[de Knop, Sabine] Cognitive Approaches to Pedagogi(BookFi.org)

Cognitive Approaches toPedagogical Grammar: AVolume in Honour of Ren

Dirven

Edited bySabine De KnopTeun De Rycker

Mouton de Gruyter

Cognitive Approaches to Pedagogical Grammar

Cognitive Approaches to Pedagogical Grammar

Applications of Cognitive Linguistics

9

EditorsGitte KristiansenMichel AchardRene Dirven

Francisco J. Ruiz de Mendoza Ibanez

Mouton de GruyterBerlin New York

Applications of Cognitive Linguistics

9

EditorsGitte KristiansenMichel AchardRene Dirven

Francisco J. Ruiz de Mendoza Ibanez


Cognitive Approachesto Pedagogical Grammar

A Volume in Honour ofRene Dirven



Cognitive Approachesto Pedagogical Grammar

A Volume in Honour ofRene Dirven



Mouton de Gruyter (formerly Mouton, The Hague)is a Division of Walter de Gruyter GmbH & Co. KG, Berlin

Printed on acid-free paperwhich falls within

the guidelines of the ANSIto ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Cognitive approaches to pedagogical grammar / edited by Sabine DeKnop, Teun De Rycker.

p. cm. (Applications of cognitive linguistics ; 9)Includes bibliographical references and index.ISBN 978-3-11-019595-8 (hardcover : alk. paper)1. Language and languages Study and teaching. 2. Cogni-

tive grammar. 3. Corpora (Linguistics) I. Knop, Sabine De.II. Rycker, Teun De.

P53.C568 2008407dc22

2007052034

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

ISBN 978-3-11-019595-8ISSN 1861-4078

Copyright 2008 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin

All rights reserved, including those of translation into foreign languages. No part of this bookmay 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.

Printed in Germany

Mouton de Gruyter (formerly Mouton, The Hague)is a Division of Walter de Gruyter GmbH & Co. KG, Berlin

Printed on acid-free paperwhich falls within

the guidelines of the ANSIto ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Cognitive approaches to pedagogical grammar / edited by Sabine DeKnop, Teun De Rycker.

p. cm. (Applications of cognitive linguistics ; 9)Includes bibliographical references and index.ISBN 978-3-11-019595-8 (hardcover : alk. paper)1. Language and languages Study and teaching. 2. Cogni-

tive grammar. 3. Corpora (Linguistics) I. Knop, Sabine De.II. Rycker, Teun De.

P53.C568 2008407dc22

2007052034

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

ISBN 978-3-11-019595-8ISSN 1861-4078

Copyright 2008 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin

All rights reserved, including those of translation into foreign languages. No part of this bookmay 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.

Printed in Germany

Acknowledgements

It is only fitting that someone who has made such a substantial and lastingcontribution to the field of linguistics over the past fifty years should becelebrated with a volume in his honour. When Ren Dirven turned sixty in1992, a Festschrift was published to mark the occasion, reflecting in itscontributions from all over the world his wide-ranging academic interests,passions as well as his numerous friendships and connections. His retire-ment from full-time professorship five years later has not slowed, however,his appetite for language research. Over the past decade Ren has beeninvolved in an impressive number of book projects, to which can be addedhis usual output as a researcher, writer and inquisitive thinker. More thananything else perhaps we feel it is his continued support and encourage-ment of others in their own (cognitive) exploration of language and lin-guistics that has been invaluable.

When Ren contacted us in 2006 to start work on this volume, we hadno idea of the journey ahead, neither in terms of the workload or what itwould mean to collaborate so closely with someone of his intellectual cali-ber, with someone who is so au fait not just with the literature but also withnewly emerging trends. We would like to thank Ren sincerely for his ex-tensive and insightful guidance during the production of this volume, and itis with pleasure that we dedicate it to him. It is largely thanks to his criticalbut constructive comments on many of the earlier drafts that the final arti-cles are of such a high quality. Talking of these articles, we really enjoyedworking together with the eighteen authors who so generously contributedtheir research. We would like to express our gratitude to them for theirprofessionalism, commitment and patience. Our thanks also go to theanonymous reviewer whose detailed assessment has helped us edit andfine-tune the manuscript in its final stages.

We would also like to thank Birgitta Meex and Tanja Mortelmans, whowere both involved in getting the project on the rails but who had to leavethe editorial team due to other professional obligations. Finally, we greatlyappreciate the help received from Birgit Sievert, managing editor of theApplications of Cognitive Linguistics series, as well as our other Moutonde Gruyter contacts and especially editor-in-chief, Anke Beck, for her be-lief in our projects potential.

Table of contents

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

By way of introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Sabine De Knop and Teun De Rycker

Part I: Cognition and usage: Defining grammar, rules, models andcorpora

The relevance of Cognitive Grammar for language pedagogy . . . . 7Ronald W. Langacker

Some pedagogical implications of cognitive linguistics . . . . . . . . 37John R. Taylor

Cognitive linguistic theories of grammar and grammar teaching . . 67Cristiano Broccias

Corpora, cognition and pedagogical grammars: An account ofconvergences and divergences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Fanny Meunier

Part II: Tools for conceptual teaching: Contrastive and error analysis

Cross-linguistic analysis, second language teaching and cognitivesemantics: The case of Spanish diminutives and reflexiveconstructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Francisco Jos Ruiz de Mendoza Ibez

Spanish middle syntax: A usage-based proposal for grammarteaching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155Ricardo Maldonado

viii Table of contents

What can language learners tell us about constructions? . . . . . . . . 197Javier Valenzuela Manzanares and Ana Mara Rojo Lpez

Conceptual errors in second-language learning . . . . . . . . . . . . . . . . 231Marcel Danesi

Part III: Conceptual learning: Construal of motion, temporalstructure, and dynamic action

Motion events in Danish and Spanish: A focus on form pedagogicalapproach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259Teresa Cadierno

Motion and location events in German, French and English: Atypological, contrastive and pedagogical approach . . . . . . . . . . . . . 295Sabine De Knop and Ren Dirven

Making progress simpler? Applying cognitive grammar to tense-aspect teaching in the German EFL classroom . . . . . . . . . . . . . . . . 325Susanne Niemeier and Monika Reif

Aspectual concepts across languages: Some considerations forsecond language learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357Barbara Schmiedtov and Monique Flecken

The use of passives and alternatives in English by Chinesespeakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385Liang Chen and John W. Oller, Jr.

Author index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427

List of contributors

Cristiano BrocciasDepartment of Linguistic andCultural CommunicationUniversit di [email protected]

Teresa CadiernoInstitute of Language and Com-municationUniversity of Southern [email protected]

Liang ChenCommunication Sciences andSpecial Education College ofEducationUniversity of GeorgiaAthens, [email protected]

Marcel DanesiUniversity of [email protected]

Sabine De KnopDepartment of Germanic Lan-guages and LinguisticsFacults universitaires Saint-Louis

[email protected]

Teun De RyckerLessius [email protected]

Ren DirvenUniversitt Duisburg-EssenGermany [email protected]

Monique FleckenSeminar fr Deutsch alsFremsprachenphilologie,University of [email protected]

Ronald W. LangackerDepartment of LinguisticsUniversity of CaliforniaSan Diego, [email protected]

Ricardo MaldonadoUniversidad Nacional Autnomade Mxico and Universidad Au-tnoma de QuertaroMexico

x List of contributors

[email protected]

Fanny MeunierCentre for English Corpus Lin-guisticsUniversit catholique de [email protected]

Susanne NiemeierUniversity Koblenz-LandauCampus KoblenzEnglish [email protected]

John W. Oller, Jr.Department of CommunicativeDisordersUniversity of [email protected]

Monika ReifUniversity Koblenz-LandauCampus LandauEnglish [email protected]

Ana Mara Rojo LpezDepartment of English PhilologyUniversity of [email protected]

Francisco Jos Ruiz de MendozaIbezDepartment of Modern Lan-guagesFaculty of Humanities and Edu-cationUniversity of La [email protected]

John R. TaylorDepartment of EnglishUniversity of OtagoNew [email protected]

Barbara SchmiedtovSeminar fr Deutsch alsFremdsprachenphilologieUniversity of [email protected]

Javier Valenzuela ManzanaresDepartment of English PhilologyUniversity of [email protected]

By way of introduction

Sabine De Knop and Teun De Rycker

The notion of pedagogical grammar is ambiguous, being used grammati-cally either as a mass noun or as a count noun. The mass noun standsmetonymically for a process, i.e., research into grammar pedagogy, and asa count noun it refers to the concrete outcome of that research in the formof a pedagogical grammar package for a specific age and proficiency levellearning a given target language. In this volume, the term is mostly used inthe first sense, i.e., it most of all refers to the research process. Of course,if continued systematically, some studies in the present volume may even-tually lead to a partial pedagogical grammar in a certain area of the lan-guage concerned. It may be important to emphasize that a pedagogicalgrammar, as the term package suggests, is different from a didactic gram-mar or school grammar in the sense that it can never be equated with, norbe intended as, a practical manual for teachers, let alone learners. Put in thesimplest way, a pedagogical grammar is both an inventory of all the form-meaning units of the target language, and a didactic approach to their ac-quisition. Therefore a pedagogical grammar will always consist of twointerwoven layers. The first layer is a representation of the main units of alanguage, which according to the theoretical viewpoint taken can be lexi-cal, grammatical or constructional in nature, and possibly also representa-tions of more abstract metacategories such as subject, noun or verb, orrepresentations of abstract schemas and rules. This will form the maintopic of the more theoretical papers in Part I. Interwoven with a survey ofall the form-meaning units is the development of an optimal teaching andlearning infrastructure to facilitate their acquisition. In fact, these two lay-ers of a pedagogical grammar come in most research programs as two ormore separate phases. Indeed, one of the first necessary steps is the exami-nation of the contribution that theoretical, descriptive and contrastive lin-guistics can make towards the inventory of items to be learned and thefacilitation of their learning and teaching in guided learning situations.

