prepositions: A case study. Journal of Child Lan-
guage 14: 79-98.
1992 First Verbs: A Case Study of Early Grammatical Development.
Cambridge: Cambridge University Press.
Tomasello, Michael and Josep Call
1997 Primate Cognition. Oxford: Oxford University Press.
Travis, Catherine
1998 Omoiyari as a core Japanese value: Japanese-style empathy? In:
Angeliki Athanasiadou and Elzbieta Tabakowska (eds.), Speaking
of Emotions: Conceptualization and Expression, 55-82. Berlin:
Mouton de Gruyter.
Wierzbicka, Anna
1972 Semantic Primitives. Translated by A. Wierzbicka and J. Bese-
meres. Frankfurt/M.: Athenäum Verlag.
1980 Lingua Mentalis: The Semantics of Natural Language. Sydney:
Academic Press.
1985 Lexicography and Conceptual Analysis. Ann Arbor: Karoma.
1995 Universal semantic primitives as a tool for the study of language
acquisition. Unpublished MS.
1996 Semantics, Primes and Universals. Oxford: Oxford University
Press.
Wilkins, David
1986 Particles/clitics for criticism and complaint in Mparntwe Arrernte
(Aranda). Journal of Pragmatics 10(5): 575-596.
No preposition required.
The role of prepositions for the understanding of
spatial relations in language acquisition
Katharina J. Rohlfing
1. Introduction
they can nevertheless let us get an insight into the learning processes:
not the language-specific content of the spatial knowledge is more
focused, but the (probably) universal processes which include some
cultural-dependent aspects.
2. Non-linguistic strategies
that are put on the plate (food or cups) obtain an active relational
character.
Chris Sinha (1982, et al. 1999) determined the relational character
of objects on the basis of cultural differences. He focuses on the ca-
nonical orientation of objects as artefacts in a given culture. If, for
example, an object is used as a container, the canonical orientation
will be as follows: the object is supported by a surface, its top is open
and the infants are inclined to put things in this object. Therefore it is
very unusual for infants to use this object upside-down as a support-
ing object and to put things on the bottom of it. The canonical rules
help children to handle the objects, i.e. they know in which orienta-
tion to use the objects to achieve something (e.g. how to hold a pen
to write). Observing this, children develop an extensive knowledge
about objects and the activities associated with them and try to pre-
serve it. For the task in my study it means that as soon as the child
recognizes an object (the trajector and/or the landmark), correspond-
ing background knowledge (Sinha 1983) is activated. This back-
ground knowledge about one object (e.g. the landmark) can dominate
over the relational behavior in the sense that the object serves as a
target and motivates the motion needed in order to reach this target in
a relation, while it does not matter what the trajector is and if it suits
this relation (e.g. the TR-object is too big for the LM-object).
To summarize, the non-linguistic strategies discussed here have
the claim in common that they more strongly determine infants' be-
havior than their lexical knowledge3 does. To examine how strong
the influence of language is, the following study was designed.
3. Experiment
certain spatial relation; the other also consists in a request, but in this
case it is syntactically incorrect: the preposition is omitted.
The choice of preposition was guided by the order of prepositional
acquisition in German (e.g. Thiel 1985) and in English (e.g. Clark
1973, Sinha et al. 1999). According to the order, the preposition IN is
acquired first, the preposition ON / AUF second. This can also be
observed in Polish for the prepositions DO (IN) and NA (ON)
(Rohlfing 1998: pilot study). In Polish, however, there are two prepo-
sitions equivalent to English ON: a dynamic and a static one. Both
are verb-derived but the latter depends on the motion of the verb and
describes the spatial movement, whereby the former stands for a
fixed location or a state. The dynamic or static character of a prepo-
sition determines the inflection of the cases (see below); in the case
of the IN-preposition there are also two different lexemes: DO for a
dynamic IN-preposition and NA for a static one. In this study only
the dynamic form NA has been used, because the static form involves
more syntactic complexity in the instructions4. In the following it will
be marked with "->" to distinguish it from the static form.
