Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Perceptual and cultural salience in noun classification : the puzzling case of standard Thai lêm Placzek, James Anthony 1984

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1984_A1 P63.pdf [ 29.2MB ]
Metadata
JSON: 831-1.0076994.json
JSON-LD: 831-1.0076994-ld.json
RDF/XML (Pretty): 831-1.0076994-rdf.xml
RDF/JSON: 831-1.0076994-rdf.json
Turtle: 831-1.0076994-turtle.txt
N-Triples: 831-1.0076994-rdf-ntriples.txt
Original Record: 831-1.0076994-source.json
Full Text
831-1.0076994-fulltext.txt
Citation
831-1.0076994.ris

Full Text

PERCEPTUAL AND CULTURAL SALIENCE IN NOUN CLASSIFICATION: THE PUZZLING CASE OF STANDARD THAI lein By JAMES ANTHONY PLACZEK B . A . , The U n i v e r s i t y of Windsor, 1965 M . A . , The U n i v e r s i t y of B r i t i s h Columbia, 1978 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES ( I n t e r d i s c i p l i n a r y Studies) We accept th i s thes i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 198^ (c) James Anthony Placzek , 1984 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of GRADUATE STUDIES  The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date 17 SEPTEMBER, 198*+ >E-6 (3/81) ABSTRACT i In Standard Thai (ST), the national language of Thailand, the noun c l a s s i f i e r lem applies to books, oxcarts and candles, as w e l l as a v a r i e t y of sharp cutt i n g implements. This c l a s s i f i c a t i o n seems to make no sense whatsoever to Western l o g i c and i n fact i s inexplicable to Thais themselves. In t h i s d i s s e r t a t i o n I work out a probable solution to t h i s puzzle, and in so doing produce insights into the ST c l a s s i f i e r system, and into the p r i n -c i p l e s of noun c l a s s i f i c a t i o n i n general. The orientation o f • t h i s d i s s e r t a t i o n i s empirical, using h i s t o r i c a l , l i n g u i s t i c and ethnographic data in a perceptual prototype model (cf. Hunn 1 9 7 6 ) , which is. i n turn part of a semantic theory of c u l t u r a l stereotypes or Idealized Cognitive Models (cf. Lakoff 1982). The viewpoint of the naive r e a l i s t observer (.cf. Lyons 1977) i s also an important t h e o r e t i c a l f a c t o r . F i r s t , word association data on the above group of a r t i f a c t s was c o l -l e c t e d from u n i v e r s i t y students i n Thailand and submitted to a computer-programmed c l u s t e r analysis i n which, the item "book" emerged as the contem-porary prototype. Next I investigated the shapes, construction procedures and symbolic or r i t u a l associations of these objects. Perceptual s i m i l a r i t i e s rather than symbolic r e l a t i o n s or generic notions dominated the structure of the grouping. The next enquiry was into actual usage among speakers of a sample Thai d i a l e c t which, divides up the ST domain among two d i f f e r e n t c l a s s i f i e r s , one a cognate of ST lem. I t was found that the other c l a s s i f i e r applied more c l e a r l y to the i r o n t o o l s , and the cognate of ST lem more to the bound com-plex objects. In a broader comparison to other languages of the Tai family, i t was i i i found that most languages appeared to apply cognates of ST lem according to a c r i t e r i o n of length or sharpness, and a focus on books and binding was p e c u l i -ar to ST and neighbouring languages strongly influenced by ST. As a r e s u l t of these i n v e s t i g a t i o n s , and i n accord with the t h e o r e t i c a l p r i n c i p l e s outlined i n previous chapters, the best scenario f o r the develop-ment of t h i s c l a s s i f i e r was presented. From an o r i g i n a l grouping around a simple bamboo kn i f e prototype, the group underwent a l a t e r r e s t r u c t u r i n g around a bark paper manuscript prototype, with the s h i f t due to c u l t u r a l and h i s t o r i c a l f a c t o r s . In a summary the main contributions of t h i s d i s s e r t a t i o n were reviewed: ( i ) the importance of the Basic Level of reference and i t s attendant canonical view to the r e l a t i o n s h i p between the perceptual prototype and the c u l t u r a l stereotype (or ICM), ( i i ) the use of word association data and prototype theory i n the study of noun c l a s s i f i c a t i o n , and ( i i i ) the unexpected f i d e l i t y of t h i s important Asian c l a s s i f i e r system to shape c r i t e r i a despite the i n -fluence of strong symbolic and generic f a c t o r s . TABLE OF CONTENTS TITLE i ABSTRACT i i TABLE OF CONTENTS i v LIST OF TABLES ' LIST OF FIGURES LIST OF APPENDICES ACKNOWLEDGMENTS DEDICATION 1 INTRODUCTION 1 . NOTES TO CHAPTER ONE . 1 0 I 2 NATURE OF THE PROBLEM 11 2.1 LACK OF A MEANING FOR lfm 12 2.2 LACK OF A COMMON CRITERION OF CLASSIFICATION . . . . . 15 NOTES TO CHAPTER TWO ' 21 3 THEORETICAL POSITION .22 3.1 SOME BASIC ASSUMPTIONS 22 3.2 GRADING 25 3.2.1 THE FOCUS OF GRADED CATEGORIES: PROTOTYPES . . . . 27 3.2.1.1 PROTOTYPE AS IMAGE . . . . , . 30 3.2.2 THE EXTENSION OF GRADED CATEGORIES: DEGREES OF SIMILARITY . 31 3.2.2.1 THE QUEST FOR A WELL-DEFINED CONCEPT 33 3.2.2.2 THE QUEST FOR A CONCEPTUAL CORE 36 3.2.3 APPLICATIONS TO THE PROBLEM 39 3.3 BASIC LEVEL CONCEPTS U7 3.3.1 CHARACTERIZING THE BASIC LEVEL U7 3.3.2 SHIFTING OF THE BASIC LEVEL 52 3.3.3 SHIFTING OF INDIVIDUAL TERMS : . "PROMOTION" TO LIFE-rFORM LEVEL . . . . . . , . 58 3.3.3.1 PROMOTION OF PROTOTYPES 58 3.3.3.2 PROMOTION OF RESIDUALS 60 3.3.k SUMMARY OF THE BASIC LEVEL 62 3.3.5 APPLICATIONS TO THE PROBLEM .6k 3.k CO-EXISTENCE OF GRADING AND TAXONOMY 69 3.H.1 THE DUAL THEORY 70 3.k.2 WHICH CAME FIRST - THE PROTOTYPE OR THE LOGICAL CATEGORY? 73 3.k.3 ACHIEVING A BALANCE BETWEEN NATURAL AND TECHNICAL CATEGORIES 75 3.h.k KINDS OF FEATURES 78 3.h.5 SUMMARY OF CO-EXISTENCE . . . 8 1 3.k.6 APPLICATIONS TO THE PROBLEM 81+ 3.^.6.1 ARE CLASSIFIERS NATURAL OR TECHNICAL CATEGORIES? . . . 8k 3.k.6.2 CULTURAL SALIENCE AND GENERIC CLASSIFIERS 87 3 .U.6.3 PERCEPTUAL SALIENCE AND SHAPE CLASSIFIERS 93 3.5 CULTURAL SALIENCE . ' .- 102 3.5.1 CULTURAL SALIENCE AND LEXICALIZATION (NAMING) . . . 103 3.5.1.2 PERCEPTUAL VERSUS CULTURAL FACTORS IN NAMING . . . . . 106 3.5.2 DEFINING CULTURE . 1 0 9 3.5.2.1 CULTURE AND MEANING I l l 3.5.2.2 CULTURAL PROTOTYPES . . . " Ilk 3.5.3 APPLICATIONS TO THE PROBLEM 117 3.6 SUMMARY OF THEORETICAL APPLICATIONS 118 NOTES TO CHAPTER THREE 121 V 4 THE WORD ASSOCIATION STUDY 125 4.1 WORD ASSOCIATION 125 4.1.1 CONSISTENCY 125 4.1.2 FREQUENCY OF OCCURRENCE IN DAILY USAGE 126 4.1.3 SYNTACTIC CLASS 130 h.l.k CULTURAL SALIENCE, CULTURAL PROTOTYPE, AND WORD ASSOCIATION . 130 4.1.5 SUMMARY OF WORD ASSOCIATION 133 4.2 CLUSTER ANALYSIS 135 4.3 INFORMANTS l U 5 4.4 ITEMS FOR THE QUESTIONNAIRE l47 4.5 VARIABLES " 148 4.6 PROCEDURE 150 4.6.1 ADMINISTRATION . . „ 150 4.6.2 CODING . . . . ' . . . 150 4.6.3 CLUSTER ANALYSIS . . 152 4.7 RESULTS . . . .' 153 4.8 INTERPRETATION ' 153 4.8.1 REGIONAL DISTINCTIONS 153 4.8.2 THE DOMINANCE OF BOOKS 154 4.8.3 THE CLUSTERING OF THE ANOMALOUS ITEMS l 6 2 4.9 SUMMARY OF THE WORD ASSOCIATION STUDY l 6 5 NOTES TO CHAPTER FOUR . l 6 8 5 THE ETHNO-HISTORICAL STUDY 170 5.1 CARTS . 170 5.1.1 KINDS OF CARTS . . . . 170 5.1.2 CONSTRUCTION OF CARTS 173 5.1.3 CULTURAL SIGNIFICANCE OF CARTS 175 5.1.4 A CONTEMPORARY CULTURAL STEREOTYPE FOR CARTS . . . . 176 5.1.5 SUMMARY FOR CARTS 178 5.2 KNIVES 178 5.2.1 KINDS OF KNIVES 179 5.2.2 CONSTRUCTION OF KNIVES . l 8 l 5.2.3 CULTURAL SIGNIFICANCE OF KNIVES 185 5.2.4 A CONTEMPORARY CULTURAL STEREOTYPE FOR KNIVES . . . . 192 5.2.5 SUMMARY FOR KNIVES 195 5.3 BOOKS ' . . . . 197 5.3.1 KINDS OF BOOKS 197 5.3.2 CONSTRUCTION OF BOOKS . 2 0 3 5.3.3 CULTURAL SIGNIFICANCE OF BOOKS 205 5.3.4 SUMMARY OF CULTURAL SIGNIFICANCE OF BOOKS 213 5.3.5 A CONTEMPORARY CULTURAL STEREOTYPE FOR BOOKS . . . . 2 1 4 5.4 ETHNO-HISTORICAL EVIDENCE ON THE APPLICABILITY OF lfm . . 217 5.4.1 PERCEPTUAL SALIENCE 217 5.4.1.1 CARTS AND SLEDGES . . . . . 218 5.4.1.2 KNIVES, SWORDS, AND OTHER CUTTING INSTRUMENTS . . . . 219 5.4.1.3 BOOKS 220 5.4.1.4 CANDLES AND FANS 221 5.4.1.5 SUMMARY OF PERCEPTUAL SALIENCE 222 5.4.2 CULTURAL SALIENCE 223 5.4.2.1 CARTS AND SLEDGES . . . . . . . . . . . 223 5.4.2.2 KNIVES, SWORDS, AND OTHER CUTTING INSTRUMENTS . . . . 2 2 5 5.4.2.3 BOOKS 226 5.4.2.4 CANDLES AND FANS . . . . • . . 2 2 8 5.4.3 THE INTERACTION OF PERCEPTUAL AND CULTURAL SALIENCE . . . 2 2 9 5.4.4 DIRECTIONALITY 230 NOTES TO CHAPTER FIVE 232 v i 6 THE COMPARATIVE-HISTORICAL STUDY 237 6.1 THE TAI FAMILY- 237 6.1.1 GENETIC RELATIONS 237 6.1.2 HOMELANDS 2l+2 6.2 APPLICABILITY OF lem IN NE: A LOCAL VIEW 2*+5 6.2.1 TRANSCRIPTION 21+5 6.2.2 THE DIALECT 2U6 6.2.3 INFORMANTS . 2h6 6.2.1+ FIELD METHODS 2l+7 6.2.5 BASIC DATA 2L8 6.2.6 ANALYSIS 250 6.3 COMPARISON WITH ST lem 256 6.1+ . ETYMOLOGY OF ST lem 259 6.1+.1 CORRESPONDENCES 259 6.1+.1.1 THE SOUTHWESTERN BRANCH . . . . . . . ' . . . 259 6.1+.1.2 THE CENTRAL BRANCH 2 6 l 6.1+.1.3 THE NORTHERN BRANCH 2 6 l 6.1+.1.1+ DISTRIBUTION OF COGNATES 263 6.1+.2 RECONSTRUCTION - 2 6 3 6.1+.2.1 PHONOLOGICAL SEGMENTS . 261+ 6.1+.2.2 SEMANTIC VALUE 267 6.1+.2.2.1 WORD FAMILIES 267 6.1+.2.2.2 SEMANTIC RECONSTRUCTION 272 6.1+.2.2.2.1 A COMPONENTIAL VIEW 278 6.1+.2..2.2.2 A PROTOTYPICAL VIEW 282 6.1+.2.2.2.3 METAPHORICAL VIEWS 293 6.1+.2.2.2.1+ A SYNTHETIC VIEW 297 NOTES TO CHAPTER SIX 307 7 SUMMARY AND CONCLUSIONS 312 7.1 COGNITIVE PRINCIPLES IN ST NOUN CLASSIFICATION . . . 312 7.2 SEMANTIC RELATIONS BETWEEN ST CLASSIFIERS AND HEADWORDS . 313 7.3 DISTINGUISHING AND EVALUATING PERCEPTUAL VERSUS CULTURAL CRITERIA OF CLASSIFICATION . 3l6 7.^ DEVELOPMENT OF CLASSIFIERS AND CLASSIFIER SYSTEMS . . . 320 7.5 REFINING THE THEORY OF NOUN CLASSIFICATION . . . . 3 2 8 NOTES TO CHAPTER SEVEN 333 BIBLIOGRAPHY 33^ APPENDICES 355 v i i LIST OF TABLES 1. LINES OF ENQUIRY IN THIS DISSERTATION AT GENERAL AND SPECIFIC LEVELS 5 2. NOUNS CLASSIFIED BY lem IN STANDARD THAI . . . . . . . . 11 3. POSSIBLE CRITERIA OF CLASSIFICATION AS lem 15 k. ST SPEAKERS' CONSCIOUS CATEGORIZATION OF ST lem ITEMS IN A "FREE SORT" TASK . 1 8 5. ST lem ITEMS NOT PRIMARILY CATEGORIZED AS "UTENSILS" IN TABLE k . 18 6. CRITERIA OF CLASSIFICATION AND NOMINAL SENSE OF SOME MORE COMMON ST CLASSIFIERS FOR CONCRETE OBJECTS . hi 7. NAMING CHARACTERISTICS OF FOLK-BIOLOGICAL CATEGORIES BY TAXONOMIC LEVEL, ACCORDING TO DIRECT OR INDIRECT DEPENDENCE UPON ENVIRONMENT 57 8. FREQUENCY OF SEMANTIC GROUPS AMONG NOUNS CLASSIFIED BY an i IN HAAS (1965) . . . . . . . 98 9. DEVELOPMENTAL DIFFERENCES IN CATEGORIES OF WORD ASSOCIATION RESPON-SES BETWEEN JAPANESE AND CANADIANS . . 1 3 2 10. HYPOTHETICAL MATRIX OF RESPONSE DISTRIBUTION 138 11. POPULATION SAMPLED IN THE SEMANTIC STUDY ll+7 12. QUESTIONNAIRES COMPLETED (BY MAJOR DIALECT REGION) . . . . 1 5 1 13. RESPONSES TO "BOOK" ELICITED BY SELECTED STIMULUS-WORDS (CODED DATA) 158 l i t . DISTRIBUTION OF STIMULI WHICH ELICITED lem (CODED DATA) . . . 1 5 9 15. SUMMARIZED RESPONSE FREQUENCIES FOR THE ITEM kvian (NO. 8 l ) . . 177 16. SUMMARIZED RESPONSE FREQUENCIES FOR THE ITEM m i i d (NO. 31) . . 192 17. STIMULUS-WORDS WHICH ELICITED SELECTED TERMS AS RESPONSES, RANKED BY FREQUENCY 193 18. SUMMARIZED RESPONSE FREQUENCIES FOR THE ITEM aawud "WEAPON" (NO. 58) 19 k 19. CLASSIFIER APPLICABILITY IN THE CONSTRUCTION OF CERTAIN ARTIFACTS . 217 20. TONAL SYSTEM NOTATION FOR THE LOCAL DIALECT OF NONG TEUN VILLAGE, MAHASARAKHAM PROVINCE (NE) 2^5 21. INFORMANT BACKGROUND (NE) 2k6 22. CLASSIFIER APPLICATION TO ST NOUNS IN NONG TEUN VILLAGE, MAHA-SARAKHAM, NORTHEAST THAILAND . 2U8 23. ITEMS CLASSIFIED PRIMARILY BY duan, AND BY l i m IN NE. . . . . 251 2k. OVERLAP BETWEEN HIGHER FREQUENCY duan, ITEMS AND l i m ITEMS IN TABLE 23 253 25. SIAMESE AND NORTHERN BRANCH CORRESPONDENCES OF PROTO-TAI *e_ . . 262 26. SIAMESE AND OTHER TAI CORRESPONDENCES OF PROTO-TAI * h l - O R * 1 - . . 262 27. CLASSIFIERS FOR SOME TAI LANGUAGES ACCORDING TO SEMANTIC DOMAIN . 2jk 28. COMPLEMENTARITY OF * h l e m CI AND * sen CI IN TAI LUE (XI DAI) AND TAI NEUA (DEHONG DAI) , BOTH LANGUAGES OF YUNNAN, SOUTHERN CHINA . . 2 7 6 v i i i LIST OF FIGURES 1. THE "STANDARD THEORY" OF TAXONOMIC LEVELS OF INCLUSION IN FOLK BIOLOGY 48 2. TZELTAL FOLK TAXONOMY OF "GRAINS" AFTER BERLIN ET AL. (1977 :223) 59 3. DIFFERENT LEVELS OF LEXICALIZATION IN THE MAIN LEXICON AND IN THE CLASSIFIER SYSTEM 68 4.. TAXONOMIC ANALYSIS OF ASIAN NOUN CLASSIFIER CATEGORIES FOR CONCRETE OBJECTS 89 5. CLUSTER ANALYSIS OF THE ST lem GROUP BY EMBEDDED MEASURE, WITH REPRESENTATIONAL RESPONSE 140 6. CLUSTER ANALYSIS OF THE ST l fm GROUP BY EMBEDDED MEASURE ONLY . 144 7. NORTHEASTERN AND CENTRAL STYLES OF kvian CHASSIS . . . . 173 8. BAMBOO MESSAGE SYSTEMS OF TRIBAL GROUPS NEIGHBOURING TAI AREAS IN MODERN TIMES, COMPARED TO AN ANCIENT CHINESE WRITING SYSTEM 188 9. PERCEPTUAL SIMILARITIES AMONG SOME lem GROUP ITEMS IN'BOTH-OPEN AND CLOSED POSITIONS . . . 2 2 4 10. THE TAI FAMILY TREE . 238 11. DISTRIBUTION OF THE TAI LANGUAGE FAMILY . . . . . . 2 3 9 12.. GENETIC RELATIONS AMONG KNOWN COGNATES OF ST lem IN THE TAI FAMILY 273 13. GEOGRAPHICAL DISTRIBUTION AND ASSUMED DISPERSION PATTERN OF ST l fm IN THE TAI FAMILY 277 14. COMMON SECONDARY FORMS OF BAMBOO USED IN SOUTH EAST ASIAN FOLK TECHNOLOGY AND NAMED OR CLASSIFIED IN STANDARD THAI- . . . 2 8 4 15. ONE SCENARIO OF DEVELOPMENT FOR ST lem SUGGESTED BY PERCEPTUAL SIMILARITIES AMONG OBJECTS CLASSIFIED 298 16. PARALLEL PERCEPTUAL SIMILARITIES BETWEEN ARTIFACTS PRODUCED FROM PALM LEAF AND ARTIFACTS DEVELOPED FROM A BAMBOO PROTOTYPE . . 