Consequently, pedagogical grammar research can be practiced in anumber of different ways, some of which are focused upon in this volume.

2 Sabine De Knop and Teun De Rycker

Providing the repertory of all the essential linguistic units consideredrelevant by theoretical and descriptive linguistics in a given language.The discussion of what these units are is the central theme of Part I.

Pointing out the areas of overlap and contrast between a target lan-guage and the learners native language, and analysing interferencesbetween those languages in the growth of the learners interim gram-mar, i.e., the grammar that he or she develops in attaining some degreeof conceptual fluency in the target language (to borrow the term usedby Danesi in this volume). These research areas are traditionallyknown as contrastive analysis and error analysis, and form the focus ofPart II.

Delineating learning problems by means of certain conceptual areas,either differently represented in the two languages or differentlymatched in the pairing of forms and meanings. Such descriptive andteaching- or learning-oriented analyses can directly or indirectly servethe purpose of preparing teaching materials and exploiting these inL2/FL instruction.

These areas of pedagogical grammar are further structured and explored ineach of the volumes main parts as follows. Part I, Cognition and usage:Defining grammar, rules, models and corpora, tries to answer the questionof what the basic form-meaning units of language are an essential con-dition for setting up a repertory of the target language units. Within Cogni-tive Linguistics there is, in spite of the numerous common assumptionsabout language, a great divergence of opinion on what these basic unitsare. Are they mainly lexical units with grammatical dependencies or argu-ments, as assumed in Hudsons (1991) Word Grammar model? Or do they,on top of lexical and grammatical categories, also include metatheoreticalabstract categories for word classes and sentence constituents, as claimedby Langacker? Do basic units also include abstract schemas or rules, asTaylor suggests? Or do they mainly contain constructions i.e., fixed,even idiomatic sequences at phrase level or sentence level as claimed inGoldbergs (1995) Construction Grammar and in Crofts (2001) RadicalConstruction Grammar, both surveyed in Broccias contribution? Beforeone can even start pedagogical grammar research, one has to come to gripswith these basic questions. In applying the usage-based model that Cogni-tive Linguistics claims to be, it will also be necessary to define what countsas admissible data, that is, the corpora one can use in linguistics and in

By way of introduction 3

language pedagogy. This is explored by Meunier, who compares the vari-ous meeting grounds between corpus linguistics and cognitive linguistics.

Part II, Tools for conceptual teaching: Contrastive and error analysis,examines what a cognitive approach to these two traditional areas of peda-gogical grammar can contribute to the field of conceptual learning. Con-trastive analysis plays an important role in Ruiz de Mendozas andMaldonados contributions. Both look at Spanish as the target language.Thus the former explores the diminutive with its very rich conceptual dif-ferentiation in Spanish vs. its minimal significance in English. The latterexplores se constructions in Spanish mainly as variants of the ubiquitousmiddle construction, in comparison with which the se reflexive is only aminor area for learners of Spanish as a foreign language. Whereas the datain Maldonados analysis are based on error analysis as a data base, thisbecomes the main topic of Valenzuela and Rojos and Danesis papers.Valenzuela and Rojo base their error analysis of the use of the Englishditransitive construction on data from the Spanish component of the Inter-national Learner Corpus of English and show that constructions exist inthe English data of L2 learners, who acquire certain constructional con-figurations in an exemplar-based way, just as Tomasello (2000, 2003) didfor L1 acquisition. Danesi, who replaces the Chomskyan notion of a dis-embodied grammatical competence by the more person- and interaction-oriented notion of conceptual fluency and that of formal-linguistic er-ror by conceptual error, examines the errors and lack of metaphorical orother associations in the use of color terms by English learners of Italian.

Part III, Conceptual learning: Construal of motion events, temporalstructure, and dynamic action offers analyses of learning problems in thesethree areas. As Talmy (1985) has shown, motion events are conceptuallystructured quite differently in Romance and Germanic languages. The con-strual of motion events is explored from the viewpoint of Spanish speakerslearning Danish by Cadierno and of French speakers learning German orEnglish by De Knop and Dirven. The construal of external temporal struc-ture (tense) and internal temporal structure (aspect) by German learners ofEnglish is described by Niemeier and Reif, who also devise a teaching andlearning experiment based on these learning difficulties. Schmiedtov andFlecken concentrate on the fundamental differences between the two Slavicaspects (perfective vs. imperfective) and the progressive aspect in Englishand Dutch. Chen and Oller explore the different ways of construing dy-namic action by the English passive constructions and its many variantsand manage to unveil the very restricted use of them by Chinese learners ofEnglish. All in all, most papers in the volume amply demonstrate the array

4 Sabine De Knop and Teun De Rycker

of potential choices to construe a scene and the extremely restricted usemade of them by foreign learners. Conceptual learning and teaching is thefirst point on the future research agendas of language pedagogy in generaland of pedagogical grammar in particular.

Like any other comprehensive theory of language, CL is faced with theproblem of turning a rich, specialized and emerging body of applied cogni-tive linguistic research into a practical guide for foreign-language teachers,course designers and materials writers. Language pedagogy is of necessityinterdisciplinary in character, crossing over into and closely collaboratingwith, among others, psycholinguistics and educational psychology. WhatCL brings to this crowded and rather confusing field of study more thanany other contemporary form of linguistics is a powerful conceptualunity. The contributions to this volume have shown that this unity in theo-retical assumptions, basic insights and constructs produces highly relevantresearch. We hope that they may also offer rewarding avenues for furtherexploration of what it means to think before you speak in a foreign lan-guage.

References

Croft, William2001 Radical Construction Grammar: Syntactic Theory in Typological

Perspective. Oxford: Oxford University Press.Goldberg, Adele E.

1995 Constructions: A Construction Grammar Approach to ArgumentStructure. Chicago: University of Chicago Press.

Hudson, Richard A.1991 Word Grammar. Oxford: Blackwell.

Talmy, Len1985 Lexicalization patterns: Semantic structure in lexical forms. In Lan-

guage Typology and Syntactic Description, Vol. 3: GrammaticalCategories and the Lexicon, Timothy Shopen (ed.), 36149. Cam-bridge: Cambridge University Press.

Tomasello, Michael2000 First steps toward a usage-based theory of language acquisition.

Cognitive Linguistics 11: 6182.2003 Constructing a Language: A Usage-based Theory of Language Ac-

quisition. Harvard: Harvard University Press.

Part ICognition and usage: Defining grammar, rules,models and corpora

The relevance of Cognitive Grammar for languagepedagogy1

Ronald W. Langacker

Abstract

Cognitive Grammar offers a natural and promising basis for language instruction. Itadvances a conceptual account of linguistic meaning which, by showing how alter-nate expressions construe the same situation in subtly different ways, renders com-prehensible the varied means of expression a language provides. This conceptualsemantics is not confined to lexicon but also extends to grammar: every grammati-cal element or construction imposes a particular construal on the situation beingdescribed. Grammar can thus be presented as an array of meaningful options whoseranges of application are in large measure predictable. Also important is the usage-based nature of Cognitive Grammar. Language structure emerges by abstractionfrom usage events, embracing all dimensions of how expressions are understood byinterlocutors in the social, cultural, and discourse context. Implications for lan-guage learning include the importance of non-descriptive modes of speech, theneed to produce and understand appropriate expressions in a natural context, andthe dependence of fluent speech on mastery of a vast array of conventional expres-sions and phraseology.

Keywords: cognitive grammar; language instruction; centrality of meaning; mean-ingfulness of grammar; units; constructions; composition; construc-tional schema; schematization; conceptual substrate; usage-based per-spective; mental construction; metaphorical construction; fictivemotion; mental space; pedagogical implications

1. Introduction

Few would maintain that language instruction is easy. Nor can the adviceof linguists always be counted on to make it any easier. Unless they arethemselves experienced language teachers, the advice of linguists on lan-guage pedagogy is likely to be of no more practical value than the adviceof theoretical physicists on how to teach pole vaulting. What they can of-fer, qua linguists, is insight into the structure of particular languages and

8 Ronald W. Langacker

the properties of language in general. But even when limited in this fash-ion, the input of linguists cannot necessarily be trusted. They quarrel withone another about the most fundamental issues, suggesting that some ofthem (at least) must be fundamentally wrong. It is therefore unsurprisingthat the impact of linguistic theory on language pedagogy has been lessthan miraculous and sometimes less than helpful.