The other motivation for the choice of this preposition is the fea-
ture of inflection in Polish. Spatial relations are created with the aid
of a preposition and, depending on the object's gender, with inflec-
tion of the cases. An instruction which contains this relation is:
3.1. Stimuli
In the second situation in which infants were tested the infants were
confronted with an abstract wooden construction. The construction
can be said to be abstract or 'not yet specified' because the child has
never seen it before. Furthermore, it was designed not to provoke a
certain relation with its physical properties and to give as little con-
text as possible coming from the relational character of the objects
used. Especially for the choice of relational objects in this wooden
construction, it was important for them not to be associated with a
relation, as e.g. an object with a surface associates the relation ON.
Thus, two balls were used: the first was installed in a wooden con-
struction named HiK5 and covered with Velcro®; the second was mo-
bile and could be related (and attached) to the first one. Balls suggest
few spatial relations, because they have neither a horizontal surface
nor are they containers. They suggest only the well-known activity of
"rolling".
Figure 1. The ^ - S i t u a t i o n
SITUATION
well-known abstract HiK
without preposition (INS„p) AKA, AKA 3
INSTRUCTION
with preposition (INSNA->) AKA 2 AKA 4
AKA: Number of correct performances according to the instruction
The order of the set (HiK, then sets 1, 2, 3 from the well-known
situation) corresponded to the progressive interest of the children. For
many infants the //zX-situation was too abstract and it was doubtful
whether they would have wanted to continue to play with it, after
they had seen the sets 1-3. Consequently, every child was seated in
front of the HiK first. To attract her/his attention, questions like
"What is it?" were asked. Then the names for the objects (sphere and
ball) were introduced. Finally, instructions were given that were neu-
tral in their formulation. They always began with Daj... 'Give...'.
This formulation is semantically correct in Polish. Other verbs like
Postaw... 'Put...' imply a state in which one object is supported by
another (e.g. lies on a table).
3.3. Predictions
3.4. Participants
3.5. Scoring
expected (e.g. POD [UNDER] instead of NA->). The line 'used dif-
ferent toy' shows how many children performed the NA-» relation
with different toys as provided by the set (e.g. they put the horse ON
the table instead of ON the bridge) — this performance was excluded
from the final scoring. The term 'anticipation' refers to a situation in
which a child performed the NA-» relation before the instruction
could be expressed (as soon as both objects were presented). This
reaction is also omitted for the final scoring.
INSTRUCTION
performed NA-> 14 13 14
INSwp other relations 0 1 0
used different 6 3 3
invalid 4 7 7
performed NA-> 13 9 16
INSna_» Anticipation 2 5 2
other relations 0 1+26 0
used different 3 4 3
invalid 6 3 3
performed contact 14
INS
"τ other relations 5
INSTRUCTION performed N A - > 4~
INSNA-> another action 4
another action 11
4. Results
INSwp (AKA,)
solved non-solved
INS na ^ solved 48 1
(AKA2) non-solved 1 0
Because the sum of the interesting fields for the McNemar-test for
dependent samples is 13 and therefore smaller than 60, the Binomial-
test was applied to this data. Assuming a significance level of 0.05
the results of the Binomial-test indicate that the tasks were equally
difficult. Moreover, the results suggest a tendency (-3 = 0.035) against
the second hypothesis made for the abstract situation (if the preposi-
tions supply important information, then the infants will pay more
attention to them and the task INSNA-> will be easier for them to per-
form). In the abstract situation, it seemed to be easier for the infants
to react to an instruction without preposition because this type of
instruction implies only a contact between the sphere and the ball. In
contrast, the type of instruction implying a concrete location of the
ball on the sphere was difficult for the infants to carry out. Thus, the
second hypothesis that infants need prepositions to understand loca-
tive instructions in an abstract situation cannot be borne out. On the
contrary, the results tend to mean just the opposite: in the HiK-
situation the location and the appropriate preposition of the ball on
the sphere do not have a meaning; only the activity of attaching mat-
ters to the infants.