301 17. ANOTHER SENARIO OF DEVELOPMENT OF ST lem SUGGESTED BY BOTH PERCEPTUAL AND CULTURAL SALIENCE 304 ACKNOWLEDGMENTS The fo l lowing i n s t i t u t i o n s have funded t h i s re search : The U n i v e r s i t y o f B r i t i s h Columbia, the I .W. K i l l am Memorial Fund, and the S o c i a l Sciences and Humanities Research C o u n c i l o f Canada. The fo l lowing i n s t i t u t i o n s have permitted or g r e a t l y f a c i l i t a t e d the research : The Nat iona l Research Counci l o f T h a i l a n d , The Nat iona l L i b r a r y o f T h a i l a n d , the Thai Khadi Seuksa Research I n s t i t u t e o f Thammasat Univer -s i t y , Thammasat U n i v e r s i t y , Mahachulalongkorn Buddhist U n i v e r s i t y , Chula-longkorn U n i v e r s i t y , Khon Khaen U n i v e r s i t y , Sr inakhar inwirot U n i v e r s i t y , Rajadamnern Commercial C o l l e g e . The three people who have made t h i s research p o s s i b l e a re , f i r s t o f a l l T i p P laczek , whose energy and s a c r i f i c e s can never f u l l y be known, Professor M.D. Kinkade, whose confidence i n t h i s work was a constant , and Professor R . J . Pearson, whose enthusiasm and d i s c i p l i n e were e s s e n t i a l . Of the many many other people who have contr ibuted adv ice , re fe rences , contac t s , and other a s s i s t ance , the fo l lowing people must be mentioned: Nanthana Danwiwat W . J . Gedney Nanthana & T h i t i Hengrasami W i l a i & Chai-amnuay J i tphokha Chatsumarn K a b i l s i n g h Wilaiwan Kanitthanan 'Charnwit K a s e t s i r i Charles F . Keyes John Le Roy Simon Yeung Y a i Mathurodmeethanii Carol Mayer Chalee Sodphrasert Boonlert Sodsuchaad David Strecker Saowalak Vorrakuttanont & family Ngaw Khamvicha F i n a l l y , s p e c i a l thanks are due to Frank F lynn and Tomiso Okuda who helped wi th the data p r o c e s s i n g , and to the members o f the Defense Committee who, along with Penny Hanson, brought t h i s study m e r c i f u l l y to a c lo se . i x LIST OF APPENDICES A . ILLUSTRATIONS 355 B. RESPONSE DISTRIBUTIONS TO "FRUIT" IN VARIOUS WORD ASSOCIATION STUDIES 370 C. RAW RESPONSE DATA FOR THE ITEM lem (CLASSIFIER) . . . . 372 D. SUMMARIZED RESPONSES FOR ST lem 372 E. EMBEDDED RESPONSES FOR THE ST l fm GROUP 373 F . SAMPLE QUESTIONNAIRE SHEET WITH TRANSLATIONS 376 G. TRANSLATION OF QUESTIONNAIRE EXPLANATION AND INSTRUCTIONS . . 3 8 l H. COMPARISON OF THE CODED SAMPLE AND THE ENTIRE.DATA SAMPLE FOR ITEMS; 1 ("BOOK") AND 83 (lfm) 383 I . NOUNS CLASSIFIED BY VARIOUS REGIONAL CLASSIFIERS . . . . 384 J . RANGES OF APPLICABILITY OF SOME ST lem COGNATES IN TAI LANGUAGES WITHIN THAILAND 385 K. REFLEXES OF PROTO-TAI ? dan A l IN SOME TAI LANGUAGES 1 . . . 386 L. THE HYPOTHETICAL ARTIFACT SPACE OF ST lfm MAPPED ONTO THE HYPOTHETICAL CLASSIFIER SPACE OF RELATED SHAPE.CLASSIFIERS . . 387 DEDICATION to T i p 1.0 INTRODUCTION By now there have been several major studies of noun c l a s s i f i e r systems as topics i n t h e i r own r i g h t , rather than as sub-sections of o v e r a l l gram-mars of-a given language."'" A valuable survey of c l a s s i f i e r languages and a summary of the differences between noun class systems and noun c l a s s i f i e r systems i s provided by Dixon ( 1 9 8 2 ) , and a very good comparative study by Conklin ( l 9 8 l ) looks at Tai and Austronesian c l a s s i f i e r systems i n some 2 h i s t o r i c a l . p e r s p e c t i v e . A l l a n (1977) has written another wide-ranging survey which finds that the paradigm type of c l a s s i f i e r language i s the numeral c l a s s i f i e r language, a type which includes the Tai familyQwhich i n turn Jones (1970) has sugges-ted may have been the o r i g i n a l source f o r the Asian c l a s s i f i e r systems. Cer-t a i n l y the Standard Thai system i s one of the most elaborate and obligatory, and the Tai family of languages shows the greatest uniformity and s t a b i l i -ty of c l a s s i f i e r usage among East and South East Asian language f a m i l i e s . I f we accept t h i s c e n t r a l p o s i t i o n of the Tai family among c l a s s i f i e r languages, the fact that Standard Thai i s the most prominent language i n the family (as the only national language of an independent Tai state) makes the study of ST c l a s s i f i e r s l i k e l y to produce s i g n i f i c a n t generalizations with isLevance to many or most other languages with s i m i l a r constructions and ways of categorizing things. Thus i t i s not s u r p r i s i n g that many writers have considered-the Tai system d i r e c t l y (e.g. Conklin 1 9 8 l ) or i n -d i r e c t l y as a source of examples for various t h e o r e t i c a l approaches (e.g. Al l e n 1977). While data on the ST c l a s s i f i e r system has been accumulating, and many cross-language studies have u t i l i z e d such data, i t has not yet been found necessary to study the extension or h i s t o r i c a l development of a single c l a s -s i f i e r or c l a s s i f i e r category. For most c l a s s i f i e r s such a study i s un-necessary, since they have a straightforward range of a p p l i c a b i l i t y ; that i s , they apply to a group of nouns with a common and prominent a t t r i b u t e (e.g. ST med c l a s s i f i e r for small seed-like objects). As in t h i s example, i n many of these cases the c l a s s i f i e r has an al t e r n a t i v e function as noun with a cl e a r semantic value. But there are some cases where the c l a s s i f i e r has no obvious c r i t e r i o n of a p p l i c a t i o n : i t i s a puzzle, an example of the '"obscure psychology" of other cultures which Durkheim (1965:171) so despaired of ever understand-ing. It does appear that to study the extension and c r i t e r i a of a p p l i c a -t i o n of one of these puzzles of fol k c l a s s i f i c a t i o n , the enquiry must ex-plore the s o c i a l and h i s t o r i c a l influences as well as the psychological and semantic processes which have combined to produce i t . That i s what t h i s d i s s e r t a t i o n sets out to do, and that i s why i t i s interdisciplinary. But perhaps such a s p e c i f i c problem i s r e a l l y just a t r i v i a l matter which i n descriptive terms i s too language-specific, and i n essence i s just another accidental conjunction of h i s t o r i c a l , c u l t u r a l , and l e x i c a l facts which o f f e r no hope of fresh insights generalizable beyond the s p e c i -f i c case at hand. Against t h i s twin challenge ( o v e r - s p e c i f i c i t y and i n s o l v a b i l i t y ) It must be pointed out that there i s an es s e n t i a l balance required between description and theory. I t would be unnecessary to repeat such a truism were we not s t i l l experiencing the effects of extremes i n l i n g u i s t i c theo-ry which have l e d to a severe narrowing of the scope of l i n g u i s t i c enquiry 3 to issues of philosophical syntax. In an inductive approach, broad theo-r e t i c a l issues and axioms a r i s e from a consideration of data. Once a r t i -culated, these issues and axioms must return to the data for t e s t i n g and refinement or abandonment. Of course "the data" does not exist i n nature, but must be gathered and i n e v i t a b l y shaped by the assumptions of the com-3 p i l e r . Ultimately, we do not need to take a p o s i t i o n on the primacy of either data or theory, since we have i n h e r i t e d a r i c h t r a d i t i o n of both. The task i s to continue the d i a l e c t i c between them. But note the fact that the u t i l i t y of a body of data i s reduced proportionally as the r e -searcher depends upon a s p e c i f i c theory. It i s also true that data-based studies increase i n value over time as t h e i r data base becomes h i s t o r i c a l -l y obscured, while theory-oriented studies decrease in value over time as t h e i r t h e o r e t i c a l base i s r e f i n e d or abandoned. For t h i s reason I f e e l that a simple d e s c r i p t i o n , based on e x p l i c i t assumptions and r e l y i n g as l i t t l e as possible on extreme t h e o r e t i c a l positions i s valuable and worth-while i n and for i t s e l f and for the use which can be made of i t by subse-quent scholars. In t h i s d i s s e r t a t i o n , however, I apply a set of p r i n c i p l e s which, while short of a f u l l - s c a l e model of semantic structures, have d e f i n i t e implica-tions for psychological processes of organizing knowledge; that i s , human concept formation. This i s the "perceptual prototype" viewpoint outlined i n Chapter Three. Thus I hope to s a t i s f y a need for t h e o r e t i c a l o r i e n t a -t i o n i n t h i s data and i n so doing provide an opportunity . • to t e s t the theory. My aims here are to generate new t h e o r e t i c a l issues and r e f i n e e x i s t i n g ones. In doing so I hope to contribute to the i n t e -gration of what i s known of c l a s s i f i e r systems into a model of psychologi-c a l conceptualization i n a f u l l h i s t o r i c a l and s o c i a l context. Ultimately I would l i k e to open up the p o s s i b i l i t y that c l a s s i f i e r systems, l i k e tonal systems before them, can be promoted from t h e i r present p o s i t i o n as merely vexing obstacles to language analysis to a p o s i t i o n where they are a major t o o l i n h i s t o r i c a l and comparative studies of the cultures and s o c i e t i e s who use them. This i s much more of a challenge i n the f i e l d of ethno-se-mantics than i n phonology. But since both meaning (here emphasizing r e a l -world reference) and sound are the e t i c foundations upon which language stands, a p h i l o l o g i c a l enquiry such as t h i s one must accord with both the semantic facts discovered and the phonological facts as w e l l as they are known. In a h i s t o r i c a l perspective the phonological facts are best ap-proached i n the well-established methodology of historical-comparative r e -construction. To make a st a r t we need an empirical basis i n a body of data. I have chosen data from the r i c h e s t c l a s s i f i e r system i n East and South East Asia. The Thai c l a s s i f i e r s , c o n s t i t u t i n g as they do a n a t u r a l l y developed and comprehensive system, are an i d e a l place to test putative p r i n c i p l e s of human categorization. Perhaps another point of entry would have done just as w e l l ; the e s s e n t i a l idea i s that we begin with an empirical reference point and proceed from there. Assuming that the problem as outlined above is at a l l solvable, the l i n e of enquiry I hope to follow i s r e f l e c t e d i n the following series of f i v e questions, where general (= t h e o r e t i c a l ) issues re l a t e d i a l e c t i c a l l y to s p e c i f i c issues* The f i v e questions are found in Table 5 below. In Chapter Two below I review the general state of knowledge about the Standard Thai (.ST) noun c l a s s i f i e r lem and how my own e a r l i e r enquiries, (es. p e c i a l l y Placzek. 1978) have only deepened the problem. A review of the important recent advances i n the understanding of the p r i n c i p l e s of human categorization follows i n Chapter Three. From t h i s r e -view I conclude that the basic categories of natural language are prototypi c a l ; that i s , they exhibit GRADING from a perceptual prototype (e.g. apples as the p r o t o t y p i c a l f r u i t i n northern European cultures) to peripheral or marginal members of the category (e.g. r a i s i n s , coconuts and f i g s ) . This characterization, of categories f i t s well with, the problem under study here. I conclude that the group of nouns c l a s s i f i e d by ST lem do not 5 TABLE 1 LINES OF ENQUIRY IN THIS DISSERATION, AT BOTH GENERAL AND SPECIFIC LEVELS GENERAL LEVEL OF ENQUIRY CHAPTER SPECIFIC LEVEL OF ENQUIRY CHAPTER 1. What cognitive p r i n c i p l e s 2,7 What kind of category i s 2 , 3 , 4 apply i n ST noun c l a s s i f i - ST lem? cation? 2. What semantic r e l a t i o n s 3,7 What are the c r i t e r i a of 2 , 3 , 4 , 5,6 hold between ST c l a s s i - c l a s s i f i c a t i o n for ST lem?" f i e r s and t h e i r headwords? 3. Can perceptual and c u l t u - 2 ,4 , What are the perceptual 2 , 3 , 4 , 5,6 - r a l salience be d i s t i n - 1,7 and c u l t u r a l a t t r i b u t e s of guished and evaluated? the ST'.lem group? 