It remains to be seen whether language teaching will fare any betterwhen guided by notions from cognitive linguistics. There are howevergrounds for being optimistic. Compared to other approaches, cognitivelinguistics offers an account of language structure that just from the lin-guistic standpoint is arguably more comprehensive, revealing, and de-scriptively adequate (certainly I have argued this, e.g., in Langacker1995a). More to the point, the present discussion will focus on three basicfeatures of Cognitive Grammar (Langacker 1987a, 1990, 1991, 1999a) thatsuggest its potential utility as a basis for language instruction: the centralityof meaning, the meaningfulness of grammar, and its usage-based nature.Although extensive pedagogical application remains a long-term goal, Iregard its effectiveness in language teaching to be an important empiricaltest for the framework.

2. The centrality of meaning

If generative linguistics views syntax as being central to language, cogni-tive linguistics accords this honor to meaning. The latter seems far morenatural from the perspective of language users. When ordinary peoplespeak and listen, it is not for the sheer pleasure of manipulating syntacticform their concern is with the meanings expressed. This does not ofcourse imply that grammar is unimportant in language or in languageteaching. It is however helpful to realize that grammar subserves meaningrather than being an end in itself.

The centrality of meaning is reflected in a fundamental claim of Cogni-tive Grammar (henceforth CG), namely that lexicon and grammar form acontinuum consisting solely in assemblies of symbolic structures. A sym-bolic structure is nothing more than the pairing of a semantic structure anda phonological structure. It follows from this claim that grammar itself ismeaningful, just as lexical items are. Grammatical meanings are generallymore abstract than lexical meanings. This is however a matter of degree, sothere is no clear line between lexicon and grammar.

The relevance of Cognitive Grammar for language pedagogy 9

Over the years, progress in describing meanings was greatly impeded bycertain assumptions which, from the standpoint of cognitive semantics,were simply gratuitous. First is the notion that a given element, such as alexical item, has just a single linguistic meaning. A glance at any compre-hensive dictionary makes this seem quite dubious. Although many detailshave yet to be resolved (Allwood 2003; Langacker 2006; Sandra and Rice1995; Zlatev 2003), the basic point that lexical items are frequentlypolysemous having multiple, related senses has been establishedthrough numerous case studies (e.g., Brugman 1981; Lindner 1982; Tuggy2003; Tyler and Evans 2003). The polysemy of lexical items is a specialcase of the general cognitive linguistic claim, also well established, thatlinguistic categories are usually complex: their full description takes theform of a network of related variants (Lakoff 1987; Langacker 1987a; Tay-lor 2004).

A second standard but gratuitous assumption is that a lexical itemsmeaning is circumscribed and distinct from general knowledge. Meta-phorically, it is like a dictionary entry: a short definition in a special for-mat, capturing everything speakers know about the entities denoted just byvirtue of knowing the language. The problem, pace Wierzbicka (1995), isthe absence of any non-arbitrary way to draw the line (Haiman 1980). In-stead of being distinct from general knowledge, lexical meanings recruitand exploit it, representing particular ways of viewing it and making itaccessible for linguistic purposes. Though some specifications are clearlymore central and frequently accessed than others, virtually any aspect ofour knowledge of the entities denoted can be invoked for linguistic pur-poses (Langacker 1987a, 2003).

More basic is the assumption that meaning resides in correspondenceswith the world: the set of entities a word denotes, or the conditions underwhich a sentence is true. In stark contrast to the objectivist tradition, wherethe mind is left out of the loop, cognitive semantics views meaning as amental phenomenon. It resides in conceptualizing activity, whereby weengage the world at many levels: physical, mental, social, cultural, emo-tional, and imaginative. Crucially, we have the capacity to conceive andportray the same situation in alternate ways. While there may be defaults,there is no completely neutral way to describe situations expressionsnecessarily construe them in a certain manner. An expressions meaning istherefore only partly determined by objective properties of the situationdescribed (if, indeed, it has any objective existence at all).

One dimension of construal is specificity, i.e., degree of precision anddetail. It is reflected in lexical hierarchies like thing > creature > animal >


dog > poodle, or do > act > move > run > sprint. Within such a hierarchy,there is usually a basic level (in this case dog or run) at which we tend tooperate lacking any reason to be more specific or more schematic (Rosch1978). For a given lexical choice, greater specificity can always beachieved through modifiers and more elaborate descriptions. One measureof success in language learning is facility in moving away from basic leveldescriptions and going into sub- and superordinate levels of categorization;this has been dubbed conceptual fluency (Danesi 1995, 1999; Kecskesand Cuenca 2005).

A second dimension of construal is prominence, of which there arevarious sorts. Two are especially important: profiling and the focal promi-nence of relational participants. An expressions profile is what it desig-nates, i.e., its referent within the array of conceptual content it evokes asthe basis for its meaning. For instance, roof evokes the conception of abuilding (primarily as seen from the outside), within which it profiles thestructural portion that covers it on top. It contrasts with ceiling, which isseen from the inside and covers part or all of it. It is predictable that manylanguages do not make this distinction, using the same term for both con-cepts. Similarly, it is a language-specific matter that brother-in-law candesignate either a wifes brother or a sisters husband. The word key in-vokes the conception of locks and how they function, profiling an objectcharacterized most essentially by its role in their operation. In the diagramsof Figure 1 the heavy lines indicate profiling; the less heavy lines in dia-gram (a) indicate the base. In diagrams (b)(d) the base an indefinitespatial expanse is not separately indicated.

Figure 1. Schemas for roof, on, intransitive move and transitive move

As Figure 1 shows, expressions profile either things, as in (a), or rela-tionships, as in (b)(d), assuming very general definitions of those terms(Langacker 1987b). In the case of relational expressions, a second kind ofprominence comes into play: the degree of salience conferred on the par-ticipants in the profiled relationship. There is generally a primary focal

trtr lmtr

lm

(a) roof (b) on (c) move (intrans.) (d) move (trans.)


participant, called the trajector (tr), and often a secondary focal partici-pant, called the landmark (lm). The trajector is the entity being described,located, or otherwise characterized. As shown in Figure 1(b), on profiles arelationship where (prototypically) the trajector is in contact with the uppersurface of the landmark, which supports it (double-headed arrow). Theverb move has two basic senses, sketched in diagrams (c) and (d). The in-transitive move profiles a relationship wherein the self-moving trajectorsuccessively occupies a series of locations (single arrow). As a transitiveverb, its trajector is characterized as somehow exerting force (double ar-row), acting on the landmark and thereby causing its motion.

Let us note just one more dimension of construal, namely perspective,which is also multifaceted. Among its facets are vantage point and orienta-tion, implicated in the various interpretations of Jack was sitting to the leftof Jill. For instance, if the speaker was facing both Jack and Jill, the sen-tence may indicate either that Jack was to the speakers left (and Jillsright), or else that Jack was to Jills left (hence the speakers right). An-other facet is the contrast between a local and a global perspective. Forexample, sentence (1a) is the sort of thing one would say while actuallytravelling along the road. What counts as this road is then the portion onecan see at a given moment. Use of the progressive (is winding) indicatesthat the road (so characterized) changes position through time vis--vis themountains (Langacker 1987b). On the other hand, (1b) is the sort of thingone would say when the entire configuration of road and mountain is ap-prehended as a single gestalt (e.g., in looking at a map). Use of the simplepresent tense (winds) indicates that their relationship is stable throughtime.

(1) a. This road is winding through the mountains.[local perspective]

b. This road winds through the mountains.[global perspective]

The pervasive importance of construal shows clearly that linguisticmeaning does not reside in the objective nature of the situation described,but is crucially dependent on how the situation is apprehended. Indeed, thesituation in question is very often a mental construction which has no ob-jective existence in the first place. Much of what we express linguisticallyis imaginative in nature, even in talking about actual occurrences. The sen-tences in (1), for instance, can both be used in reference to an actual con-figuration of road and mountain. They nonetheless involve what is known


as fictive (or virtual) motion: although the road is really stationary, wedescribe it with expressions (like wind and through the mountains) nor-mally used for movement along a path:

(2) A snake was winding through the grass. [actual motion]

In (1a), we construct a sense of motion by imagining distinct road segments(the portions successively in view) as being a single moving entity. In (1b),we do so by mentally scanning along the roads extension (the same scan-ning we do when conceptualizing something actually moving along it).

Fictive motion is varied, extremely common, and psychologically real(Langacker 1986; Matlock 2001, 2004; Matlock and Richardson 2004;Matsumoto 1996a; Talmy 1996). It is a special case of fictive change(Dapremont 2001; Matsumoto 1996b; Sweetser 1997). Some further exam-ples of fictive change are given in (3).