5. Discussion
The results from the well-known situation show that the ΝΑ-»
preposition does not have a prior role in understanding locative in-
structions. To solve the tasks, the relational character of objects used
in the trials was crucial. As a consequence of the instructions without
prepositions, infants carried out the NA—>· (ON)-relation as implied
by the LM-object. Only in two cases during the cup / plate-set did the
infants show other reactions. Their 'mistakes' were interesting in
their own way. Marcel (23 months) received the instruction to put
"the cup plate" (saucer) and he put it the other way around, i.e. the
plate (saucer) ON the cup instead. I wondered about his reaction:
according to Thiel's relational character, should not the plate suggest
a passive ON-relation? After consultation with his mother it turned
out that she always makes tea that way: she puts the tea in a cup and
No preposition required 243
then covers it up with a saucer to stew. Thus, while solving the task,
this behavior imitated a situation that was well-known. In this exam-
ple, the non-linguistic strategy was established by the typicality of the
situation, depending on certain experiences of a child. According to
my observations during the study, the typicality of a situation consists
of a canonical spatial relation implying a canonical orientation of
objects as noticed by Sinha (1982: 140). The child perceives the
situation task-dependently, and the emerging relation is linked to a
certain activity and a canonical relation (in Marcel's case: plate ON
cup) with a canonical orientation (plate inverted, cup upright) of ob-
jects that depends on that situation. From the task-oriented situation
follows the relational character of objects, which, as described by
Thiel (1985), is dominated by the aspect of canonical orientation of
an object. So, if the child observes a specific kind of tea-making and
gets used to it (typicality of a situation), the child will associate the
plate with the activity of covering, and the relational character of the
plate is in this case: active, ON.
The results presented here confirm Clark's partial semantics hy-
pothesis to the extent that, at this stage, the infants' behavior is based
on the combination of partial semantic knowledge and non-linguistic
strategy. However, more than just the two rules formulated by Clark
(1973) that are based on physical properties of objects are involved in
the non-linguistic strategies; the typicality of a situation also belongs
to them, which is an extension of the relational character of objects
formulated by Thiel (1985) and includes cultural aspects as the ca-
nonical orientation of objects described by Sinha (1982).
The tendency to refute the second hypothesis on the basis of the
results from the abstract situation suggests that in a new and abstract
situation the prepositions in an instruction do not supply any addi-
tional information. Moreover, in this procedure, a task in which chil-
dren were expected to respond to a preposition directly and to specify
the location accordingly, seemed to be more difficult. Nevertheless, a
longer familiarization period could help to find out more about in-
fants' behavior in the ///X-situation. For the time being, the results
suggest that infants employ non-linguistic strategies to understand
locative instructions, and a linguistic instruction cannot be under-
244 Katharina J. Rohlfmg
Acknowledgements
The research reported in this paper took place in the Graduate Pro-
gram "Task-Oriented Communication" (GK 256) at the University of
Bielefeld, Germany and was supported by the Deutsche Forschungs-
gemeinschaft.
No preposition required 245
Notes
References
Bowerman, Melissa
1996a The origins of children's spatial semantic categories: cognitive
versus linguistic determinants. In: John J. Gumperz, and Stephen
C. Levinson (eds.), Rethinking Linguistic Relativity, 145-176.
Cambridge: Cambridge University Press.
1996b Learning how to structure space for language: A crosslinguistic
perspective. In: P. Bloom, M. Peterson, L. Nadel and M. Garrett,
(eds.), Language and Space, 385-436. Cambridge, Mass.: MIT
Press.
Choi, Soonja and Melissa Bowerman
1991 Learning to express motion events in English and Korean: The in-
fluence of language-specific lexicalization patterns. Cognition 41:
83-121.
Choi, Soonja, Laraine McDonough, Melissa Bowerman and Jean M. Mandler
1999 Early sensitivity to language-specific spatial categories in English
and Korean. Cognitive Development 14: 241-268.
Clark, Eve V.
1973 Non-linguistic strategies and the acquisition of word meanings.
Cognition3: 161-182.
Grieve, Robert, Robert Hoogenraad and Diarmid Murray
1977 On the young child's use of lexis and syntax in understanding
locative instructions. Cognition 5:235-250.
Langacker, Ronald W.
1991 Concept, Image, and Symbol. The Cognitive Basis of Grammar.
Berlin and New York: de Gruyter.
1998 Topic, subject, and possessor. Linguistic Notes from La Jolla 19:
1-28.
Sinha, Chris
1982 Representational development and the structure of action. In: G.
Butterworth and P. Light (eds.), Social Cognition: Studies of the
Development of Understanding, 137-162. Brighton: Harvester.