4. What constraints are : 3,7 What i s the best scenario 2,6 there on the development of the development of ST of c l a s s i f i e r s and c l a s s i - lem? f i e r systems i n ST 5. How can the theory of 7 What predictions can be 7 noun c l a s s i f i c a t i o n be re- made about the development fined? of ST lem i n the future? NOTE: Underlined chapter numbers show the place where the problem i s p r i m a r i l y addressed. constitute a homogeneous C l a s s i c a l (or A r i s t o t e l i a n , here c a l l e d "Technical") category, produced by necessary-and-sufficient features^- and therefore t h i s group must have degrees of membership, fuzzy edges and a f o c a l item or proto-type. This conclusion i s reinfo r c e d by an analysis of the most common ST c l a s -s i f i e r s . This analysis reveals that even those c l a s s i f i e r s which seem to op-erate on a generic c r i t e r i o n ( i . e . , kind of thing, natural kind, essence, etc.) are i n every case c l o s e l y t i e d to the much more widespread perceptual c r i t e r i o n ( e s p e c i a l l y shape). Evidence of p r o t o t y p i c a l i t y alone i s i n s u f f i c i e n t to explain the cross-c u l t u r a l differences and s i m i l a r i t i e s i n human categorization. P r o t o t y p i c a l -i t y must be combined with another factor, the empirically-established notion of a Basic Level of human i n t e r a c t i o n with e n t i t i e s i n the environment, i n terms of perception, sensorimotor manipulation and naming. Different cultures may have d i f f e r e n t Basic Levels for some domains, yet despite i n t e r c u l t u r a l , i n t e r l i n g u a l and interpersonal v a r i a t i o n i n t h i s Basic Level of reference there appears to be a common Basic Level available to a l l members of a speech community which provides a shared frame of reference within which to communi-cate. This Basic Level:is then characterized as a culture-wide s i t u a t i o n a l stereotype or prototype. Category prototypes (e.g. apples) occur at the Ba-s i c Level, and the Basic Level i n turn i s equivalent to the Folk-Generic l e -v e l of f o l k taxonomies i n Ethnoscience (at l e a s t for r u r a l people). In order to understand how a p r o t o t y p i c a l c l a s s i f i e r category relates to the r e a l world, i n p a r t i c u l a r to the a r t i f a c t s of material culture which the ST lem group includes, i t i s necessary to r e l a t e the c l a s s i f i e r system to the Basic Level. Simple nouns (e.g. "apple") are at the Basic Level. But c l a s s i f i e r s r e f e r to a t t r i b u t e s of Basic Level objects which are usually attended to and named only i n a close i n t e r a c t i o n with the object. In the case of a r t i f a c t s , such close i n t e r a c t i o n may be taken for granted. However, we s t i l l need to account for the existence of a c l a s s i f i e r system as a sepa-rate and p a r a l l e l l a b e l i n g system operating alongside the ordinary lexicon of Standard Thai, for example the c l a s s i f i e r f o r "round t h i n g " (luug) along-side the common adjective "round" (klom). A functional j u s t i f i c a t i o n for the seemingly redundant c l a s s i f i e r sys-tem i s found i n the fact that c l a s s i f i e r s only occur i n certain r e s t r i c t e d environments where precise i n d i v i d u a t i o n i s required. This special-purpose c l a s s i f i c a t i o n requires a special-purpose set of l a b e l s , the c l a s s i f i e r sys-tem i t s e l f . This special-purpose function also explains why a t t r i b u t e s of shape so completely dominate most c l a s s i f i e r systems.. It i s the shape which, supplies the countable, structured, bounded en t i t y which i s being counted, l i s t e d , selected, contrasted or otherwise i n d i v i d u a l l y r e f e r r e d t o . I take these conclusions as support for the claims that ST lem must have a perceptual prototype which may or may not be i d e n t i c a l with i t s o r i g i n a l 7 prototype, and that the o r i g i n a l prototype almost c e r t a i n l y was an a t t r i b u t e of shape, or a complex of shape a t t r i b u t e s . In Chapter Four I set out to f i n d the contemporary prototype for ST lem, using word association data to produce a graphic representation of the degrees of s i m i l a r i t y between a l l items c l a s s i f i e d by ST lem and between each item and the c l a s s i f i e r i t s e l f . I assume word association frequency and dominance to be d i r e c t indicators of p r o t o t y p i c a l i t y , since word associations d i r e c t l y tap the c u l t u r a l stereotype. The r e s u l t s are unequivocal in showing the item "book" (and not, as expected, any of the c u t t i n g instruments) to be the proto-type. This evidence i s supported by other data discussed i n t h i s chapter. In Chapter Five 1 examine the kinds, the perceptual salience, the meth-ods of construction^and some of the r i t u a l and symbolic s i g n i f i c a n c e of the objects i n the ST lem group, focusing on books and carts as prominent anoma-l i e s , and knives as representative of the majority of items which are cutt i n g implements of some kind. In conjunction with the word association data,books were shown to have an extremely high value both i n s o c i o - p o l i t i c a l terms and as symbols of ethnic i d e n t i t y . Knives and carts were found to be much l e s s prominent. An a d d i t i o n a l factor of binding, p a r a l l e l structure, stacking or f o l d i n g was also found to be s a l i e n t i n many of the items; i t may be a secon-dary r e s u l t of the p r o t o t y p i c a l i t y of books. In Chapter Six I look at the a p p l i c a b i l i t y of a cognate of ST lem i n the c l o s e l y r e l a t e d d i a l e c t of Nong Teun v i l l a g e i n Mahasarakham, North East Thailand. In contrast to the word association data, t h i s data i s from middle-aged and e l d e r l y v i l l a g e r s . This l o c a l d i a l e c t was found to s p l i t the ST lem group into at least two main groupings: one group was c l a s s i f i e d by duan,, and included the metal cutt i n g tools plus some a r t i f a c t s with an e s s e n t i a l mecha-nism made of metal. The other group was c l a s s i f i e d by l i m , cognate of ST lem, and seemed to comprise non-metal a r t i f a c t s i n bound p a r a l l e l structures. \ 8 When these r e s u l t s are contextualized i n a comparative study of cog-nates of ST lem i n the Tai family i t i s cl e a r that cognates of t h i s c l a s -s i f i e r are probably shape-based, that books and cuttin g t o o l s are c l a s s i -f i e d i n diverse ways i n d i f f e r e n t Tai languages, and that Standard Thai i s the ONLY Tai language to combine just books, c a r t s , knives, etc. i n a singl e c l a s s i f i e r grouping. F i n a l l y , two ways of assessing the comparative data are o u t l i n e d : a componential analysis and a comparison to well-known metaphorical themes. These two approaches were found to have d i f f e r e n t and sometimes contradic-t i n g r e s u l t s , and the Standard Thai c l a s s i f i e r system was found to be p a r t i c u l a r l y aberrant i n i t s p a r t i c i p a t i o n i n the plant-parts metaphor so widespread i n languages and language families of the area. The ST c l a s s i -f i e r lem was found to be aberrant within t h i s already unusual system. A more useful approach focuses on basic shapes prominent i n the f o l k -technology. In the t r a d i t i o n a l technology bamboo was paramount, with other kinds of wood, stone, iron and s h e l l s also i n use. The name for some com-mon secondary forms of bamboo are also c l a s s i f i e r s . A study of the more common secondary forms showed a d e f i n i t e c l a s s i f i e r f o r most of these, but the common form of a bamboo s l i p (used e i t h e r f or w r i t i n g , t a l l y i n g , send-ing messages, or cutting) was found to have no widely-recognized c l a s s i f i e r other than an, the general c l a s s i f i e r . In a synthesis of these findings and of the other conclusions reached above, a probable scenario of the de-velopment of t h i s c l a s s i f i e r i s out l i n e d , and the supporting evidence as-sessed. In the f i n a l chapter, the value of the t h e o r e t i c a l approach i s as-sessed, e s p e c i a l l y the dual theory of categorization processes, Hunn's perceptual model, and the notion of noun c l a s s i f i c a t i o n as a s p e c i a l pur-pose c l a s s i f i c a t i o n at a finer-grained l e v e l of perceptual i n t e r a c t i o n with the environment. The p o s s i b i l i t y of an " a r t i f a c t space" constructed from perceptual and l e x i c a l semantic continua was presented. The metho-dologies of word association, e l i c i t a t i o n of c l a s s i f e r usage, and d i f f e r e n t historical-comparative approaches are also assessed, and recommendations for future research are put forward. F i n a l l y , the contribution o f t h i s intensive l o n g i t u d i n a l study of a single c l a s s i f i e r and i t s domain of ap-p l i c a t i o n i s evaluated. To s t a r t at the beginning, then, we must examine and delimit the na-ture and extent of the problem at hand, the puzzling facts about the ST lem- class... ' ._ .-10 NOTES TO CHAPTER ONE 1. ( p . l ) C l a s s i c works of t h i s sort are the papers by Adams and Conklin (1973, 191b), Adams et a l (1975), Becker ( 1 9 7 5 ) , A l l a n (1977), B e r l i n ( 1 9 6 8 ) , B u r l i n g ( 1 9 6 5 ) , Denny (l9lb, 1 9 7 5 , 1 9 7 6 a , 1 9 7 6 b ) , Greenberg (1975) , Hass ( 1 9 ^ 2 , 1 9 6 5 ) , Hla Pe ( 1 9 6 5 ) , Jacob ( 1 9 6 5 ) , Jones ( 1 9 7 0 ) , Mathiot ( 1 9 6 2 ) , Placzek ( 1 9 7 8 ) , Saul ( 1 9 6 5 ) , Schaefer ( 1 9 ^ 8 ) , T ' sou (1976) , K i l l i n g l e y ( 1 9 8 0 , 1982) among many others . 2 . ( p . l ) Other works by Greenberg (1977) Lehman (197 *0 and Adam's d i s -ser ta t ion ( 1982) have been consulted here only i n d i r e c t l y through se-condary re ferences . 3. ( p . 2 ) Winograd ( 1977) has documented the step-by-step degeneration o f what he c a l l s the "Chomskian Paradigm". S t r u c t u r a l i s t extremism ( i . e . a concern with the i n t e r n a l organiza t ion of a system at the expense o f invaded the realm o f Anthropology but so far does not seem to have r u l e d out the study o f how r e a l language i s used by people i n r ea l -wor ld s i t u a t i o n s . 11 2.0 THE NATURE OF THE PROBLEM In Standard Thai (ST), the na t iona l language of Tha i l and , there i s a cate-gory which groups the fo l lowing objects"'": TABLE 2 NOUNS CLASSIFIED BY lem IN STANDARD THAI nanjsii books ( in general) phraa bush knives (machetes) samud khgoj t r a d i t i o n a l books mi id koon razors (manuscripts of khem needles bark paper) w i i fans (obsolete term) th i an candles w i i combs majphaaj paddles khraad harrows, rakes siam spades siw c h i s e l s cpob hoes (or mattocks) khwaan axes khan thaj plows daab swords khiaw s i c k l e s hgog spears ,halberds , lances takra j s c i s sor s kwian oxcarts mi id knives ( i n general) thuub incense At a d e s c r i p t i v e l e v e l , the problem i s m u l t i p l e : Table 2 shows. ( i ) no apparent common c r i t e r i o n o f c l a s s i f i c a t i o n to account for the fact that these items a l l occur with the c l a s s i f i e r lem_3 ( i i ) no c l e a r d e l i n e a t i o n of which items are , and which are not , c l a s s i f i e d by lem ( i i i ) no "best example" o f the group ( iv ) no c l e a r meaning for lem, despite the fact that the major i ty of ST c l a s s i f i e r s are e a s i l y suppl ied wi th a g lo s s . Data i s a d d i t i o n a l l y d i f f i c u l t to gather on t h i s problem for other reasons. Objects which are new to Thai cu l ture do not seem to enter the lem category (see Placzek 1978:20). Most new items of technology take an, the general c l a s -s i f i e r commonly used for r e s i d u a l items or items of uncerta in c l a s s i f i c a t i o n , or they take an a l t e r n a t i v e c l a s s i f i e r such as khr^an, "machine, d e v i c e , e t c . " ) . Farmers and less-educated people of the Centra l P l a ins region (whose dia--l e c t ought to be c loses t to ST) use lem very l i t t l e . They tend to use the general c l a s s i f i e r an for many of the items of Table 2 , even more than do speakers of ST and other major r eg iona l d i a l e c t s . These aspects of the de-s c r i p t i v e problem are discussed i n more d e t a i l below. 