(3) a. The companys president keeps getting younger.b. broken line (cf. broken pencil); scattered villages (cf.

scattered marbles); sunken bathtub (cf. sunken ship)

The president who defies the laws of nature is a fictive entity analogousto the road in (1a). We mentally construct this person by treating differentinstantiations of the role as if they were a single individual. Past participleslike broken, scattered, and sunken normally designate the situation result-ing from a change-of-state process, e.g., a broken pencil is one that hasundergone the process of breaking. The uses cited, however, do not involveany actual change scattered villages have never been clustered together,nor has a sunken bathtub ever actually sunk. The change implied is onlyvirtual, representing the conceived departure of the situation describedfrom the one regarded as canonical.

Everyday language is replete with references to fictive entities, ofteninvoked for describing actual situations, as in (4).

(4) a. He doesnt have a sister.b. If she writes a novel, she will try to publish it.c. A kitten likes to chase its tail.d. Each boy was holding a frog.

If we take the sentences in (4) as truthful statements of what the world isactually like, the nominals in bold nevertheless have referents which are


only virtual they are conjured up as parts of mental constructions thatrelate to actuality in various ways. In sentence (4a), a sister is conjured upprecisely in order to specify what is not the case. The novel referred to in(4b) exists only in the hypothetical situation introduced by if. A genericstatement like (4c) represents a generalization about the worlds essentialstructure. As such, it does not refer to any actual kitten or its tail, but ratherto fictive instances of these types which correspond to open-ended sets ofactual ones. Similarly, (4d) makes a generalization about a delimited set ofactual boys. The entities directly expressed linguistically are only virtual,however: each boy is not any actual boy, nor does a frog designate a spe-cific frog; these virtual entities are invoked to describe a virtual relation-ship taken as corresponding to a contextually determined set of actual ones.Despite their fictivity, these nominals all have referents in the linguisticallyrelevant sense. Observe that the pronoun it refers back to the novel in (4b),and to the kitten in (4c).

Much of what we talk about is constructed metaphorically (Kvecses2005; Lakoff and Johnson 1980; Lakoff and Nez 2000; Turner 1987). In(5), for instance, achieving a goal is metaphorically construed as hitting atarget (aim at), drug dependency as captivity (free from), and the lure ofdrugs as fishing (hooked). Even the preposition on is metaphorical: to beon drugs is to be in contact with them and require them for support.

(5) The therapy is aimed at freeing him from the drugs he ishooked on.

Metaphor is a special case of blending (Fauconnier and Turner 2002),where selected elements of two conceptions are projected and integrated toform a third, which is often purely imaginative but nonetheless real as anobject of thought. One such case is a cartoon character, e.g., a dog thatbehaves like a person and thinks in English. Brunch is a blend of breakfastand lunch (both conceptually and phonologically). Blending in turn is aspecial case of mental space configurations (Fauconnier 1985, 1997; Fau-connier and Sweetser 1996). In both thought and discourse, we divide ourmental world into separate representational spaces, which are connectedto one another in particular ways but nonetheless retain a measure ofautonomy. In (4b), if introduces a hypothetical space distinct from reality.Generic statements like (4c) pertain to a space representing the worldsessential nature (Goldsmith and Woisetschlaeger 1982; Langacker 1999b).In (5), aim at implies a goal and thus a person who entertains that goal. The


goal is a special mental space within the space representing that personsthoughts.

Though generally implicit, these various kinds of mental constructionsare crucial to both the form and meaning of expressions. They are facets ofan elaborate conceptual substrate that supports and makes coherent thenotions overtly expressed. This substrate gives us the freedom to be selec-tive in what we explicitly code linguistically, knowing that the listener canfill in the gaps. It also gives us the freedom to focus on what is salient oreasily expressed, at the expense of full accuracy. This is the basis for me-tonymy, in which some entity is invoked by overtly mentioning another,associated entity which calls it to mind (Kvecses and Radden 1998; Lan-gacker 1993; Panther and Radden 2004). Metonymy is utterly pervasive inlanguage. A few examples are given in (6).

(6) a. Im parked down the street.[I my car]

b. Chicago made the playoffs.[Chicago team from Chicago]

c. She heard a truck.[a truck sound emitted by a truck]

d. The omelet is impatient.[the omelet the customer who ordered the omelet]

e. I phoned my lawyer.[phoned talked to on the phone]

Since language is all about meaning, the findings of cognitive semanticsare clearly relevant to language teaching. Most broadly, they show the bla-tant inadequacy of the conventional CONDUIT metaphor (Reddy 1979) em-ployed in thinking and talking about language itself: meaning is conceivedmetaphorically as a substance (meaning is a mass noun); words are con-tainers (cf. empty words) which hold only limited quantities of this sub-stance; expressions holding ideas are conveyed from the speaker to thehearer (get the idea across); so understanding an expression is just a matterof opening the containers and combining the thoughts they hold. But it issimply not true that an expressions meaning is in its words. Actually,words are merely prompts for an elaborate process of meaning constructionthat draws on the full range of our mental resources. An appreciation of therichness and flexibility of these resources would seem essential for effec-tive language instruction, especially at advanced levels. It is also a worthyeducational goal, even for first language instruction.


More specific pedagogical implications are limited only by the avail-ability of cogent and detailed analyses. For example, Kvecses (2001) hassuggested and to some extent shown empirically that the learning ofidioms is facilitated by apprehension of their metaphorical motivation.Along the same lines, Kurtyka (2001) reports positive results in using cog-nitive semantic descriptions as a basis for teaching phrasal verbs. It standsto reason that the teaching and/or learning of these verb + particle combi-nations would be aided by the realization that the choice of particle, ratherthan being arbitrary, virtually always has a semantic rationale (Lindner1982; Rudzka-Ostyn 2003). It is likewise pertinent to realize that a particu-lar element does not have a single, fixed meaning, but rather an array ofsenses related in principled ways to its prototypical value. Consider on.Prototypically it designates a physical relationship in which the trajector issupported by the landmark and in contact with its upper surface (e.g., thecat on the mat). Other uses are then obtained by suspending one or more ofits specifications: that the surface be an upper one (the painting on thewall); the notion of support (the spots on that cow); or that of contact (theaccent marks on these vowels). Still other uses can be seen as metaphoricalapplications of these same features (e.g., contact and support for on drugs).

The basic point is that conventional usage almost always has conceptualmotivation. Though it has to be learned, it represents a particular way ofconstruing the situation described. With proper instruction, the learning ofa usage is thus a matter of grasping the semantic spin it imposes, a farmore natural and enjoyable process than sheer memorization. The peda-gogical challenge is then to determine the optimal means of leading stu-dents to this understanding (see Boers and Lindstromberg 2006).

3. The meaningfulness of grammar

A conceptualist semantics that properly accommodates construal makespossible a symbolic account of grammar. Like lexicon, with which it formsa gradation, grammar reduces to form-meaning pairings. All the elementscorrectly invoked in grammatical description should thus have semanticimport, however schematic they might be in terms of conceptual content.The meaningfulness of grammar must obviously be recognized in devisingstrategies for teaching it effectively.

From the CG perspective, the first order of business in analyzing gram-mar is to ascertain the meanings of grammatical structures and the ele-ments invoked to describe them. These include both general descriptive


notions (e.g., noun, verb, subject, object, and clause) and the grammaticalformatives (markers or function words) of particular languages. Semanticcharacterizations are possible when it is recognized that such elements areoften polysemous, that their meanings are usually quite abstract, and thatmeaning resides in how we conceptualize the world (not simply in corre-spondences to it).

These points invalidate standard arguments for the doctrine found inevery linguistics textbook that basic grammatical notions like noun, verb,subject, and object cannot be semantically characterized. It is presupposedthat the only possible characterizations would be notions like (physical)thing, event, agent, and patient, pertaining to objective propertiesof the entities involved. The arguments then consist in showing that suchdefinitions are inappropriate for many instances, e.g., a noun like explosion(which refers to an event) or the subject of a passive (which is non-agentive). But this is comparable to looking for your car keys under thestreet light instead of where you dropped them. Any general definitionwould have to be considerably more abstract, and would have to be basedon conceptual factors rather than objective properties. The characteriza-tions proposed in CG are based on factors (e.g., the directing of attention)justified independently as being necessary for semantic description.

Briefly, an expressions basic category depends on the nature of theprofile it imposes on evoked scenarios (not its overall conceptual content).A verb profiles a process, characterized schematically as a relationshipfollowed sequentially in its evolution through time. Lets have a look at thefollowing figure:

Figure 2. Schemas for move and mover

Move is thus a verb, as it profiles the change through time in a spatial rela-tionship. On the other hand, mover is a noun, even though it evokes thesame conceptual content. It is a noun because it profiles a thing, specifi-cally the trajector of the verb it derives from.

(a) move (verb) (b) mover (noun)

tr trlm lm


A thing is defined abstractly as any product of two fundamental concep-tual operations, namely grouping and reification. (Many nouns designategroups: team, stack, herd, set, convoy, etc. Reification is the treatment of agroup as a unitary entity for higher-level cognitive purposes e.g., in as-sessing a herd as large, even when all the animals constituting it are small.In the case of physical objects, the prototype for nouns, these operationsoccur automatically at a low level of processing, leaving us unaware ofconstitutive entities.)