1983 Background knowledge, presupposition and canonicality. In: T.
Seiler and W. Wannenmacher (eds.), Concept Development and
the Development of Word Meaning, 269-296. Berlin: Springer.
Sinha, Chris, Lis Thorseng, Mariko Hayashi and Kim Plunkett
1999 Spatial language acquisition in Danish, English and Japanese. In:
P. Broeder and J. Murre (eds.), Language and Thought in Devel-
opment. Cross Linguistic Studies, 95-125. Tübingen: Gunter
Narr.
No preposition required 247
Slobin, Dan I.
1973 Cognitive prerequisites for the development of grammar. In:
Charles A. Ferguson and Dan I. Slobin (eds.), Studies of Child
Language Development Crosslinguistic. Evidence for the Lan-
guage-Making Capacity, 175-276. New York: Holt, Rinehart and
Winston, Inc.
Thiel, Thomas
1985 Räumliches Denken und Verständnis von Lokativen beim Spra-
cherwerb [Spatial thinking and understanding of locatives in lan-
guage acquisition]. In: Harro Schweizer (ed.), Sprache und Raum
[Language and Space], 184-208. Stuttgart: Metzler.
Wilcox, Stephen and David S. Palermo
1974 "In", "on", and "under" revisited. Cognition 3: 245-254.
The 'Graded Salience Hypothesis' in
second language acquisition
Istvan Kecskes
1. Objectives
3. The Study
3.1.1. Implicatures
The participants were given a test which consisted of two parts. The
first part focused on implicatures while the second part tested the use
of SBUs.
3.2.1. Implicatures
The first task presented the students with ten dialogs that included
implicatures with their immediate context whose meaning the stu-
dents were expected to explain.
1/2. Amy: Don't you think Jim drinks a bit too much?
Billy: Is the Pope Catholic?
1/6. Alan: Do you think Mr. Herbert will give us a quiz today?
Bob: Does the sun rise in the East?
1/8. Adam: How is Bill doing at his new job in the bank?
Jim: Oh, quite well, I think. He hasn 't been thrown in prison
yet.
Test 1
In Test 1 students were asked to describe the meaning of situation-
bound utterances without context:
Test 2
In Test 2 students were given dialogs with SBUs most of which have
both a literal and a figurative meaning as salient. The participants
were expected to recognize the most salient meaning in the given
situation. Some expressions from the previous list (S/2, S/3, S/5, S/6)
were repeated here to find out how student responses change if the
expressions are used in a situational context.
S/l 9. - Jim, do you think you can repair this coffee machine?
- Piece of cake.
Test 3
The last part of the test focused on SBUs in which the figurative
meaning is usually the most salient meaning. Some of these expres-
sions (S/25, S/29) are rarely used in their literal meaning, others
(S/26, S/27), however, quite frequently keep their original meaning.
In the dialogs below these formulae were given in their literal mean-
ing, and participants were expected to consider the literal meaning of
the expressions as intended meaning in the given situation.
S/25. - Bill, you must show me how to use this word processor.
- Sorry, I am busy. Help yourself on that.
S/26. - Jim, can I spend the night in your apartment? It's, kind
of, too late to go home now.
- You know jOH are always welcome.
S/27. - Mary, why don't you sit down? There is a chair over there.
- OK, OK, I prefer to stand.
S/28. - Andy, can you tell me why you left me in the store without a
word?
- Not now. Don't you see I am busy. I'll talk to you later.
4. Hypothesis
5. Data analysis
5.1. Implicatures
Table 1. Implicatures
One and the same function ("Isn't this obvious for you?") can be
lexicalized in a lot of different ways. This kind of implicature is not
necessarily culture-specific but the way it is lexicalized can be, and
this may result in an absolute misinterpretation of the implicature in
L2 processing. The pragmatic unit "Is the Pope Catholic?" is a good
The 'GradedSalience Hypothesis' 259
1/1. Mr. R: Have you finished with Mark's term paper yet?
Mr. M: Yeah, I read it last night.
Mr. R: What did you think of it?
Mr. M: Well, I thought it was well typed.
260 Istvan Kecskes
1/8. Adam: How is Bill doing at his new job in the bank?
Jim: Oh, quite well, I think. He hasn't been thrown in
prison yet.