2.1 LACK OF A MEANING FOR lem The majority of Thai noun c l a s s i f i e r s can "be given a gloss, either because they "have a meaning" for native speakers (often with independent function as a noun) or because the range of nouns which they c l a s s i f y transparently r e -veals some common c r i t e r i o n of c l a s s i f i c a t i o n . At the other extreme, a small r e s i d u a l group of c l a s s i f i e r s has no a t t r i b u t a b l e meaning. Placzek (1978) l a -beled t h i s group the "extended" c l a s s i f i e r s . Of those c l a s s i f i e r s which are glossable, Placzek (1978) labeled as " F u l l Repeaters" those which merely repeat the head noun"'in the c l a s s i f i e r phrase, for example khaabsamud SQOQ. khaabsamud isthmus two ISTHMUS^ "two isthmuses" These items are mostly geographical formations, body parts and abstractions. Another group was labeled " P a r t i a l Repeaters". These repeat part of the compounded head noun, a part which may be either a class (generic) term or refer to the shape of the object named: tua bun, saam tua body-shape c a t e r p i l l a r 3 BODY-SHAPE "Three c a t e r p i l l a r s " ST.lem, however, i s never repeated and r a r e l y , i f ever, occurs as a generic. But t h i s c l a s s i f i e r does have a meaning of a vague sort. Most speakers of ST, when queried about the meaning of lem, w i l l say that i t means a "long 6 thing". Unfortunately, there are approximately 39 other ST l e x i c a l items r e -f e r r i n g to "long things" which can function i n the c l a s s i f i e r p o s i t i o n , and even i f we r e s t r i c t these to c l a s s i f i e r s other than repeaters and p a r t i a l r e -peaters and further to only concrete objects, there are s t i l l 13 other c l a s s i -f i e r s which r e f e r to various kinds of "long things". Among these the follow-ing c l a s s i f i e r s seem to be competitors for the ' semantic space of "long thing • kaan ( e l f . f o r ) "stem, stalk", daam "handle, h i l t , h o l d e r " , e l f . for pens 13 dun ( e l f . fo r ) "p ieces of firewood, bone, etc'.' thasaen, "bar , ingot " thpon " s e c t i o n , par t , p i e c e " Among the c la s s terms which funct ion as P a r t i a l Repeaters we also have l ikely-competitors : khan "paddy f i e l d dyke, long handle, e t c . " s i i " long items i n p a r a l l e l , spaced arrangement" dgog "blossom", e l f . for incense s t i c k s , arrows, etc . lam e l f . for bamboo, condui t s , boats , channels , e tc . A componential ana lys i s might be able to sort these few c l a s s i f i e r s i n -to a s t r u c t u r a l set , but the ana lys i s would require a separate d i s t i n g u i s h i n g feature for each i tem. With the r a r i t y , overlapping a p p l i c a t i o n and v a r i a b l e usage of c l a s s i f i e r s across i n d i v i d u a l s and d i a l e c t s there would u l t i m a t e l y be no recourse for v e r i f i c a t i o n o f the f ea tura l ana lys i s beyond the ana ly s t ' s own subject ive reckoning , no matter how rigorous or elegant the outcome. A point against accepting " long t h i n g " as a meaning for lem i s the fact that severa l o f the items c l a s s i f i e d wi th in t h i s group are demonstra-b l y not l o n g , or are long i n d i f fe rent senses: compare the " l e n g t h " o f ox-c a r t s , books, harrows, plows, fans and combs. Other predominantly long objects are not c l a s s i f i a b l e with lem at a l l ( e . g . bamboo p o l e s ) . A second reason for the inadequacy of " long t h i n g " as a meaning for t h i s c l a s s i f i e r i s the fact that a sense such-as " long t h i n g " i s the second most common shape c r i t e r i o n used i n a widespread group of c l a s s i f i e r languages, a f t e r a sense o f "round t h i n g " (see Adams & Conkl in 1973:3 ,5) . In such a case we would expect each o f the above c l a s s i f i e r s for d i f f e rent 7 ~ subsets o f long things to overlap the domain o f t h e i r superordinate , lem. Such overlapping i s extremely l i m i t e d , however: lem and khan overlap as c l a s s i f i e r for forks and spoons, daam appl ies only marg ina l ly to these ob-ject s and to knives . The use o f lem for s t a l k s , f irewood, ingot s , bars or r i b s , blossoms or lengths o f bamboo would be c a l l e d incorrec t or i r r e g u -l a r by ST speakers. Thus there i s no taxonomic r e l a t i o n here between c l a s -l U s i f i e r s , and ST lem contrasts on an equal l e v e l with the other "long t h i n g " c l a s s i f i e r s l i s t e d above, not as t h e i r superordinate. Clearly the range of a p p l i c a b i l i t y of ST lem i s r e s t r i c t e d to only c e r t a i n kinds of long things: the cutting instruments and the books, c a r t s , candles, fans, etc. The enqui-ry i s thus back to square one. The discussion so f a r has ra i s e d at le a s t two important questions: one i s the precise appearance of the items c l a s s i f i e d , and the other i s the sense or senses of "long" which might be relevant. For i l l u s t r a t i o n s of the items see Appendix A. For an e x p l i c i t system of conveying the senses of "long", " f l a t " and "round", i n c l a s s i f i e r systems I r e l y on Denny ( 1976a) who uses " s a l i e n t l y one dimensional" (SID), " s a l i e n t l y two dimensional" (S2D), and " s a l i e n t l y three dimensional" (S3D), respectively. This allows for the c l a s s i -f i c a t i o n of cubes, f o r instance, as "round thing", a not uncommon occurrence (cf. Thai c l a s s i f i c a t i o n of di c e ) . As " s a l i e n t l y one dimensional", objects are seen as being longer than they are wide or deep. In t h i s sense books might be long, e s p e c i a l l y t r a d i t i o n a l books (see Appendix A), and oxcarts might as well. But here we return to the problem that i f such a general notion of length i s the primary semantic moti-vation for t h i s c l a s s i f i c a t i o n , why aren't other c l a s s i f i e r s for "long things" more c o n s i s t e n t l y applicable to t h i s group of nouns, and why are native speakers inconsistent about t h i s meaning for lem and doubtful about i t s a p p l i c a t i o n to some of the items of- the group? Although there does seem to be a vague sense of "long thing" for t h i s c l a s -s i f i e r , to accept t h i s meaning seems to be obscuring several remaining problems and seems i n t u i t i v e l y unsatisfactory. I f the consideration of the i n -tension of the category name i s so d i s s a t i s f y i n g , perhaps a consideration of i t s extension w i l l be more useful. That i s , we w i l l now look at the range of things to which the term applies, rather than look for i t s "meaning". 2.2 LACK OF A COMMON CRITERION OF CLASSIFICATION It i s now well-known that many c l a s i f i c a t i o n s are not based on a common c r i t e r i o n o f class-membership; r a ther , the p r i n c i p l e might be a family resem-blance , a chain o f a s soc i a t ions , a metaphorical resemblance, or a r e l a t i v e degree o f membership. This appl ies to standard c las terms w i t h i n a language and to symbolic c l a s s i f i c a t i o n s of r i t u a l or mythic s i g n i f i c a n c e . Never-t h e l e s s , we cannot be sure o f our ana lys i s unless we can d e f i n i t e l y discount the presence o f a common c r i t e r i o n o f c l a s s i f i c a t i o n . This i s e s p e c i a l l y t rue i n t h i s study since the major i ty o f ST c l a s s i f i e r s , as s tated above, have assessed the lem group for such a common c r i t e r i o n among members and ._ the re su l t s are presented below i n Table 3. TABLE 3 POSSIBLE CRITERIA OF CLASSIFICATION AS lem OBJECT CUT, RAD. BOUND COMPRESS PARALLEL FLAT L O N G CYLINDR. PEGGED S O L I : PIERCE CENTER FOLD SERIES ROLLED k n i f e X (x) (x) X X X (x) (x) X bushknife X (x) (x) (x) X X (x) X r a z o r X (x) (x) X X X X axe X X X (x) X p i n X (x) X X X needle X (x) (x) X X X sword X. (x) X (x) X X X (x) X shove l X X X (x) X hoe X X X X s i c k l e X (x) X X (x) X l ance X (x) X X X plow X (x) (x) X X c h i s e l X (x) X X (x) X comb (x) X X X (x) (x) harrow X X x X X '.x candle (x) X X (x) X X X X incense (x) X X X X X Tha i book X X X X X (x) oxcart X X X (x) X X (x) s ledge (x) X X X X (x) fan (x) X X X (x) (x) paddle X X X (x) X N O T E : (x) means that the c r i t e r i o n may apply , but I f e e l less ce r t a in than I do i n cases marked x. "Rad. center" means " r a d i a t i n g center " , a common c l a s s i f i e r cate gory for the sun, moon and s t a r s , and other l i g h t sources. For i l l u s t r a t i o n s o f many o f these objects see Appendix A . In Table 3, we see again the wide a p p l i c a b i l i t y o f " long t h i n g " , found above to be l e s s than t o t a l l y s a t i s f a c t o r y . The reasons for s e l e c t i n g the other c r i t e r i a considered i n Table 3 w i l l become apparent i n the d i scuss ion below of c u l t u r a l value and f o l k technology. Beyond t h i s , Table 2 of fers no fresh ins ight s at a l l . An a d d i t i o n a l problem which ar i ses i n examining the extension o f the c l a s s i f i e r .lem, i s that some items are agreed upon by most speakers and some by only a few. Other items are l i s t e d i n d i c t i o n a r i e s or grammars but may now be obsolete as members- o f the category. In shor t , the lem-class has fuzzy edges. My s e l e c t i o n o f items i n Table 1 i s based p r i m a r i l y upon Haas's 1964 d i c t i o n a r y , but includes a d d i t i o n a l items based on i n d i v i d u a l informants ' c l a s s i f i c a t i o n . These items are " t r a d i t i o n a l book", "paddle" , " incense" and "plow" (see a l so footnote 2 ) . I f a category i s seen to have unclear boundaries i t i s more l i k e l y to be organized around a foca l member or members (c f . Chapter Three below). A cur-sory i n v e s t i g a t i o n o f the ST lem group again reveals very l i t t l e : speakers, when asked to give the best example o f the lem group, are u sua l ly puzzled by the request . Of those who do volunteer a best example, most suggest " k n i f e " or "book", but few are d e f i n i t e about t h e i r cho ice . One informant pointed out that although she can make judgments about the t y p i c a l i t y o f kind.s o f b i rds by mental ly comparing them to her concept of " b i r d i n genera l " , no such concept was a v a i l a b l e to her for lem. This impl ies that i t i s i n c o r r e c t , or at l ea s t mis lead ing , to consider the c l a s s i f i e r lem as a category l a b e l . To do so implies a taxonomic r e l a t i o n between the c l a s s i f i e r and (the names of) the objects c l a s s i f i e d . That i s , i t implies that an oxcart , say, i s a KIND OF lem. This i s c l e a r l y not the case. The c l a s s i f i e r seems to l a b e l a set of c h a r a c t e r i s t i c s o f the items c l a s i s f i e d , as impl ied i n the Thai grammati-c a l term for t h i s syntac t i c form-class : lagsananaam, l i t e r a l l y the "characte-r i s t i c of the noun". This f a c t . i s very important and has' wide t h e o r e t i c a l impl ica t ions discussed i n Chapter Three. It i s not c l e a r whether the fact that a c l a s s i f i e d item i s i t s e l f a c lass g term inf luences speakers' judgments about i t s t y p i c a l i t y o f membership i n a category. At any rate the few judgments I have been able to e l i c i t by d i rec t quest ioning have been i n c o n c l u s i v e . Thai scholars have on occasion addressed the problem o f de l inea t ion o f the complete set o f c l a s s i f i e r s i n ST, but no one has attempted a d e f i n i t i v e l i s t i n g (c f . Warotamasikkhadit 1970:23), and Haas (196k) remains the best ( i f r e s t r i c t e d ) source outs ide o f my own research. Dixon (1980:214-218) d e t a i l s how most c l a s s i f i e r systems are open-ended (unl ike noun c lass systems) and explains why i t i s i m p r a c t i c a l or at l eas t e s s e n t i a l l y a r b i t r a r y to define !.•*. c l a s s i f i e r s and thereby to l i s t them. There i s a. gradual merger with p l a i n nouns on the one hand and with measures on the other . In cases where no meaning or c l ea r c r i t e r i o n o f c l a s s i f i c a t i o n i s ava i l - , ab le , a l l that the lex icographer can do i s l i s t the items commonly c l a s s i f i e d . As for the case-of ST lem, - both Haas (1964:1*91) and Manitcharoen (19.64:852) opt to d iv ide the group into c u t t i n g implements and the heterogenous re s idue , and i n her a n a l y s i s , Hiranburana (1979) considers the d i f ference wide enough to pos i t two c l a s s i f i e r s , lem^ and lem^. Perhaps i t i s the case, then , that the lem group represents a combina-t i o n o f at l ea s t two other groups. This may be the re su l t o f a d i r ec t bor-rowing from another language or Ta i d i a l e c t , or the re su l t of r e s t r u c t u r i n g according to an evolv ing world view when an o r i g i n a l s ing le c r i t e r i o n no long-er app l i ed . C e r t a i n l y , i f the anomalous items (the oxcar t s , books, and per-haps candles) are removed, the remainder share a common feature of sharpness and c u t t i n g func t ion , with allowances for the k ind of c u t t i n g p e c u l i a r to fans, paddles , s c i s s o r s , combs and sledges. 18 In fact a c r i t e r i o n of "sharpness" or "sharp t h i n g " was quite prominent i n 9 a study I d id of ST speakers' own conscious ca tegor i za t ion of the 1 em-class nouns. The r e s u l t s are shown i n Table k: TABLE k ST SPEAKERS' CONSCIOUS CATEGORIZATION OF ST lem ITEMS IN A "FREE SORT" TASK SPEAKERS* OWN CRITERION NO. OF ITEMS IN CATEGORY (MAX 782) u t e n s i l s . khcygn, chaj .329 weapons. . a a w u d . . . . . .121 sharp th ing . • khcjpn, m i i khom .107 t o o l . . . Jkhrian, m i i • 97 education . s igsaa . . . . . 3h v e h i c l e . .phaahana . . . . . 31 a g r i c u l t u r e . .kaan kaseed . 29 w r i t i n g mater ia l s .khrian, khian . • 27 r e l i g i o n .saadsanaa . 25 man-made .manud pradid . • 23 movable • k h l i a n t h i i . 22 metal . l o o h a 9 . . . . . 13 sharp pointed . lieaem. . . . . . 12 na tura l . .thammachaad . 1 0 NOTE: Tota l s l e s s than 10 not inc luded. Table 3 shows that most of the i n d i v i d u a l items were c lassed by most of the i n -formants as " u t e n s i l s " ( l i t e r a l l y " t h i n g to use") . Among the items whose p r i -mary ca tegor i za t ion was NOT as " u t e n s i l " are the fo l lowing : TABLE 5 ST lem ITEMS NOT PRIMARILY CATEGORIZED AS "UTENSILS" IN TABLE k ITEM PRIMARY CATEGORY NO. OF CHOICES (MAX. 3k) book education 12 t r a d i t i o n a l book education 11 axe weapon 16 sword weapon 16 spear weapon 16 oxcart v e h i c l e 13 sledge v e h i c l e 13 Again we seem to have confirmed the expected and unexcept ional . The r e s u l t s i n Tables k and 5 do not appear to o f fer us any ins ight s in to the s o l u t i o n to the "puzz le " o f the lem-class . In fact the most s t r i k i n g th ing about the re su l t s here i s the p r e d i c t a b i l i t y : the student informants seem to c l a s s i f y these items much as Westerners would c l a s s i f y them. The importance of t h i s 19 p r e d i c t a b i l i t y i s discussed c a r e f u l l y below. For now l e t me merely emphasize that e s p e c i a l l y when informants make conscious ca tegor i za t ion judgments t h e i r groupings seem to us to be very l o g i c a l and even t r i v i a l . 