Likewise, the correlation of subject and object with the semantic rolesagent and patient, or subject with discourse topic, is at best only prototypi-cal. Schematic characterizations valid for all instances have to be inde-pendent of specific conceptual content and discourse status. From the CGstandpoint, they find a natural basis in the focal prominence of relationalparticipants: a subject specifies the trajector of a profiled relationship, andan object specifies the landmark. It is further suggested that this primaryand secondary focal prominence resides in the participants being invokedas initial and subsequent reference points in building up to the full concep-tion of the profiled relation (Langacker 1993, 1999c, 2001a, 2001b). Thisdynamic characterization meshes with Chafes description of a subject asthe starting point for apprehending a clause (Chafe 1994).2

The conceptual definability of basic grammatical notions makes evidentthe meaningfulness of grammatical formatives. Among these are deriva-tional elements, such as -er, which combines with a verb to derive a noun.Since these categories are meaningful, so are elements effecting a changeof category membership. What -er contributes semantically to a form likemover is an aspect of construal: the specification that it profiles the verbstems trajector a thing rather than the overall process it designates.Similarly, any elements affecting the choice of subject or object are mean-ingful by virtue of conferring focal prominence on particular relationalparticipants. For instance, a passive marker confers trajector status on whatwould otherwise be a clausal landmark.

Conceptual characterizations have been offered in CG for numerous pa-rade examples of purely grammatical markers, e.g., do, of, and Spanishse. The auxiliary verb do is semantically equivalent to the schematic de-scription of the verb class: it profiles a process, i.e., a relationship trackedthrough time. Since it does not specify any particular kind of process, it isused as a pro form for clauses, referring back to a previously mentionedprocess (just as a pronoun refers back to a nominal referent): She does;They did. The preposition of indicates that the relationship between itstrajector and landmark is somehow intrinsic rather than contingent (Lan-


gacker 1992). Thus it is used for part-whole relations (the sole of my foot),for identity (the month of January), for intrinsic functions (the father of thebride as against the brides dress), to specify the participants in a reifiedprocess (the shooting of the hunters), etc. Note that we say the color of thelawn but the brown spot in (*of) my lawn, the difference being that the spotis not supposed to be there. Another case is the (non-reflexive) se that oc-curs with many Spanish verbs: sentarse sit, caerse fall, lavarse wash,enojarse get mad, ahogarse drown, etc. Though polysemous, it hasbeen shown by Maldonado (1988, 1999) to be consistently meaningful. Itsvarious senses are natural extensions from a prototype that is intermediatebetween a prototypical transitive and a prototypical intransitive (hence theterm middle voice). While the trajector has patient-like properties, thereis also a notion of force or agentivity, without however invoking an agentdistinct from the trajector.

Grammar consists primarily in patterns for combining simpler expres-sions into more complex ones. Complex expressions are called construc-tions, and the patterns they instantiate are constructional schemas. In CG, aconstruction is simply an assembly of symbolic structures linked by corre-spondences. Consider expressions where a prepositional phrase modifies anoun: the table by the window, a cabin in the woods, the roof on thathouse, etc. The schema for such expressions is sketched in Figure 3.

Figure 3. Schema for the roof on that house

X Y

X Y

Y

N P+N

P+NN

P N

tr lm

lmtr


Two levels of organization are shown. At the lower level, a preposition (P)combines with a nominal expression (N) to form a prepositional phrase(P+N); P and N are component structures at this level, while P+N is thecomposite structure resulting from their combination. At the higher level oforganization, the component structures N and P+N combine to yield thecomposite structure N P+N. The two nominal expressions profile things(represented as circles), with X and Y abbreviating their additional seman-tic content. The preposition designates a relationship (shown as an arrow)between two things characterized only schematically.

Correspondences, given as dotted lines, indicate how component struc-tures at each level are conceptually integrated to form the composite struc-ture. At the lower level, the profile of the nominal (that house) correspondsto the prepositions landmark. The composite structure (on that house)results from merging the specifications of the corresponding elementswhile preserving the profile of the preposition (on). Because it inherits itsprofile from the preposition (which is thus the head at this level), the com-posite expression is a prepositional phrase (rather than a nominal). Thisprepositional phrase then modifies the other nominal element (roof) at thehigher level of organization. Here the nominal profile corresponds to theschematic trajector of the prepositional relation. At this level it is thenominal that imposes its profile (making it the head), so the overall expres-sion is nominal rather than relational. Taken as a whole, for example, theroof on that house profiles the roof.

The various aspects of construal needed for describing lexical meaningsalso figure in the description of complex expressions. In particular, profil-ing and trajector/landmark organization are pivotal in grammatical con-structions. On that house is a prepositional phrase because it profiles arelationship rather than a thing, and that house is the prepositional objectbecause its profile corresponds to ons landmark. Similarly, the overallexpression is nominal rather than prepositional because it profiles a thing,and the head noun roof is the subject (as broadly defined in CG) with re-spect to the prepositional phrase because its profile corresponds to the lat-ters trajector. The profiling and correspondences in Figure 3 are specifiedby the constructional schema describing this grammatical pattern, hencethey are characteristic of all the expressions which instantiate it. They con-stitute the constructional meaning of the pattern, which particular instan-tiations elaborate based on the lexical elements chosen.

Constructional schemas can also be thought of as patterns of semanticcomposition, since they specify how the meanings of constitutive elementscombine to form the meaning of the whole. As viewed in CG, however,


semantic structure is only partially compositional, since we draw on manyresources in arriving at the full composite meaning it is not true that anexpressions meaning is in its words. This is so even for seeminglystraightforward examples like the roof on that house. On the most likelyinterpretation, it is not precisely parallel to the vase on that table, since thevase and the table are distinct objects (hence we cannot say *the vase ofthat table), whereas the roof and the house are not distinct entities. Basedon our knowledge of typical scenarios in everyday life, we presume thatthe roof is an inherent part of the house, not that the roof from some otherbuilding is resting on the houses own roof. (Given a proper context, e.g.,in describing the aftermath of a hurricane, this default assumption may ofcourse be overridden.) Just because the trajector and landmark are non-distinct, this usage of on is non-prototypical. We would normally say theroof of the house, recognizing the intrinsic nature of their part-whole rela-tionship. By using on instead, we highlight the role of the house in support-ing the roof, as well as suggesting that their relationship is non-intrinsic(i.e., the roof in question is not the only one the house might have had itis considered in relation to other possible sorts of roof).

On the CG account, therefore, grammar consists in patterns for assem-bling not just complex expressions, but also complex meanings. It incorpo-rates particular ways of construing conceptual content and symbolizing theconstrual imposed on it. Besides the sorts of prominence already indicated,the construal embodied in grammar includes the related factors of perspec-tive and sequentiality in the evocation of conceived entities. A local vs. aglobal perspective was exemplified in (1) [is winding vs. winds]. Anotherdifference in perspective is whether a situation is viewed in terms of inter-acting participants or in terms of how the scene is sequentially accessed.Illustrating the latter are numerous constructions that first introduce a set-ting or location and then indicate what is found there. In some of theseconstructions the setting or location is focused as trajector, making it thegrammatical subject:

(7) a. The streets were lined with spectators.[cf. Spectators lined the streets.]

b. This year has seen some big events.[cf. Some big events have occurred this year.]

c. It appears that shes quite smart.[cf. She appears to be quite smart.]

d. There are problems with that theory.[cf. That theory has problems.]


The setting in (7b) is temporal rather than spatial. The it in (7c) is an ab-stract setting, something like the scope of awareness for the followingjudgment (Langacker 2004). In (7d), there is an abstract realm of existence.

If not the grammatical subject, a setting or location can nonetheless beinitial in terms of word order:

(8) On the mat was a lazy cat. [cf. A lazy cat was on the mat.]

This represents the discourse strategy of first directing attention to a knownlocation as a way of introducing a new participant found there. It exempli-fies sequentiality as a dimension of construal. Another, more striking ex-ample is the contrast in (9):

(9) a. The cookies are in the pantry, on the bottom shelf, in aplastic container.

b. The cookies are in a plastic container, on the bottomshelf, in the pantry.

Both sentences use the same locative phrases to describe exactly the samespatial configuration. They are nevertheless semantically distinct by virtueof imposing different ways of mentally accessing it: the conceptual experi-ence is that of zooming in vs. zooming out. This is just one of manygrammatical phenomena whose essential import resides in sequence ofmental access (Langacker 1993, 2001b).

These brief examples merely hint at the conceptual richness embodiedand reflected in grammatical structure. The fact that grammar is meaning-ful, not an autonomous formal system, creates the potential for new anddifferent approaches in teaching and learning it. Being conventionally de-termined, differently in each language, the proper form of expression doesof course require instruction. Learning grammar does not however have tobe the soulless internalization of arbitrary restrictions. If properly analyzed,every grammatical element makes a semantic contribution and everygrammatical distinction has conceptual import. Awareness of these factorsoffers a basis for effective language instruction aimed at their full exploita-tion in thought and communication. How can this be achieved? Not beingqualified to make specific pedagogical recommendations, I will limit my-self to some general considerations.