Testi
In the first test SBUs were listed without context. According to the
GSH for information to be salient, i.e. to be foremost on one's mind,
it needs to be stored or coded in the mental lexicon (Giora 1997,
forthcoming). Stored information enjoys cognitive priority over un-
stored information such as new information or information inferable
from context. Consequently, it is the most salient sense of a word or
expression that is directly computable from the mental lexicon before
any extra inferences based on contextual information are put to play .
This approach claims that context has a limited role in the activation
of salient meanings because salience is a matter of convention, fre-
quency, and familiarity. Research (Rayner, Pacht and Duffy 1994,
Giora 1997) suggested that even when prior context is heavily biased
in favor of the less salient (e.g. less frequent, familiar and conven-
tional) meaning of an ambiguous expression, salient meaning is acti-
vated first. So salient meanings cannot be bypassed in LI processing.
Familiar metaphors and frequent SBUs often have both a figura-
tive and a literal meaning as salient. They activate both their literal
and figurative meanings simultaneously if both meanings are equally
salient. However, one of the meanings is canceled based on contex-
tual clues. In the first test the expressions could have more than one
salient meaning. However, in most cases the figurative meaning was
expected to be considered salient.
Both the NS and NNS responses support GSH. The majority of re-
spondents found the most salient meaning without context. The num-
bers demonstrate how important frequency of encounters with the
expression is for both NSs and NNSs. Even the NSs made their deci-
sions based on frequency when two salient meanings were possible.
For instance: "Take it easy", "I'll talk to you later". Where parallel
processing would be needed because two meanings are equally sali-
ent (for instance: "Get out of here", "Give me a break"), NSs are di-
vided in their decisions since there are no contextual cues to rely on.
NNSs, however, prefer literal meaning to figurative meaning, which
makes sense, and supports our hypothesis about NNS language proc-
262 Istvan Kecskes
In some cases there are clear differences between NSs and NNSs.
The expression "Give me a break" was interpreted by NSs as "I dis-
agree with you" (47%), or "Stop kidding" (53%). The NNSs, how-
ever, thought that the intended meaning was either "Let me rest"
(31%), or "Leave me alone" (27%). This also demonstrates that
NNSs are usually more comfortable with a meaning that is close to
the compositional meaning of the expression.
The fact that NSs and NNSs do not share a common socio-cultural
background also led to different interpretations of one and the same
expression. Quite a number of the NSs (20%) said that the expression
"Tim is going out with that blonde" was derogative. None of the
NNSs thought that way.
Test 2
In the second test context was biased for the figurative meaning of
SBUs no matter whether the figurative meaning or the literal mean-
ing of the expression was salient such as, for instance, in S/16 ("OK,
shoot"), S/17 ("Get out of here"), S/18 ("Come on").
"Give me a break"
Test 1. S/14. "Rest": 31% "Leave me": 27% Misinterpr.: 42%
Test 2. S/20. "Don't bother me": 61% "Want rest": 18% Misinterpr.:
21%
Test 3
In the third test SBUs were used in their less salient meaning which
was the original literal meaning. Usually it is very difficult to find
situations where an SBU is used in its original literal meaning. This
is what Gibbs (1980) called "literal uses of highly conventional ex-
pressions." Responses of NSs demonstrated that the less salient
meaning was processed sequentially with no problem if the context
was clear. In a couple of cases, however, where the situation was
open-ended, and the context was unclear sequential processing did
not always occur in the responses of participants because some NSs
and many NNSs processed the most salient meaning directly. From
this respect the S/31 "Give me a hand" situation deserves attention.
The 'GradedSalience Hypothesis' 265
NSs NNSs
SBUs Literal Figurative Literal Figurative Misinter.
down?