2.3 SUMMARY OF THE PROBLEM At t h i s point our explorat ion of the problem has produced a ser ies o f c r o s s - c u t t i n g analyses : ( i ) my own subject ive assessment (Table. 3) ( i i ) the i m p l i c a t i o n o f a f oca l item or items (though none was immediate-l y apparent) ( i i i ) the l ex icographer ' s d i v i s i o n in to sharp things and an anomalous res idue ( iv) conscious ca tegor iza t ions by ST speakers which include "sharp t h i n g " , but do not a f ford i t prominence, and furthermore d i s t r i b u t e the ano-malous res idue across widely d i f f e rent categories (Tables 'k and 5) It i s c l ea r that t h i s type of i n v e s t i g a t i v e ana lys i s for i n t e r n a l s t ructure could be p r o l i f e r a t e d endles s ly , according to d i f f e ren t emphases and d i f f e ren t 10 p r i n c i p l e s . Note that the conscious r a t i o n a l judgments of ST speakers appear to be very s i m i l a r to those of any other modern educated speaker of any other language. Note a lso how Thai and American lexicographers can s p l i t the l i s t i n an i d e n t i c a l way. In t h i s sort of ana lys i s Thais seem to be at no p a r t i c u -advantage over non-Thais . I f we are to improve such an a n a l y s i s , i t w i l l have to s a t i s f y the fo l low-ing c r i t e r i a : ( i ) I f i t deals with contemporary ca tegor iza t ion i t w i l l have to be based upon speakers' judgments rather than non-speakers' analyses . ( i i ) I t w i l l have to depend upon p r i n c i p l e s other than simple l o g i c a l de-ductions based on inspec t ion of the l i s t o f nouns c l a s s i f i e d . ( i i i ) I t w i l l have to adopt an approach other than d i r e c t quest ioning of speakers' conscious knowledge. ( iv ) I t w i l l have to be h i s t o r i c a l l y and c u l t u r a l l y informed. 20 These requirements w i l l "be addressed i n subsequent chapters . For now, i t should be manifest that at l ea s t at a d e s c r i p t i v e l e v e l , there i s quite a prob-lem here indeed. Note a lso that merely i n sounding out the depths of the prob-lem we have a lready gone further than previous scholars , Thai or non-Thai . The d e s c r i p t i v e problem, at i t s most basic and concrete l e v e l , i s t h i s : why are oxcarts and books c l a s s i f i e d together with the other items of t h i s category? 21 NOTES TO CHAPTER TWO 1. ( p .11) T r a n s c r i p t i o n follows Haas ( 1 9 6 4 ) , but n=n,, E=ae, d.crf, and o=g. Glosses are approximate. 2. ( p .11) " incense" , "p low", "paddle" and some other items are i n fact seldom i f ever c l a s s i f i e d i n t h i s category i n modern ST. These items are inc luded here for comprehensiveness because o f l i s t i n g s i n some e a r l i e r d i c t i o n a r i e s or other reference works (e .g . McFarland 1944), and because they were accepted as members of the lem-class by at l ea s t some informants. The item p in ( " p i n , h a i r p i n " ) could a lso have been inc luded on the l i s t , as w e l l as samud "notebook"'. 3. ( p .11) See Haas (1942) for an exce l lent d e s c r i p t i o n of c l a s s i f i e r usage i n T h a i . 4. ( p .12) I am aware that "meaning" i s f a r more than something which l e x i -c a l items "have" i n some mechanical or absolute way. I am using the term here as a general summary term for t o t a l semantic fo rce , e i ther i n i s o l a t i o n or i n f u l l context . I use i t expressly to s i m p l i f y the presentat ion for an i n t e r d i s c i p l i n a r y readership . When necessary, d i f -vferent theor ies o f what i t i s w i l l be d i f f e r e n t i a t e d d e s c r i p t i v e l y , thereby avoiding the p r o l i f e r a t i o n o f jargon and r e d e f i n i t i o n . 5. ( p .12) A f te r Adams & Conklin (1973) c l a s s i f i e r glosses w i l l be given i n c a p i t a l s . 6. ( p .12) Genera l iza t ions about Thai c l a s s i f i e r s are based upon Haas ( 1 9 6 4 ) , as summarized i n Placzek ( 1 9 7 8 ) . Some modif icat ions to Haas's data have been made i n accordance with my own research. 7. ( p .13) I have inves t iga ted the problem o f overlapping c l a s s i f i e r s i n Placzek (1978) and ( 1 9 8 0 ) , and conclude only that some c l a s s i f i e r s are of wider a p p l i c a t i o n than others , and there i s genuine indeterminacy for some items for most; speakers. Furthermore there i s no cont rad i c t ion nece s s a r i ly impl ied by the use o f d i f fe rent c l a s s i f i e r s for the same noun. In most cases d i f fe rent c h a r a c t e r i s t i c s were be ing focused upon by each usage with a d i f fe rent c l a s s i f i e r . (see Haas 1942, Becker 1975, Erbaugh i n press , among o ther s ) . 7. ( p .17) Although " r a z o r " (miid koon) contains " k n i f e " (miid) morpholo-g i c a l l y , i t does not nece s sa r i ly do so semant ica l ly . That i s , there are degrees o f opaci ty i n compounds. The examples I present here , however, are i n my opinion a l l transparent cases. 9. ( p .18) Done i n 1982. Informants were 34 students i n f i r s t year L i n g u i s -t i c s at Thammasat U n i v e r s i t y , Bangkok. They were given a l i s t o f 108 items i n c l u d i n g the ST lem items and asked to f r e e - s o r t ; i . e . , d iv ide the items in to categories according to any c r i t e r i a they f e l t were r e l e -vant , and they were then to name the c r i t e r i a used. 10. ( p .19) In Placzek (1978) I a l so presented a t en ta t ive recons t ruct ion of a poss ib le semantic extension of the range o f nouns c l a s s i f i e d by lem. That system, I c l e a r l y s ta ted , was as subject ive as the approaches con-s idered above, and depended upon " s i c k l e " as a f o c a l i tem. 3.0 THEORETICAL POSITION 3.1 SOME BASIC ASSUMPTIONS The t h e o r e t i c a l o r i e n t a t i o n of t h i s d i s s e r t a t i o n i s i m p i r i c a l . It - i s concerned with questions o f the r e l a t i o n of p s y c h o l o g i c a l , l i n g u i s t i c , s o c i a l and c u l t u r a l facts to the " r e a l wor ld " which may be genera l ly con-s idered to be the world o f na tura l phenomena as perceived by a naive mem-ber o f a given c u l t u r e . Fol lowing Lyons (1977:440, 442) I w i l l adopt the "minimal and r e l a t i v e l y uncont rover s i a l " assumptions o f naive rea l i sm (summarized i n Placzek 1978:6, Rosch 1978: 41-46), the primary assumption being that a r e a l world can and does ex i s t as an independent inf luence upon mind and language. Lyons (1981:69) a l so a t t r ibu te s to r e a l i s t o r i e n -ta t ions (both naive and ph i lo soph ica l ) a b e l i e f i n the p o s s i b i l i t y o f NATURAL KINDS or groups o f r e a l world e n t i t i e s (and events) which share the same "essence". I can a l so subscribe to t h i s b e l i e f , on condi t ion that "essence" i s equatable wi th prototype-based generic concepts of the s p e c i f i c type I descr ibe i n t h i s chapter."'' Naive rea l i sm approximates a non-urban non-technica l viewpoint assumed to be dominant i n the e a r l i e s t stages o f language evolut ion and perhaps ac-q u i s i t i o n (c f . Keesing 1974:85 on the assumption of a " f a m i l i a r space" i n ear ly developmental s tages) . Concern for the r e a l world., then, e n t a i l s a naive observer who i s i n fact an abs t rac t ion o f a s o r t . "Of a s o r t " i s used advisedly because the naive observer ' s view (and h i s set o f meanings) i s t h e o r e t i c a l l y available'-as a s t a t i s t i c a l l y measurable c u l t u r a l consensus. To be sure, the task o f o p e r a t i o n a l l y de f in ing any c u l t u r a l norm i s not an easy one. In t h i s d i s -ser ta t ion I opt for word a s soc ia t ion as the most v i a b l e and v a l i d measure o f the culture-wide values o f l e x i c a l items. Assuming the naive observer viewpoint , explanation then en ta i l s t r ans -l a t i n g the p u z z l i n g aspects o f the cu l ture under study ( in t h i s case Thai • cu l ture) in to a form which i s consistent and p l a u s i b l e to a n o n - s p c i a l i s t member of the i n v e s t i g a t o r ' s own cu l ture ( in t h i s case the Canadian var iant of the l i t e r a t e Western European t r a d i t i o n ) . Trans l a t ion requires some k ind o f " e t i c " g r i d against which to compare the two systems being r e l a t e d . Here the o n t o l o g i c a l bas i s ( c f . Lyons 1977) or e t i c substructure i s i n part the assumption o f naive r e a l i s m . The other part o f the o n t o l o g i c a l sub-structure i s the r e g u l a r i t y o f d i s c o n t i n u i t i e s i n perceivable phenomena i n the r e a l wor ld . This aspect of the o n t o l o g i c a l basis i s discussed fur-ther below. The " p u z z l i n g " aspects o f a cu l ture here imply a c u l t u r a l uniqueness which i s assumed to p e r s i s t over time and across generations as a part of the language/culture her i t age . The example at hand i s the Thai c l a s s i f i e r category assoc ia ted wi th lem. Since oppos'ites define each other and ex-ceptions prove the r u l e , the study o f c u l t u r a l uniqueness i s a use fu l ap-proach to de f in ing those aspects which are considered to be common. I do not speak d i r e c t l y of univer sa l s since the only r e a l universa l s are based on s t a t i s t i c a l measures, or g loba l surveys, or are p h y s i o l o g i c a l . S t i l l the establishment o f more widespread patterns i s an e s s e n t i a l preparatory step before undertaking expensive and labor ious g loba l surveys . Though I r e l y here on a t h e o r e t i c a l view which i s best a t te s ted by psycholog ica l data, I do not consider t h i s to be a study o f the mind or of mental processes. Here I attempt only to UTILIZE psycho log ica l data along wi th other kinds o f data to describe and expla in a c u l t u r a l l y unique phe-nomenon. The aim i s to use the input of mental processes to study the f o r -3 mation o f a unique c u l t u r a l c h a r a c t e r i s t i c , a product of supra - ind iv idua l processes. Another major r e s t r i c t i o n i n t h i s study i s the focus on " t h i n g s " or "mater ia l c u l t u r e " ra ther than broad c u l t u r a l values or s o c i o l o g i c a l s i g -n i f i c a n c e . In part t h i s i s a r e su l t of t o p i c s e l e c t i o n ; that i s , ST lem deals e x c l u s i v e l y with a r t i f a c t s . But there are at l ea s t two other reasons for t h i s r e s t r i c t i o n . F i r s t , i t i s mani fes t ly more d i rec t to study human c l a s s i f i c a t i o n as i t appl ies t o concrete objects than as i t appl ies to ab-s t rac t or metaphysical ones s ince i t i s a common l i n g u i s t i c (perhaps cog-n i t i v e ) device to model expressions for abstract ions upon expressions for concrete things and r e l a t i o n s . Second, as i n the case o f a rchaeolog ica l study, i f we wish to attempt reconstruct ions of p r e - h i s t o r i c developments we must deal with the mater i a l cu l ture i f that i s a l l that remains, and learn to i n t e r p r e t the s i g n i f i c a n c e o f form and l o c a t i o n o f the a r t i f a c t s found. S i m i l a r l y , i f a l l we have as d i r ec t evidence for e a r l i e r periods o f the T a i language family are the surv iv ing daughter languages, we must l e a r n to i n t e r p r e t the s i g n i f i c a n c e o f forms and s tructure among these languages. Where words as names r e l a t e d i r e c t l y to " t h i n g s " we also gain ins ight s i n t o the mater i a l cu l ture o f e a r l i e r per iods . Deal ing wi th con-crete e n t i t i e s i s , i n the end, the most su i t ab le domain for work wi th in the naive r e a l i s t viewpoint mentioned above. F i t t i n g l y , an empi r i c i s t view i s c l o s e l y a l l i e d with a h i s t o r i c a l (developmental or d iachronic ) view. Although i t i s important to unders'tand r e l a t i o n s w i t h i n a system at a point i n time ( s t a t i c or synchronic view) complete understanding depends upon knowing the processes which produced the system at that point i n 'time, and which are s t i l l at work. Thus a h i s t o ^ r i c a l view impl ies a goal o f at l e a s t p r o b a b i l i s t i c p r e d i c t i o n of future r e -su l t s o f the ongoing processes. Note, however, that d iachronic study depends on as complete assessment as poss ib le of the current synchronic state and o f e a r l i e r synchronic states i f data i s a v a i l a b l e . 