A point not yet mentioned is the extensive iconicity of grammar (Givn1991; Haiman 1983, 1985). The symbolic nature of grammar does not im-ply that the form-meaning pairings are simply arbitrary. On the contrary,


the symbolization of grammatical notions invariably has some kind or de-gree of conceptual motivation. Usually, for example, the more tenuousmeanings of grammatical elements are paralleled by their more tenuousphonological realizations: the latter tend to be shorter than lexical items,and are often affixes or inflections rather than independent words. A well-known iconic principle is that elements which belong together semanticallygenerally occur together phonologically. To state this more precisely in CGterms, the conceptual integration effected by grammatical constructionstends to be symbolized by phonological adjacency. With respect to Figure3, the semantic relationship between a preposition and its object is symbol-ized by placing the object nominal directly after the preposition. Likewise,a prepositional phrase is placed directly after the nominal expression itmodifies to symbolize that relationship. A further dimension of iconicity isthe tendency for the sequencing of words to mirror the sequence of eventsor some other conceptual ordering (Langacker 2001b). In (8), word orderimplements the discourse strategy of using a known location to anchor theintroduction of a new participant. The contrasting orders in (9) representalternate natural paths for arriving at the full conception of a complex spa-tial configuration. However it might be exploited pedagogically, the iconic-ity of grammar is an important factor in its learnability (Langacker 1987a:Section 9.3.2).

Cognitive linguistic notions are also potentially helpful in showing con-nections among the varied uses of a grammatical element or construction.Why should possessives, for instance, be used in expressions like Lincolnsassassination (since Lincoln was hardly the owner of this event, evenmetaphorically)? The answer lies in the conceptual characterization of sand the possessive construction (Langacker 1995b, 2001a; Taylor 1996).Their wide range of conventional usage, exemplified in (10), covers farmore than ownership, part-whole, and kinship relationships, which all havesome claim to prototypicality:

(10) Sams house; my neck; the girls mother; our town; the catsfleas; your bus; the stores location; their anxiety; Zeldasproblem; his height; the years biggest event; my driving;anyones guess; Booths assassination [of Lincoln]; Lincolnsassassination [by Booth]

The prototypical senses are special cases of a schematic characterizationbased on a mental operation rather than any specific conceptual content.Possessives manifest our capacity for invoking one conceived entity as a


reference point in establishing mental contact with another, i.e., for men-tally accessing a target via and in relation to the reference point. Out of allthe houses in the world, Sams house directs attention to a particular in-stance of that type, namely the one mentally accessible via Sam. Thiswould typically be the house he lives in, but given the proper context itmight also be the house he owns and rents out, the one he dreams of own-ing someday, the one he designed, the one he is scheduled to paint nextweek, etc. Recall now that the focal prominence of trajector and landmarkwas ascribed to their functioning as conceptual reference points for pur-poses of building up to the full conception of a profiled relationship. Onthis account both the agent and the patient of assassinate are referencepoints for purposes of conceptualizing (mentally accessing) the profiledevent. For this reason either one can be coded as possessor of the derivednoun assassination, obtained from assassinate by conceptual reification:Booths assassination; Lincolns assassination.

More generally, it is helpful to realize that grammatical elements of anysort are likely to be polysemous, having a prototypical meaning as well asan array of other, less central values possibly susceptible to schematiccharacterization. It may not be true that the possessive construction carriesthe meaning of ownership, but neither is it appropriate to ignore the cen-trality of this particular kind of reference point relationship (nor that ofpart-whole and kinship relations). Indeed, full mastery of the constructionimplies awareness of a wide range of usage possibilities like those in (10),since there is no assurance that every type of usage that conforms to theschematic meaning is conventionally exploited in the language. Similarly,while it is important to realize that nouns are not simply the names forphysical objects, and that subjects are agents only prototypically, it is alsouseful to know that these notions are indeed central to the categories, aconceptual core which is grounded in physical experience and provides abasis for extension to more abstract or less typical values.

Finally, grammar is less mysterious when it is recognized that the formof expressions is shaped by an elaborate, multifaceted conceptual substrate.Grammatical peculiarities are commonly the visible trace of tacit, oftenimaginative mental constructions which are readily grasped by nativespeakers even when they fail to cross the threshold of explicit awareness.A simple example concerns the prepositional phrase in (11):

(11) Its pretty through that valley.


Normally a through-phrase modifies a motion verb and specifies the pathtraversed (as in (2)). But (11) has no such verb indeed, its pretty de-scribes a static situation. So what does through the valley modify? Insteadof modifying an overt element, it itself evokes the conception of someonetravelling through the valley and observing its appearance. Moreover, thistravel scenario may be entirely fictive. It cannot be identified with ongoingmovement by the current speaker (as it might if that were changed to this).Nor is it necessarily based on memory or knowledge of any specific jour-ney. It may be just an imagined journey, on the part of an imagined repre-sentative viewer, conjured up as a way of describing the valley.

The general point is that an expressions semantic and grammatical co-herence is often dependent on imagined scenarios or other tacit concep-tions. As one more example, this is the key to understanding non-presentuses of the English present tense (Langacker 2001c), as in (12):

(12) a. I come home last night and a stranger opens the door.b. The children leave for camp next week.c. A kitten is born with blue eyes.

Inter alia, the so-called present tense is used for past events (the historicalpresent), anticipated future occurrences, and timeless statements (suchas generics). The label is not really a misnomer, however. These uses arebased on particular mental constructions in the context of which an in-stance of the profiled process coincides with the time of speaking. Thecoincident process is not however actual but rather virtual a fictive oc-currence that serves as a representation of one or more actual ones. In(12a), the profiled events belong to a mental replay of a previous epi-sode. Sentence (12b) does not refer directly to the actual event of leaving(as would be the case with the future will); it amounts instead to readingoff an entry on a conceptual plan or schedule, where each entry representsa planned future occurrence. Invoked in (12c) is something like a blue-print, a supposed representation of the worlds essential nature. In thisgeneric expression, the kitten and the process of its being born are virtualentities corresponding to an open-ended set of actual instances. These sen-tences are in the present tense because the profiled virtual occurrences arepart of a mental construction the replay, the schedule, or the blueprint immediately available at the time of speaking.


4. A usage-based approach

A final aspect of CG that makes it relevant for language teaching is itsusage-based nature (Barlow and Kemmer 2000; Langacker 2000). A lan-guage comprises an immense inventory of conventionally established ele-ments (lexical items, formatives, grammatical constructions, sound pat-terns, etc.) which fluent speakers learn as units, i.e., they are thoroughlymastered and can thus be employed in largely automatic fashion. Conven-tional units are abstracted from usage events: actual instances of languageuse, in their full phonetic detail and contextual understanding. A crucialfactor in this process is the reinforcing of features that recur across a suffi-cient number of such events. Since every usage event is unique at the levelof fine-grained detail, the recurring commonalities are apparent only at acertain level of abstraction. To some extent, therefore, all linguistic units even the most specific are schematic relative to the usage events in whichthey figure. Fine-grained differences fail to be reinforced and are thereforefiltered out as units emerge.

The emphasis is thus on actual learning. Whatever might be its innatebasis, mastering a language requires the specific, usage-based learning of avast array of conventional units. This itself has pedagogical implications(which may seem obvious, but are not so in every linguistic theory). It sug-gests the importance of providing the learner with sufficient exposure torepresentative uses of a given unit. Ideally, moreover, this exposure shouldoccur in the context of meaningful exchanges approximating socially andculturally normal usage events. In this respect the usage-based approachresonates with the natural approach to language teaching (Achard 2004).

From a usage-based perspective, a basic question is the degree of ab-straction relative to usage events. Given that all units derive from suchevents by schematization, it remains to determine just how schematic theymight be. There is no real doubt that linguistic units run the gamut fromhighly specific to highly schematic. We have noted this in regard to themeanings of lexical items (e.g., do > act > move > go > run > sprint). InCG, grammatical structures are schematic relative to instantiating expres-sions; the constructional schema in Figure 3, for example, is a schematizedrepresentation of particular expressions like a cabin in the woods, the roofon that house, etc. Phonologists find good reason to characterize sounds atdifferent levels of abstraction, for instance [i] < [HIGH VOWEL] < [VOWEL]< [SEGMENT].