I'll talk to you later explain don't bother explain don't bother
57% 43% 57% 35% 8%
Stick around 100% 100%
I am not going out busy
with you now 100% 84% 10% 6%
OK, give me a hand extend a help extend a help
hand hand
76% 24% 41% 45% 14%
266 Istvan Kecskes
6. Conclusions
NNSs can hardly apply the principle of salience in the target lan-
guage. The lack or low level of conceptual fluency in the L2 forces
NNS to rely on linguistic signs rather than conceptualizations while
processing L2. The study demonstrated this very well. NNSs usually
had no difficulty identifying the compositional meaning of SBUs and
implicatures. Problems occurred when literal meaning was not the
most salient meaning. Then NNSs mapped target language expres-
sions on LI conceptualizations, which often resulted in misinterpre-
tation of expressions. Using the principle of salience, LI speakers
processed figurative meanings directly without falling back on literal
The 'GradedSalience Hypothesis' 267
The GSH claims that context affects comprehension after highly sali-
ent information has been accessed (Giora 1997, forthcoming). There
is empirical research that supports the hypothesis that available in-
formation is accessed initially in LI processing, regardless of con-
textual fit or speaker's intent (Keysar 1998; Keysar, Barr, Balin and
Paek 1998). This is not so in L2 processing where contextual cues
seem to have priority over salience for reasons discussed above. It is
usually the linguistic context that NNSs rely on, and this is a direct
consequence of the compositional interpretation of words and ex-
pressions in the target language.
Note
References
prototype 109, 123, 125-127, 129, speech event 18, 22-27, 32, 36, 43,
13 If, 134, 137, 148, 150-152 68f
proximal-distal spatial relations 65, spoken discourse 44, 110-112, 121,
78, 80, 82f, 86, 88f, 95, 99 125
pseudocleft 109-117,119f, 122-125 subject and topic prominence 139,
-piece 112-116,119-121,123f 142,152f
relational character of objects 232f, syntactic category 167-70, 173-177,
236-238,242-244 182, 189
response concepts 68-70, 75-80, 83, syntax 46, 58, 109, 142, 149, 167-
88f, 99 171, 174, 177f, 185-187, 189,
role marking system 194, 220f
-direct 151 temporality 57
-indirect 151 tense 3, 1 If, 16, 21-30, 32-37, 41-
salience 63-66, 69f, 72, 74, 77, 80, 43,45-59, 63-74, 80-83, 88-100,
84, 88f, 91-93, 96f, 99, 249, 257, 114, 170, 198
261,266f English - 63, 65-457, 69f, 91, 96-
principle o f - 249,257,266 98
salient meaning 249f, 253, 255-257, present - 11, 21-30, 32-37, 45,
260f, 264-266 47-50, 52-54, 56, 58, 67, 69-72,
second language acquisition 153f, 89f, 92-94, 97, 100
249f, 267 time-reference 63-66, 69, 71, 77f,
segmentation 184,189 80-82, 88-92, 94-96, 99
semantic transitivity parameter 131, 141, 152,
- prime(s) 194f, 197f, 202, 209, 157
216,218,222 transitivity parameter and prominence
-transparency 131,141,157 typology 131,152
sequential processing 250,264 two-word stage 198,212,217,222
situation-bound utterances (SBU) typicality of the situation 243f
250-256,261-266 understanding of instructions 244
socio-cultural background 252, 259, universal relation 229f
263 usage-based model 3, 5
spatial variable constituent 172-175, 183,
- configuration 13,229 185-187, 189f
- relation 65, 229f, 232, 234, 236, viewing arrangement 15-17, 19, 21,
243f 26f, 29f, 32-37,41-46, 69f, 93
Cognitive Linguistics Research
Edited by René Dirven, Ronald W. Langacker and
John R. Taylor
Mouton de Gruyter · Berlin · New York
This series offers a forum for the presentation of research within the per-
spective of "cognitive linguistics". This rubric subsumes a variety of con-
cerns and broadly compatible theoretical approaches that have a common
basic outlook: that language is an integral facet of cognition which reflects
the interaction of social, cultural, psychological, communicative and func-
tional considerations, and which can only be understood in the context of
a realistic view of acquisition, cognitive development and mental process-
ing. Cognitive linguistics thus eschews the imposition of artificial bound-
aries, both internal and external. Internally, it seeks a unified account of
language structure that avoids such problematic dichotomies as lexicon vs.
grammar, morphology vs. syntax, semantics vs. pragmatics, and synchrony
vs. diachrony. Externally, it seeks insofar as possible to explicate language
structure in terms of the other facets of cognition on which it draws, as well
as the communicative function it serves. Linguistic analysis can therefore
profit from the insights of neighboring and overlapping disciplines such as
sociology, cultural anthropology, neuroscience, philosophy, psychology, and
cognitive science.