25 3.2 GRADING It i s by now a well-documented f i n d i n g that subjects overwhelm-i n g l y agree i n t h e i r judgments o f how good an example or c l e a r a case members are of a category even for categories about whose boun-daries they disagree (Rosch 1978:36). As a simple i l l u s t r a t i o n , grading i s the phenomenon whereby a r o b i n i s c l e a r l y a be t te r example of the category " b i r d " than i s , say, an owl , and an owl i s be t ter than a penguin. S i m i l a r l y , the sequence apple-pear-orange-peach-banana-apr icot-p ineapple-date-ra i s in probably accurate ly r e f l e c t s degrees o f t y p i c a l i t y i n the category " f r u i t " \ This f i n d i n g has been repeatedly con-firmed and i s " robus t " under var ious condit ions and i n s t r u c t i o n s , and for a wide range o f semantic domains(Mervis & Rosch 1981:95). Even researchers with d i f f e ren t t h e o r e t i c a l views have accepted t h i s fact (see Armstrong et aL,1983:5, Smith & Medin 1 9 8 l : l U U , Carey 1983:353 for examples). Furthermore, grading has been shown to a f fec t a l l ttie major i n d i c a t o r s used i n psycholog i -c a l research: speed o f process ing , free l i s t i n g , use i n na tura l and composed sentences, d i r e c t i o n a l i t y i n measures o f s i m i l a r i t y ( less t y p i c a l members are more s i m i l a r to more t y p i c a l members than v i c e ver sa ; thus " c h a i r " i s more l i k e " t a b l e " than " t a b l e " i s l i k e " c h a i r " ) , l e a rn ing of categories (most t y p i -c a l members learned f i r s t and fas tes t ) (Mervis & Rosch 1981:96-98, A n g l i n 1970, Rosch 1978:38). Rosch (1978:28) o u t l i n e s the basic p r i n c i p l e s o f c a tegor i za t ion which mo-t i v a t e the grading phenomenon: categories tend to form which provide the maxi-mum information with the l e a s t cogn i t ive e f fo r t . This i d e a l i s approached as the categories map rea l -wor ld d i s c o n t i n u i t i e s . Maximum information of a category i s provided by a combination of maximum shared a t t r i b u t e s among category members and minimum shared a t t r ibu te s with members o f other categories (Rosch & Mervis 1975). Rosch (.1978:37) i s not 26 sure whether t h i s tendency i s a d i r e c t r e s u l t of the way a t t r ibu te s tend to co-occur and c l u s t e r i n nature or whether i t i s a r e s u l t of the i n i t i a l stages of the ca tegor i za t ion process ; nevertheless i t i s c l e a r that EITHER the degree of membership or the degree o f d i s t i n c t i v e n e s s from other categories w i l l produce a grading ef fect i n ca tegor ie s , therefore i t i s not necessary to t r y to d i s t i n -guish these two factors (see Merv i s , C a t l i n & Rosch 1976, Mervis & Rosch 1 9 8 l : 99). Rosch goes on to emphasize that The fact that p r o t o t y p i c a l i t y i s r e l i a b l y ra ted and i s c o r r e l a t e d with category s t ructure does not have c l e a r impl ica t ions for p a r t i c u l a r pro-cess ing models nor for a theory o f cogni t ive representat ions o f catego-r i e s . . . (Rosch 1978:38). The fact of grading i n category membership can be represented by a v a r i e t y of formal models i n c l u d i n g p r o p o s i t i o n a l or image systems; the grading phenom-enon only CONSTRAINS such t h e o r e t i c a l models, without determining which type i s p re fe r rab le . Rosch (1978:28) s p e c i f i c a l l y states that grading has no i m p l i -cat ions for the development of categories i n c h i l d r e n , or the process which adults use to process novel items. E s p e c i a l l y i n view of the c h i l d a c q u i s i -t i o n data c i t e d above and that discussed below, . s h e may be over ly cau-t ious on t h i s po int . But I agree with Rosch that the one k ind of theory which "would probably be incapable o f handl ing a l l the present ly known f a c t s " (Rosch 1978:Ul) i s a theory of necessary and s u f f i c i e n t c r i t e r i a l a t t r i b u t e s for cate-gory membership, a type of theory popular at l e a s t from A r i s t o t e l i a n times down to very r e c e n t l y , when i t became unmistakably apparent that an a n a l y t i c decom-p o s i t i o n in to features i s an inadequate way o f approaching ca tegor i za t ion (see Armstrong et a l 1983:3, 2k; Smith & Medin 1 9 8 l ; Carey 1983; Hunn 1982; Needham 1975, among o ther s ) . Yet such a decomposition seems fundamental to our very sense of understanding and has been very d i f f i c u l t , i f not im-pos s ib l e , to abandon. 3.2.1. THE FOCI OF GRADED CATEGORIES: PROTOTYPES The work of B e r l i n & Kay (1969) on bas ic colour terms has been modi-f i e d severa l times to accommodate exceptions and new p h y s i o l o g i c a l data , but the o r i g i n a l not ion has withstood the t e s t o f time and c r i t i c i s m . This i s the not ion o f widespread uni formity i n the f o c a l re ferents of the f i r s t colour terms acquired i n language e v o l u t i o n . See Witkowski & Brown(l9?8) for a recent review. The c r u c i a l point here i s why there should be foca l items i n a cate-gory at a l l . Hunn (1976:513, 5l4) has given an explanation i n terms o f induct ive processes operat ing upon c lu s t e r s o f a t t r ibu te s and deductive processes operat ing upon s ing le prominent a t t r i b u t e s . To take the l a t t e r f i r s t : DEDUCTION generates categories by s t a t ing a de f in ing r u l e . It requires only one or at l ea s t very few a t t r ibute s that are o f wide a p p l i c a b i l i t y independent o f the current act of c a t e g o r i -za t ion . That i s , the a t t r ibu te s must have p r i o r existence and some sort o f broad a p p l i c a t i o n to a wide range o f e n t i t i e s . This i n turn implies that these a t t r ibu te s would themselves have some sort o f sa l ience and ab-s t r a c t i o n , probably bear ing a l i n g u i s t i c l a b e l . For example, an object i s categorized as a " red b a l l " i f and only i f i t i s perceived to be red i n colour and spher i ca l i n shape. The r u l e creates the category and depends on the prominent, wide ly-app l i cab le and already l abe l ed notions of redness and spher i ca l shape. This deductive process tends to create categories without grading. Thus i f i t i s red and a b a l l i t HAS to be a red b a l l . ' ' This i s abso lute ly true of a l l examples which q u a l i f y , abso lute ly true o f a l l which d o n ' t . It i s a uniform category with no foca l prototype and with c l e a r l y def ined l i m i t s . An important point i s that the r u l e generates the category, but the motivat ion for the r u l e i t s e l f i s based, I mainta in , :•> upon the PRIOR establishment o f induct ive categories UPON WHICH the r u l e 28 operates . Deductive ru les presuppose induct ive categories (see also Bybee and S lobin 1982:267-268 ) . v This i s a lso the bas is o f prototype-plus-rules models where the prototype i s a schema accompanied by ru les to r e l a t e the various poorer examples (Foss & Hakes 1978:162). ~~ ~~ For an account o f a t t r i b u t e c l u s t e r i n g i n INDUCTIVE processes , I f o l -low Hunn i n r e f e r r i n g to Bruner, Goodnow & Aust in (1956:47): . . . i n fact the de f in ing . fea tures of most objects and events are redundant with respect to each other . A b i r d has wings and a b i l l and feathers and c h a r a c t e r i s t i c l e g s . B u t . . . i f i t has wings and feathers , the b i l l and legs are h i g h l y p r e d i c t a b l e . In coding or ca tegor iz ing the environment, one b u i l d s up an expectancy o f a l l these features being present together . It i s t h i s un i ta ry concept t i o n that has the conf igura t iona l or Gestal t property o f "b i rd-nes s " . . .When the conception i s w e l l enough e s t ab l i shed , i t takes on the property o f being able to serve as a d i scr iminable and seemingly i r r e d u c i b l e a t t r i b u t e o f i t s own. Rather than merely operat ing upon the conf igura t ion o f a t t r i b u t e s , the induct ion process produces the conf igurat ion by abs tract ing^ across many examples. The conf igurat ion i s the prototype. That the prototype i s indeed an abs t rac t ion i s demonstrated by an ex-periment conducted by Franks and Bransford i n 1971. The experimenters de-v i s e d a " b a s i c " v i s u a l pattern and showed only systematic d i s t o r t i o n s of i t to subjects . In r e c a l l , "the item that the subjects were most c e r t a i n that they had seen was the prototype -an item that had never been presented" (Foss & Hakes 1 9 8 l : l 6 3 ) . These inves t iga tor s a l so performed a s i m i l a r experiment us ing only compo-nent sentences i n presentat ions . Subjects were most confident that they.had heard the composite sentence which incorporated these component sentences. Aga in , the composite was the only sentence which had NOT prev ious ly been pre-sented, a r e s u l t s i m i l a r to the r e s u l t s wi th the v i s u a l pattern (see Palermo 1978:161-164 for d e t a i l s ) . The main points to s tress at t h i s time are the basis i n percept ion of induct ive categories and the dependence o f deductive categories upon induct ive ones. Deductive categories are based on a s i n g l e (or few) wide ly-app l i cab le named a t t r i b u t e s , with a c l e a r ru l e of membership and c l e a r boundaries ( a l l - o r - n o t h i n g membership). They presuppose induct ive catego-r i e s . Inductive ca tegor ie s , i n cont ra s t , are based upon s i m i l a r i t y to a complex of a t t r ibu te s which tend to co-occur i n nature and which funct ion as s ing le genera l ized attributes such as "b i rd -nes s " , "bush-ness", e t c . This complex, or prototype , although i t i s e a s i l y u t i l i z e d by a l l speakers of a language, cannot be adequately descr ibed without recourse to examples or demonstration. Inductive ca tegor ie s , then , exh ib i t c l e a r grading and unclear boundaries . That i s , they gradual ly fade out into less and less t y p i c a l examples which share fewer and fewer o f the const i tuent a t t r ibu te s o f the prototype, as the category " f r u i t " fades out through b r e a d f r u i t , o l i v e s , coconuts and tomatoes. The d i f ference between induct ive and deductive processes and catego-r i e s i s not as c l e a r cut as portrayed i n the d i scuss ion above. In p a r t i c u -l a r there i s a process c a l l e d " a t t r i b u t e r e d u c t i o n " whereby a l i m i t e d • subset o f a t t r ibu te s (or perhaps a s ing le prominent a t t r ibu te ) may come to represent the abstract prototype (see Hunn 1976:513). The r e s u l t i s a deductive category although the., process may w e l l be i n d u c t i v e . A l l con-t r a s t s break down upon close i n s p e c t i o n . For the present , however, i t i s necessary to e s t ab l i sh the induct ive and deductive contrast and characte-r i z e the t y p i c a l r e s u l t s o f both processes. Since I have used the term " n a t u r a l " for induct ive categories I could use " a r t f i c i a l " for deductive ca tegor ie s , but s ince I d i f f e r s l i g h t l y wi th Hunn (1982 ) i n h i s use of t h i s l a b e l , I w i l l use the l a b e l " T e c h n i c a l " ins tead , c a p i t a l i z e d for formal sense. For reasons made c l e a r i n the. next s e c t i o n , " T e c h n i c a l " turns out to be a very appropriate l a b e l indeed. 30 3.2.1.1 PROTOTYPE AS IMAGE Before we come to a more d e t a i l e d cons idera t ion of the processes o f cate-gory-bu i ld ing upon prototypes and the unclear boundaries they generate, i t i s necessary to consider the not ion that the prototype at tr ibute-complez i s a- '•— mental image. Rosch h e r s e l f cautions (1978 :213-214) that imagery i s not a way o f avo id-ing complicated problems. S t i l l she does not r e j e c t the p o s s i b i l i t y and r e -fers to Kosslyn who reviews various p o s s i b i l i t i e s for formal models, conclud-ing that There does not seem to be anything mys t i ca l about the not ion that people have mental images, nor does It seem to r.equire "hocus pocus" i n order to t a l k about how images might funct ion i n the storage and process ing o f information (Kosslyn 1978:254). Whatever the process and nature of the image, I w i l l assume i t to be the v e h i c l e of prototype storage, whether as images of ear ly or common examples or of a composite, genera l ized representat ion o f the prototype i t s e l f . Some caut ions , however, accompany t h i s assumption. F i r s t , the prototype image i s not a s t r i c t l y v i s u a l or sensorimotor image, although v i s i o n does tend to dominate our other senses, and primary among v i s u a l a t t r ibu te s of ob-jects i s t h e i r shape ( M i l l e r & Johnson-Laird 1976:214, F r i e d r i c h 1970, E . Clark 1976, for examples). V i s u a l percept ion of shape, then , dominates the induct ive imagery which provides our p r o t o t y p i c a l f o c i o f n a t u r a l cate-gor ie s , and t h i s i s s t r i k i n g l y c l ea r i n the c r i t e r i a of noun c l a s s i f i e r categor ies . Nevertheless the prototype as image contains more than v i s u a l or sensorimotor input . It i s something akin to the "perceptual paradigm" of M i l l e r & Johnson-Laird (1976:214-215, 221, 224-225). This i s a f acu l ty which 31 . . .presumably emerges i n d u c t i v e l y from experience with many exam-plars whose l a b e l s were t o l d to us , o r , a f t e r a c e r t a i n l e v e l o f l i n g u i s t i c competence i s a t t a ined , from a more d i rec t verba l s p e c i -f i c a t i o n . o f the paradigm. M i l l e r & Johnson-Laird note that imaginary images can occur as w e l l as those from past percept ions . What I want to suggest here i s that the image must be one of the f u l l experience o f perce iv ing the example and l ea rn ing i t s l a b e l , i n c l u d i n g input from a l l the various perceptual f a c u l -t i e s and i n c l u d i n g t o t a l context of s i t u a t i o n and speech. A f f e c t i v e fac-tor s are a lso important. Nor does t h i s view necess i ta te a s ing le and s p e c i f i c moment o f l e a r n i n g : the image contains the f u l l context of l e a r n -ing i n a l l i t s occasions and on-going developments. The reasons for tak-ing t h i s p o s i t i o n w i l l become more apparent below i n the sect ion on c u l t u -r a l s a l i e n c e . 