In the case of grammar, degree of abstraction figures in several conten-tious theoretical issues. The first is whether as claimed in CG gram-


matical structures are in fact simply schematized representations of com-plex expressions. On this account they consist in assemblies of symbolicstructures fully reducible to form-meaning pairings. With certain qualifica-tions (Langacker 2005a, 2005b), CG shares this outlook with ConstructionGrammar (Fillmore, Kay, and OConnor 1988; Goldberg 1995). It is nothowever standard linguistic doctrine. A second issue is the relative impor-tance of higher- versus lower-level schemas, and a third is whether welearn both schemas and specific instantiations of them. These latter issuespertain to the abstractness and hence the generality of linguistic knowl-edge. Theorists have a natural tendency to look for the broadest generaliza-tions, corresponding to highly schematic descriptions. Broad generaliza-tions should of course be sought. Still, it cannot be taken for granted thatspeakers have the same proclivity. One aspect of the usage-based approachis the notion that speakers rely extensively on low-level schemas and spe-cific learned expressions, even when these conform to general patterns.

The learning of specific forms is obviously necessary in cases of irregu-larity or limited productivity. At the extreme we find the students night-mare of complex morphological paradigms that simply have to be memo-rized, e.g., the conjugations of irregular verbs. For this CG and the usage-based approach have no magic cure. They do however suggest the reason-ableness of what usually occurs in practice. What typically does not occuris that a student thoroughly learns all the forms of a complex paradigm e.g., all the person, number, tense, and mood inflections of an irregularverb and instantaneously retrieves them as needed in actual use. Instead,students tend to do what children presumably do in learning a languagenatively: the forms they learn first and learn best are those which occurmost frequently. Aiding the learning process are many low-level generali-zations (e.g., several verbs forming their past tense in analogous fashion).Eventually an awareness develops of all the dimensions represented inparadigms, patterns (morphological schemas) emerge for constructing anydesired regular form, and a large number of irregular forms are learnedwith different degrees of thoroughness. This is not to say that paradigmsshould never be studied as such. But that should not be thought of as theprimary means of learning what is needed for fluent speech.

The opposite extreme, that of full productivity and regularity, is far lessprevalent than theoretical attitudes would lead one to expect. Much of ourknowledge of grammar resides in intermediate cases: patterns usable withmultiple lexical items, perhaps even an open-ended set, but not with everymember of a basic category. A well-known construction of this sort is theEnglish ditransitive, where a verb has two object-like complements, e.g.,


She gave her brother a watch. Naturally, the verbs occurring in this patternare limited to those involving three central participants. A further restric-tion is that the verbal action has a causal relationship in regard to the situa-tion of the first object possessing or having access to the second (e.g., thegiving results in her brother having a watch). Observe that this causal rela-tionship may be negative rather than positive (She denied them permissionto interview me), and instead of actually bringing about the situation it mayrepresent only a commitment to do so (She promised her brother a watch).Quite a number of verbs satisfy this schematic characterization. Yet notevery verb which does is conventionally used in ditransitives. For instance,while get appears in this construction (She got her brother a watch), obtaindoes not (*She obtained her brother a watch). Also excluded are provide,supply, donate, contribute, deprive, etc.

Hence the distributional facts do not lend themselves to a single, fullypredictive generalization. But neither is the picture one of randomness orcomplete idiosyncrasy. If the construction cannot be adequately describedby means of a global generalization, promising absolute predictability, itnonetheless exhibits considerable regularity in the form of local generaliza-tions, which offer degrees of motivation for the observed distribution(Goldberg 1992, 1995). In other words, the ditransitive construction repre-sents a complex category whose full description consists in a network ofrelated variants centered on a prototype. The prototypical pattern is for theverb to profile an act of transfer from agent to recipient (e.g., give, send,hand, mail, throw, bring). In another basic pattern, the agent creates anobject with the intent of the recipient having access to it (cook, bake, knit,build, make, write). Somewhat more peripheral are verbs of commitment,either positive (promise, owe, permit, allow, guarantee) or negative (re-fuse, deny). Alternate groupings may of course be proposed, as well asvarious subgroups. The point, however, is that the distribution is anchoredby particular, fairly frequent verbs well-established in the construction.These verbs form clusters on the basis of semantic similarity, and certainclusters give rise to constructional subschemas capable of being used pro-ductively.

This organization provides the basis for devising pedagogical strategy.The obvious suggestion is to start with the prototype and then move on toother major clusters, in each case focusing initially on the most frequentand basic verbs. Full mastery of the construction, with native-like knowl-edge of the conventional range of usage, will come about only graduallythrough long-term practice with the language. But the same is true of itslearning by native speakers.


The ditransitive is not atypical of the pedagogical challenge posed bygrammar overall. Fully general and exceptionless rules are themselves theexception. Instead of being monolithic, most constructions exist in multiplevariants, with schemas abstracted to represent what is common to differentsets of them. At the lowest level are constructional variants that incorpo-rate a particular lexical item well-established in the pattern. Such units thusinclude both specific and schematic elements, providing one argumentagainst any sharp distinction between lexicon and grammar. Thoroughknowledge of a construction resides in the entire network of variants(rather than any single unit), and lower-level structures are often the mostimportant in determining conventional usage.

This leads to a final but possibly crucial pedagogical implication of theusage-based approach. A substantial proportion of what is needed to speaka language fluently tends to be ignored because it is part of neither lexiconnor grammar as these are traditionally conceived. What I have in mind arethe countless units representing normal ways of saying things. Nativespeakers control an immense inventory of conventional expressions andpatterns of expression enabling them to handle a continuous flow of rapidspeech (Langacker 1987a: 3536, 2001d). While they can certainly be in-cluded, I am not referring to lexical items of the sort found in dictionaries,nor even to recognized idioms. At issue instead are particular ways ofphrasing certain notions out of all the ways they could in principle be ex-pressed in accordance with the lexicon and grammar of the language.These units can be of any size, ranging from standard collocations to largechunks of boiler-plate language. They can be fully specific or partiallyschematic, allowing options in certain positions.

In fluent speech, piecing together these prefabricated chunks is at leastas important as productively invoking lexical and grammatical units. Na-tive speakers cannot avoid using them here are some examples from theprevious paragraph: leads to; a substantial proportion; speak a languagefluently; tends to be ignored; as traditionally conceived; what I have inmind; continuous flow; rapid speech; I am not referring to; nor even; atissue; out of all the ; in principle; in accordance with; of any size; rang-ing from to ; large chunks of ; boiler-plate language. Given theprevalence of conventional expressions, as well as their critical role influency and idiomaticity, finding effective ways to facilitate their learningwould seem essential.


5. Conclusion

Of necessity (given its source), this discussion has been long on theory andshort on practical recommendations. Nor have the sample descriptionsbeen anything more than fragmentary. If we nevertheless suppose that CGconcepts and descriptions are relevant for language instruction, a basicquestion remains to be addressed: Just who are they relevant to? The mainpossibilities are the student, the instructor, and those responsible for de-signing curricula or developing teaching materials. Assuming their validity,explicit awareness of CG notions would seem most clearly helpful for thelatter. With effective materials and a sensible curriculum, an instructorwith lesser awareness of CG insights can nonetheless still exploit them.And students would hopefully benefit even if they are never exposed totheoretical concepts or explicit analyses. I cannot help thinking, however,that the cognitive linguistic view of language is a matter of universal inter-est, and that its conceptual descriptions of linguistic phenomena are suffi-ciently natural and revealing to be widely appreciated. In some form, I canimagine these ideas being an integral part of general education or first lan-guage instruction. I can further imagine them as being useful in secondlanguage learning, especially at more advanced levels.

Notes

1. This is a slightly changed version of the paper called Cognitive Grammar as abasis for language instruction published in 2007 by Lawrence Erlbaum Asso-ciates in Handbook of Cognitive Linguistics and Second Language Acquisi-tion, edited by Nick Ellis and Peter Robinson. Used with permission by Law-rence Erlbaum Associates.

2. This characterization of the notions subject and object is claimed to bevalid for all languages. This universalist definition of Cognitive Grammar iscontrasted to a language-specific definition in Crofts Radical ConstructionGrammar by Broccias in this volume.

References

Achard, Michel2004 Grammatical instruction in the natural approach: A cognitive gram-

mar view. In Cognitive Linguistics, Second Language Acquisition,


and Foreign Language Teaching, Michel Achard and Susanne Nie-meier (eds.), 165194. (Studies on Language Acquisition 18) Ber-lin/New York: Mouton de Gruyter.

Allwood, Jens2003 Meaning potentials and context: Some consequences for the analysis

of variation in meaning. In Cognitive Approaches to Lexical Seman-tics, Hubert Cuyckens, Ren Dirven, and John R. Taylor (eds.), 2965. (Cognitive Linguistics Research 23) Berlin/New York: Moutonde Gruyter.

Barlow, Michael, and Suzanne Kemmer (eds.)2000 Usage-based Models of Language. Stanford, California: Center for

the Study of Language and Information Publications.Boers, Frank, and Seth Lindstromberg

2006 Cognitive Linguistic applications in second or foreign language in-struction: Rationale, proposals and evaluation. In Cognitive Linguis-tics: Current Applications and Future Perspectives, Gitte Kristian-sen, Michel Achard, Ren Dirven, and Francisco Ruiz de MendozaIbez (eds.), 305358. (Applications of Cognitive Linguistics 1)Berlin/New York: Mouton de Gruyter.