3.2.2. THE EXTENSION OF GRADED CATEGORIES: DEGREES OF SIMILARITY I have already described the fuzzy edges which r e s u l t with. Inductive ca te -g o r i z a t i o n . Mervis & Rosch (1981:101) report c l e a r evidence that there i s not only disagreement between i n d i v i d u a l s about boundaries (cf . B e r l i n & Kay 1969, Labov 1973, Kempton 1978, among o ther s ) , but there i s a lso disagreement between successive judgments by the same Ind iv idua l ( B e r l i n & Kay 1969, McClosky & Glucksberg 1978). I have found both kinds of Indecisiveness i n my own f i e l d -work on c l a s s i f i e r s . The ob jec t ion has been r a i s e d that the i n c l u s i o n of obvious ly a t y p i c a l items w i t h i n l i s t s o f stimulus-words to be ranked for t y p i c a l i t y (such as " foo t " i n a l i s t o f weapons) i s the r e a l source of grading and i t s fuzzy boundaries. However, Rosch has removed the problematic items and s t i l l demon-s t ra ted grading (see Armstrong et a l . 1983:26,27) , and McClosky & Glucksberg (1978) have demonstrated the extreme u n l i k e l i h o o d that each subject had a -clear and d i s t i n c t d e f i n i t i o n only at the time of t e s t i n g . C r u c i a l to the development of fuzzy-edged categories i s the mechanism o f making judgments o f s i m i l a r i t y to the prototype. Several researchers have con eluded that s i m i l a r i t y judgments are g loba l i n nature. For example Hunn (.1976 515) pos i t s that although people make judgments i n terms of percept ib le a t t r i -butes , the judgments are based on perceptions o f " o v e r a l l s i m i l a r i t y and d i f -ference" (see a l so Needham 1975:355, Brown 1977, 1979, B e r l i n 1973:261, e t c . ) . Fi l lenbaum & Rapoport ( 1 9 7 1 : 2 4 l ) found " l i t t l e d i f f e rence " whether t h e i r subjects used s i m i l a r i t y or d i s s i m i l a r i t y c r i t e r i a i n t h e i r judgments. They d i d f i n d that d i f f e ren t subclusters w i t h i n a semantic domain may be organized by d i f f e rent p r i n c i p l e s (e .g . taxonomic versus paradigmatic) and thus s i m i l a r -i t y based upon SPECIFIC a t t r i b u t e s , i f i t occurred at a l l , was probably l o c a l -i zed (see 'Tversky & Gat i 1978:97, F i l lenbaum & Rapoport 1971 : 2 4 4 ) . Theor i s t s i n b i o l o g i c a l c l a s s i f i c a t i o n have developed a c r i t e r i o n of category membership by complexes o f features (see Needham 1975:353). They spec i fy that i f there i s a l a rge number o f a t t r ibu te s invo lved , and each a t t r i b u t e occurs i n a l a rge number of examples, (given that there i s a l a rge number o f i tems) , then a l l the items w i l l be perceived as s i m i l a r . These views depend, of course, on an unspec i f i ed number as " l a r g e " . Judgments o f s i m i l a r i t y a l so depend' on which item i s the c r i t e r i o n (the s i m i l a r i t y o f (a) to (b) may not be the same as the s i m i l a r i t y o f (b) to (a) ) . See Tversky & Gat i (1978:97). Work by M i l l e r (reported i n Angl in 1977:7) has also produced a very important p r i n c i p l e i n s i m i l a r i t y judgments: he found that when adults do so r t ing ta sks , they group items by s e l e c t i v e l y ignor ing features which d i f f e r e n t i a t e items and at tending to features which are shared by items. This s e l e c t i v e ignorance i s an important, factor i n c u l t u r a l s a l i e n c e . In summary, " s i m i l a r i t y i s indeed r e l a t i v e and v a r i a b l e , but i t var ies i n a lawful manner" (Tversky & Gat i 1978:98). Most judgments of s i m i l a r i -t y are hased on o v e r a l l perceptions o f na tura l a t t r ibu te complexes. S imi-l a r i t y judgments based on' 'one or a few a t t r ibutes are u s u a l l y . l o c a l i z e d wi th in a semantic subdomain, or occur i n experimentally c o n t r o l l e d contexts with or thogona l ly-var iab le a t t r ibu te s (see Rosch 1975:196-197). The p r i n c i p l e of " s e l e c t i v e ignorance" i s an important one for i t s i m p l i -ca t ion that we categor ize on the basis o f s i m i l a r i t i e s rather than d i s s i m i l a r i -t i e s (although both i n t e r a c t ) and that we categorize i n an ac t ive way; that i s , our categories are not nece s s a r i ly perceptua l ly g iven. Further to t h i s , Tver-sky & G a t i (1978:98) point out that although we may categor ize on the bas is o f s i m i l a r i t i e s , our notions o f s i m i l a r i t y "are constant ly updated by experience to r e f l e c t our ever-changing p i c ture of the wor ld " . Thus even t h i s perceptu-a l ly -based psycho log ica l mechanism seems open to c u l t u r a l knowledge. Having reviewed the evidence for fuzzy categories (and thus for induct ive ca tegor iza t ion) we can turn to the evidence for c learly-bounded and w e l l -def ined deductive c a t e g o r i z a t i o n . 3.2.2.1. THE QUEST FOR A WELL-DEFINED CONCEPT Armstrong et a l . .(1983:273) argue that grading i n some categories i s not nece s sa r i ly evidence that these categories have a f u l l prototype orga-n i z a t i o n ( i . e . , are " n a t u r a l " categories by the above d e f i n i t i o n ) . They argue that to make the connection between grading and p r o t o t y p i c a l i t y , A necessary part of the proof requires f i n d i n g some categories that DO have d e f i n i t i o n a l d e s c r i p t i o n s , and showing as w e l l that subjects patent ly know and assent to these d e f i n i t i o n s ; and, f i n a l l y , showing these categories DO NOT y i e l d the graded outcomes. I argue below that t h i s cond i t ion does not nece s sa r i ly apply j r a t h e r , t h e out-come, should i t appear that we l l -de f ined categories are not to be found, would merely be a question of INTERPRETING the omnipresence of grading. In short that i s exact ly the outcome of t h e i r enquiry. They assume that we l l -de f ined concepts ex i s t ("are there d e f i n i t i o n a l concepts? Of 3b course" . (Armstrong et a l . 1983:274)'). Then they se lec t examples and i n a c a r e f u l l y c o n t r o l l e d ser ies of experiments f i n d confirmation of grading pheno-mena i n every task i n v o l v i n g t h e i r "we l l -de f ined" concepts, i n c l u d i n g cases where only students who had prev ious ly s tated that i t d i d n ' t "make sense" (Arm-strong et a l . 1983:286) to grade such items were se lec ted to do the grading. For these"researchers , committed to an assumption of we l l -de f ined concepts, these r e s u l t s requ i re a s p e c i a l e f fo r t of i n t e r p r e t a t i o n . Now the task of de-c i d i n g whether i t "makes sense" to judge objects for degrees of membership i n a given category i s an e n t i r e l y d i f f e r e n t th ing from a c t u a l l y judging those ob-j e c t s . Armstrong et a l . (1983:290) r e a l i z e t h i s and conclude that judging exemplar iness ' i s not the same as judging class membership. P lay ing t h e i r own d e v i l ' s advocates Armstrong et a l . provide an exce l lent explanation for t h e i r f indings i n the form of a "dua l theory" i n v o l v i n g on the one hand grading, the i d e n t i f i c a t i o n of ob jec t s , and prototypes , and on the other hand c la s s membership, a "conceptual core" and (presumably) necessary and s u f f i c i e n t a t t r i b u t e s . This theory i s roughly the equivalent o f a pro to type-p lus- ru le s accommodation o f grading and de f in i t i ons , and has been championed by such philosophers as Kripke and Putnam, and psychologists such as P a i -v i o . For our purposes here i t i s necessary to look a l i t t l e more c l o s e l y at the unsuccessful attempt of Armstrong et a l . to f i n d a we l l -de f ined con-cept . The i r choices for such concepts were odd and even numbers, plane geometric f igures and "female" . In a more d e t a i l e d account (Placzek 1983) I have shown how these con-cepts are not at a l l we l l -de f ined i n the fo lk d e f i n i t i o n s of most people, and have given more prec i se c r i t e r i a according to which the subjects of Armstrong .et a l . MUST HAVE graded the given examples of these concepts according to a prototype. Let us take the concept "even number" as an i l -35 l u s t r a t i o n . Armstrong et a l . (1983:27 *0 assumed that "even number" means "an i n -teger not d i v i s i b l e by two without remainder" (see a lso Carey 1983: fn h) but people o r d i n a r i l y ass ign a meaning to t h i s phrase i n terms of simple count ing , not computing: i t means every second number, s t a r t i n g with the number "2". C l e a r l y the p r o t o t y p i c a l example o f an even number i s "2". Thus people judge the "evenness" of numbers by the fo l lowing c r i t e r i a : ( i ) "c loseness " to 2 among s ing le d i g i t s ( i i ) presence of 2 or k ( i i i ) smallness o f the number ( iv ) presence of round numbers Cv) presence o f other even numbers On the bas is o f t h i s k ind o f c r i t i q u e I th ink there i s r e a l doubt about the existence o f we l l -de f ined categories IN NATURAL LANGUAGE. Here I must d i s t i n g u i s h between " w e l l - d e f i n e d " and " d i s t i n c t " s ince E . Hunn ( p . c . ) notes that p r o t o t y p i c a l categories based on decided gaps i n nature may be much more d e f i n i t e and prec i se than c l a s s i c a l categories based upon v a r i a b l e features . Considering the " w e l l - d e f i n e d " ca tegor ie s , then , note the p a r t i -cu lar type o f words Armstrong et a l . were forced to s e l e c t : t e c h n i c a l terms o f a f a i r l y simple and common nature , l e a v i n g as ide "female" the vagueness 7. of which I need not argue. Perhaps i f they had se lec ted more t e c h n i c a l terms ("cosine"? "parabola"?) they would have run the r i s k o f incomprehen-s ion or extreme var i ab le d e f i n i t i o n s among subjects . It was t h e i r s e l e c -t i o n o f s impler and common t e c h n i c a l terms which allowed the subjects to use t h e i r fo lk d e f i n i t i o n s and grade the examples. Apparently the bas ic numerals are n a t u r a l concepts probably ava i l ab l e to infants (L. Gleitman, p . c . ) . C l e a r l y t h i s i l l u s t r a t i o n i s t e l l i n g us that grading ind ica te s NATURAL concepts, and we l l -de f ined concepts ( "Technica l concepts" i n my usage) are confined to t e c h n i c a l language. Note the s o c i o l i n g u i s t i c impl ica t ions as w e l l , impl ica t ions for which concept i s more widespread, more common and undoubtedly e v o l u t i o n a r i l y and ontogenet ica l ly p r i o r . 3.2.2.2. THE QUEST FOR A CONCEPTUAL CORE Armstrong et a l . (1983: .291-295) explore the dual theory mentioned above and suggest that grading may be the r e su l t o f an i d e n t i f i c a t i o n 8 function used to make quick sorts of t h i n g s , scenes, and events i n the w o r l d " . Accompanying the i d e n t i f i c a t i o n function would be a (presumably feature-based) d e s c r i p t i o n o f the category to which each item belonged. Af ter M i l l e r , they c a l l t h i s mental d e s c r i p t i o n a "conceptual core " and 9 d i s t i n g u i s h i t from a s c i e n t i f i c or f ac tua l d e s c r i p t i o n . The example they o f fe r i s i n s t r u c t i v e : " g o l d " would have an i d e n t i f i c a t i o n funct ion such as "the ye l low g l i t t e r y s t u f f " (Armstrong et al .