Brugman, Claudia1981 The story of over. MA thesis, Department of Linguistics, University

of California, Berkeley.Chafe, Wallace

1994 Discourse, Consciousness, and Time: The Flow and Displacement ofConscious Experience in Speaking and Writing. Chicago/London:University of Chicago Press.

Danesi, Marcel1995 Learning and teaching languages: The role of conceptual fluency.

International Journal of Applied Linguistics 5 (1): 320.1999 Expanding conceptual fluency theory for second-language teaching.

Mosaic 6 (4): 1621.Dapremont, Elena M.

2001 Assembled data, mounting evidence: Conceptual contrasts betweenpast and present participle based change. Paper presented at the 7thInternational Cognitive Linguistics Conference, Santa Barbara.

Ellis, Nick, and Peter Robinson (eds.)2007 Handbook of Cognitive Linguistics and Second Language Acquisi-

tion. Mahwah, New Jersey: Laurence Erlbaum.Fauconnier, Gilles

1985 Mental Spaces: Aspects of Meaning Construction in Natural Lan-guage. Cambridge, MA/London: Massachusetts Institute of Technol-ogy Press.


1997 Mappings in Thought and Language. Cambridge: Cambridge Uni-versity Press.

Fauconnier, Gilles, and Eve Sweetser (eds.)1996 Spaces, Worlds, and Grammar. Chicago/London: University of Chi-

cago Press.Fauconnier, Gilles, and Mark Turner

2002 The Way We Think: Conceptual Blending and The Minds HiddenComplexities. New York: Basic Books.

Fillmore, Charles J., Paul Kay, and Mary C. OConnor1988 Regularity and idiomaticity in grammatical constructions: The case

of let alone. Language 64 (3): 501538.Givn, Talmy

1991 Isomorphism in the grammatical code: Cognitive and biologicalconsiderations. Studies in Language 15: 85114.

Goldberg, Adele E.1992 The inherent semantics of argument structure: The case of the Eng-

lish ditransitive construction. Cognitive Linguistics 3: 3774.Goldberg, Adele E.

1995 Constructions: A Construction Grammar Approach to ArgumentStructure. Chicago/London: University of Chicago Press.

Goldsmith, John, and Erich Woisetschlaeger1982 The logic of the English progressive. Linguistic Inquiry 13: 7989.

Haiman, John1980 Dictionaries and encyclopedias. Lingua 50: 329357.1983 Iconic and economic motivation. Language 59: 781819.

Haiman, John (ed.)1985 Iconicity in Syntax. (Typological Studies in Language 6) Amster-

dam/Philadelphia: John Benjamins.Kecskes, Istvan, and Isabel M. Cuenca

2005 Lexical choice as a reflection of conceptual fluency. InternationalJournal of Bilingualism 9 (1): 4967.

Kvecses, Zoltn2001 A cognitive linguistic view of learning idioms in an FLT context. In

Applied Cognitive Linguistics II: Language Pedagogy, Martin Ptz,Susanne Niemeier, and Ren Dirven (eds.), 87115. (Cognitive Lin-guistics Research 19.2) Berlin/New York: Mouton de Gruyter.

2005 Metaphor in Culture: Universality and Variation. Cambridge: Cam-bridge University Press.

Kvecses, Zoltn, and Gnter Radden1998 Metonymy: Developing a cognitive linguistic view. Cognitive Lin-

guistics 9: 3777.


Kurtyka, Andrzej2001 Teaching English phrasal verbs: A cognitive approach. In Applied

Cognitive Linguistics II: Language Pedagogy, Martin Ptz, SusanneNiemeier, and Ren Dirven (eds.), 2954. (Cognitive Linguistics Re-search 19.2) Berlin/New York: Mouton de Gruyter.

Lakoff, George1987 Women, Fire, and Dangerous things: What Categories Reveal about

the Mind. Chicago/London: University of Chicago Press.Lakoff, George, and Mark Johnson

1980 Metaphors We Live By. Chicago/London: University of ChicagoPress.

Lakoff, George, and Rafael E. Nez2000 Where Mathematics Comes From: How the Embodied Mind Brings

Mathematics into Being. New York: Basic Books.Langacker, Ronald W.

1986 Abstract motion. Proceedings of the Annual Meeting of the BerkeleyLinguistics Society 12: 455471.

1987a Foundations of Cognitive Grammar. Vol. 1: Theoretical Prerequi-sites. Stanford, California: Stanford University Press.

1987b Nouns and verbs. Language 63: 5394.1990 Concept, Image, and Symbol: The Cognitive Basis of Grammar.

(Cognitive Linguistics Research 1) Berlin/New York: Mouton deGruyter.

1991 Foundations of Cognitive Grammar. Vol. 2, Descriptive Application.Stanford, California: Stanford University Press.

1992 The symbolic nature of cognitive grammar: The meaning of of and ofof-periphrasis. In Thirty Years of Linguistic Evolution: Studies inHonour of Ren Dirven on the Occasion of His Sixtieth Birthday,Martin Ptz (ed.), 483502. Philadelphia/Amsterdam: John Benja-mins.

1993 Reference-point constructions. Cognitive Linguistics 4: 138.1995a Raising and transparency. Language 71: 162.1995b Possession and possessive constructions. In Language and the Cog-

nitive Construal of the World, John R. Taylor and Robert E.MacLaury (eds.), 5179. (Trends in Linguistics Studies and Mono-graphs 82) Berlin/New York: Mouton de Gruyter.

1999a Grammar and Conceptualization. (Cognitive Linguistics Research14) Berlin/New York: Mouton de Gruyter.

1999b Virtual reality. Studies in the Linguistic Sciences 29 (2): 77103.1999c Assessing the cognitive linguistic enterprise. In Cognitive Linguis-

tics: Foundations, Scope, and Methodology, Theo Janssen andGisela Redeker (eds.), 1359. (Cognitive Linguistics Research 15)Berlin/New York: Mouton de Gruyter.


2000 A dynamic usage-based model. In Usage-based Models of Language,Michael Barlow, and Suzanne Kemmer (eds.), 163. Stanford, Cali-fornia: Center for the Study of Language and Information Publica-tions.

2001a Topic, subject, and possessor. In A Cognitive Approach to The Verb:Morphological and Constructional Perspectives, Hanne G. Simon-sen and Rolf T. Endresen (eds.), 1148. (Cognitive Linguistics Re-search 16) Berlin/New York: Mouton de Gruyter.

2001b Dynamicity in grammar. Axiomathes 12: 733.2001c The English present tense. English Language and Linguistics 5:

251271.2001d Cognitive linguistics, language pedagogy, and the English present

tense. In Applied Cognitive Linguistics I: Theory and Language Ac-quisition, Martin Ptz, Susanne Niemeier, and Ren Dirven (eds.), 339. (Cognitive Linguistics Research 19.1) Berlin/New York: Moutonde Gruyter.

2003 Context, cognition, and semantics: A unified dynamic approach. InJob 28: Cognition in Context, Ellen van Wolde (ed.), 179230. (Bib-lical Interpretation Series 64) Leiden/Boston: Brill.

2004 Aspects of the grammar of finite clauses. In Language, Culture andMind, Michel Achard, and Suzanne Kemmer (eds.), 535577. Stan-ford, California: Center for the Study of Language and InformationPublications.

2005a Construction grammars: Cognitive, radical, and less so. In CognitiveLinguistics: Internal Dynamics and Interdisciplinary Interaction,Francisco J. Ruiz de Mendoza Ibez, and Sandra Pea Cervel(eds.), 101159. (Cognitive Linguistics Research 32) Berlin/NewYork: Mouton de Gruyter.

2005b Integration, grammaticization, and constructional meaning. InGrammatical Constructions: Back to the Roots, Mirjam Fried andHans C. Boas (eds.), 157189. (Constructional Approaches to Lan-guage 4) Amsterdam/Philadelphia: John Benjamins.

2006 On the continuous debate about discreteness. Cognitive Linguistics17: 107151.

Lindner, Susan1982 What goes up doesnt necessarily come down: The ins and outs of

opposites. Papers from the Regional Meeting of the Chicago Lin-guistic Society 18: 305323.

Maldonado, Ricardo1988 Energetic reflexives in Spanish. Proceedings of the Annual Meeting

of the Berkeley Linguistics Society 14: 153165.1999 A media voz: Problemas conceptuales del cltico se. (Publicaciones

del Centro de Lingstica Hispnica 46.) Mexico City: Universidad


Nacional Autnoma de Mxico, Instituto de InvestigacionesFilolgicas.

Matlock, Teenie2001 How real is fictive motion? PhD dissertation, Department of Psy-

chology, University of California, Santa Cruz.2004 Fictive motion as cognitive simulation. Memory and Cognition 32:

13891400.Matlock, Teenie, and Daniel C. Richardson

2004 Do eye movements go with fictive motion? Proceedings of the An-nual Conference of the Cognitive Science Society 26: 909914.

Matsumoto, Yo1996a Subjective motion and English and Japanese verbs. Cognitive Lin-

guisti

Documents

[de Knop, Sabine] Cognitive Approaches to Pedagogi(BookFi.org)