- 1983: 293), and a s c i e n t i f i c d e s c r i p t i o n : "at the present moment i n the h i s t o r y o f i n o r -ganic chemistry, atomic number such-and-such". In t h i s example, only an aphorism i s given as a conceptual core ( " a l l that g l i t t e r s i s not g o l d " ) . Perhaps Armstrong et a l . r e a l l y can supply a " c o r e " here between the i d e n t i f i c a t i o n rout ine and the s c i e n t i f i c d e f i n i t i o n , but I suggest that the aphorism occurred to them p r e c i s e l y because there i s NO core concept for t h i s i tem. The non- spec i a l i s t (the vast major i ty o f us) must r e l y upon h i s own sense-perception to i d e n t i f y go ld , or upon the word o f an expert . Between the g u l l i b l e and the goldsmith there i s no core at a l l . " ^ Rather than d i scard M i l l e r ' s not ion of a conceptual core , I would prefer to i d e n t i f y i t with the s c i e n t i f i c d e f i n i t i o n , with important empha-s i s on the v a r i a b l e range o f abs t rac t ion found i n competing theor ie s o f science wi th in speech communities and the. even greater range across cul tures It i s very i n s t r u c t i v e to examine M i l l e r ' s own notion of a "conceptual core" a l i t t l e more c l o s e l y . M i l l e r and Johnson-Laird (.1976:280) begin by i n d e n t i f y i n g the "conceptual core" with the more abstract concept which inte grates groups o f l e x i c a l concepts according to t h e i r r e l a t i v e pos i t ions i n a taxomomic h ierarchy or semantic domain. Thus the core seems to be equiva-lent to the superordinate term i n a taxonomy, as "animal " provides—ardeeper abstract organizat ion to the c o l l e c t i o n of r e l a t e d terms " b i r d " , " f i s h " , "snake", e tc . M i l l e r and Johnson-Laird a l so s tress that the s o - c a l l e d core i s an "abstract conceptua l iza t ion o f the w o r l d " , more s p e c i f i c a l l y , A conceptual core i s an organized representat ion of general knowledge and b e l i e f s about whatever objects or events the words denote the layman has some k ind o f theory - "prototheory" might be a be t ter term - that plays a r o l e i n h i s ord inary use of language that i s i n some sense analogous to the r o l e theor ies p lay i n s c i e n t i f i c d i scourse . The layman's theor ies frozen i n h i s vocabulary probably ex-press general points o f view that provide s t a r t i n g pos i t ions for the development of s c i e n t i f i c theor ie s . As science advances, the meanings o f words can change as a consequence o f increas ing know-ledge ( M i l l e r and Johnson-Laird 1976:290). This view i s supported by M i l l e r and Johnson-Laird with • language a c q u i s i -t i o n data which i s discussed below. It appears, then , that rather than a core , the concept proposed by M i l l e r can i n many cases be i d e n t i f i e d with a conscious, deduct ive , r u l e - • based d e f i n i t i o n which i s subject to the development o f the speaker's -know-ledge o f the wor ld , i . e . , h i s " s c i e n t i f i c " knowledge. Note the v a r i a b i l i -ty of t h i s " core " i n terms o f i t s development, dependence upon knowledge of the wor ld , i t s abstractness , and u l t i m a t e l y i t s degree o f s i m i l a r i t y or di f ference from the perception-based prototype. Thus I quite r e a d i l y accept t h i s va r i ab le n o t i o n , i d e n t i f y i t with a s c i e n t i f i c d e f i n i t i o n and i d e n t i f y both with my "Technica l category" d i s t i n c t i o n . The l a b e l " c o r e " , however, i s a complete 'misnomer. The p o s i t i o n so f a r , then , i s that the phenomenon o f grading i s ex-p la ined by the formation o f a p r o t o t y p i c a l image (of some type) u sua l ly based on a g e s t a l t - l i k e fusion of the a t t r ibu te s o f many perceived examples. This image, however i t i s formed and however i t functions i n the mental operations of language use, i s a record not only of a t t r ibutes but o f the f u l l contexts i n which the category l a b e l was learned . Natural categories form about the p r o t o t y p i c a l image by an e n t i r e l y induct ive process o f matching various items i n the environment against the prototype. This leads to unclear c r i t e r i a o f category membership and unclear boundaries but c l ea r judgments o f degree o f membership (contra Armstrong et a l . ) . On the assumption that we l l -de f ined ungradable concepts must e x i s t , we reviewed the search for same conducted by Armstrong et a l . and concluded that such concepts, i f they exis t at a l l , are confined to t e c h n i c a l language of vary ing degrees o f abs t rac t ion and d i s t inc tnes s from the p r o t o t y p i c a l or Natural category. The recourse of Armstrong et a l . to M i l l e r ' s not ion o f a "conceptual core" was seen to provide no core at a l l , but ra ther another species of Technica l d e f i n i t i o n . In t h i s sect ion we have discussed the phe-nomenon of grading as the under ly ing b a s i s , along with the p r o t o t y p i c a l focus, o f Natural ( i . e . p r o t o t y p i c a l ) ca tegor ies . Hereafter i t w i l l not be necessary to re fe r to grading s p e c i f i c a l l y , s ince i t i s understood to be a c h a r a c t e r i s t i c o f Natura l ca tegor ies . The general p i c ture i s now es tab l i shed o f a dichotomy between Natural categories based upon i n d u c t i v e , perceived s i m i l a r i t y to a prototype, and Technica l categories based upon deduct ively derived taxonomic h ierarch ie s (or "k ind o f " r e l a t i o n s ) . We can c a l l t h i s the "s tandard" prototype theory , 39 although to my knowledge no one else has made just t h i s synthesis of the f ac t s . L ike a l l dichotomies, with care fu l analys i s t h i s d i s t i n c t i o n w i l l be shown to be r e l a t i v e . Although Technica l categories presuppose Natural ca tegor ie s , thus accounting for the r e l a t i v e s t a b i l i t y o f Natural catego-r i e s and the extreme v a r i a b i l i t y o f Technica l ones, . apparently d i s t i n c -t i v e features may a l so be na tura l while. Hunn (1982) demonstrates the a r t i -f i c i a l nature o f some Natural ca tegor ies . 3.2.3. APPLICATIONS TO THE PROBLEM. We have noted that informants appear i n d e c i s i v e i n suggesting "book" or " k n i f e " as best examples o f the lem group. The above d i scuss ion implies s t rongly that ALL categories i n na tura l language are founded on prototypes . Lyons ( l 9 8 l : T 3 ) concurs . The hypothesis i s therefore that the lem group does too . The bas i c question i s not whether there i s a prototype for the lem group; r a ther , i t i s which members are c loses t to the prototype. The one reservat ion concerns whether the lem group i s the product of an ob l iga tory syntac t i c ru l e rather than being a s tra ightforward generic name. For the moment I w i l l deal with i t as i f i n fact i t were a genuine but poor ly- focus-ed category. We have a lso noted above the fuzzy q u a l i t y o f the lem c lass boundary. This implies that as a category i t i s more a Natural one than a Technica l one. On the other hand the r e s t r i c t i o n o f the lem group to only man-made a r t i f a c t s implies a Technica l category. However, i f lem refers to or implies any neces-sary and d i s t i n c t i v e features , they are not apparent by simple i n s p e c t i o n , the nature and i n t e r n a l r e l a t ionsh ip s among such features being the object o f en-,quiry of t h i s d i s s e r t a t i o n . We d i d note above that judgments of s i m i l a r i t y are a f fected p o s i t i v e l y where a large number of a t t r ibutes are common to a group of e n t i t i e s . With a t t r ibu te s of sharpness, l eng th , f la tness and b i n d i n g , or f o l d i n g , lem may be considered to have a " large number" of a t t r i b u t e s , and ho such a foundation i n m u l t i p l e a t t r ibute s co-occurr ing i n the r e a l world again implies a N a t u r a l , rather than a Technica l category. It i s a reasonably safe assumption that lem represents a N a t u r a l , not a Technica l category, p r i m a r i l y because, as described above, a Technica l category i s imposed upon percepts by the deductive a p p l i c a t i o n o f necessary and s u f f i c i e n t features of a wide ly-app l i cab le k i n d . In such a case the features would normally be recoverable in a deductive analys i s o f the grouping, and I have shown above ( e spec i a l ly i n I -u Chapter Two) how t h i s i s not the case: the deductive features ( i f any) are not c l e a r l y apparent. There i s always the p o s s i b i l i t y that i n a h i s t o r i c a l l y e a r l i e r stage of ST or i t s proto-languages, lem d i d apply to nouns according to Technica l c r i -t e r i a , but i n accord with, the above ontogenetic arguments p o s i t i n g Natural before Technica l ca tegor ie s , the perceptual nature and the complex family o f a t t r ibutes which, seem to motivate the lem grouping a l l imply an o r i g i n as a Natural category. Nevertheless , the r e s t r i c t i o n o f ST lem to a r t i f a c t s r e -mains, a nagging r e s i d u a l f ac t . At t h i s point the best approximation as to the status of lem as a cate-gory i s tha t , l i k e other c l a s s i f i e r s i t l abe l s an a t t r i b u t e or a t t r i b u t e c l u s -t e r prominent i n the object l abe l ed by the head-noun. This a t t r i b u t e i s used to d i s t ingui sh , and i d e n t i f y tokens or examples o f the k ind of th ing the noun names, i n genera l . Such, at tr ibutes , are u sua l ly shape-based. For h i s t o r i c a l and other reasons the o r i g i n a l prototype may have been obscured, r e s u l t i n g i n the present oddly skewed range of a p p l i c a t i o n of t h i s c l a s s i f i e r . A com-parison with other common (and glossable) c l a s s i f i e r s , i s most en l ighten ing . Some basic data i s presented i n Table 6. The important points to not i ce in Table 6 are ( i ) the r e l a t i v e uniqueness o f khon "person" ( i i ) the l i m i t e d d i s t r i b u t i o n o f Generic c r i t e r i a r e l a t i v e to Shape or other perceptual c r i t e r i a (..iii.) the l i m i t e d noun function o f most o f these c l a s s i f i e r s (only luug' " o f f s p r i n g " , khan "paddy dyke", and tua "body" occur i n t h i s s e lec-TABLE 6 CRITERIA OF CLASSIFICATION AND NOMINAL SENSES OF THE MORE COMMON ST CLASSIFIERS FOR CONCRETE OBJECTS CRITERIA OF CLASSIFICATION •• - khon tua ton luug baj lem an khan klQgn, lam sen phaen BY-SHAPE limbed SID S3D S2D sharp SID? SID c y l i n d e r SID SID S2D (primary) body point BY SHAPE head & v e r t i c a l round f l e x i - long small long . l o n g . c y l i n d e r f l e x i - t h i c k (secondary) t a i l b le sharp-edged long? hollow- long hollow b le r i g i d small GENERIC people animals plants f r u i t con- books? count- land opt i c a l conduits - -(x. r e - (x. e l - (x.very ta iners tool s? ables vehic le s i n s t r u - boat s vered) ephants ) small) (x.kwian, • ments? planes l i a n , Sci.)-smoking pipes? SEMANTIC VALUE IN < COMPOUNDS GENERIC people animate p lant ; or ig in , , source of f-spring l e a f , blade :.bar, pole conduit s t r i n g l i n e SHAPE- body- - S3D S2D - - SID c y l i n d e r c y l i n d e r SID S2D BASED shape round small long long hollow f l e x i -b le t h i c k r i g i d small SEMANTIC VALUE AS INDEPENDENT NOUN (NECESSARILY GENERIC) people body trunk, s ta lk of f-spring — — — handle, bar c y l i n d e r , o p t i c a l — l i n e — pole , i n s t r u -dyke ment? smoking pipe?  NOTE: x. means " e x c e p t ^ f p r V , and a dash (-) means " i n a p p l i c a b l e " . In severa l cases i t i s extremely d i f f i c u l t to d i s t ingu i sh sense as shape and generic sense, as discussed below. Question marks mean e i ther I am unsure of the c r i t e r i a or sense, or that a generic sense i s i n c i p i e n t . SID, S2D and S3D are explained i n Figure k, p.89 Here "gener ic " i s used i n a general sense o f "natura l k i n d " , or "essence" , ra ther than i n a s t r i c t taxonom-icsense. Generic c r i t e r i a o f c l a s s i f i c a t i o n contrast with Perceptual c r i t e r i a i n the example of the ST c l a s s i f i -er t u a : " c l a s s i f i e r for animals" (Generic) versus "body-shape" (Perceptual ) . B a s i c a l l y the d i f ference i s that Generic c r i t e r i a may include perceptual and non-perceptual c r i t e r i a to greater or l e s se r extents , w i l e Percep-t u a l c r i t e r i a are r e s t r i c t e d ( in t h i s data) p r imar i ly to Shape. 1+2 tion.) Only khon has the same semantic value as c l a s s i f i e r and noun. (iv) the close s i m i l a r i t y between each Generic c r i t e r i o n and i t s respec-t i v e perceptual base, (v) the p r o t o t y p i c a l nature of the Generic c r i t e r i a (e.g. f r u i t as proto-t y p i c a l S3D things, containers as p r o t o t y p i c a l S2D things, etc.) (vi) the close p a r a l l e l i n c r i t e r i a of c l a s s i f i c a t i o n regardless of func-t i o n as c l a s s i f i e r or class term The c l a s s i f i e r khon i s unique s y n t a c t i c a l l y as the only ST c l a s s i f i e r which.is a F u l l Repeater, P a r t i a l Repeater and f u l l noun. I t i s shown i n Table 6 to be semantically anomalous as well: i t i s the only c l a s s i f i e r for concrete objects without a shape c r i t e r i o n , except for the most general c l a s s i f i e r an, which, may have been the o r i g i n a l SID c l a s s i f i e r i n t h i s system (see below). In every case except that of khon, wherever Generic c r i t e r i a occur there i s a c l o s e l y r e l a t e d Shape c r i t e r i o n . In some cases there i s a Shape c r i t e r i o n without a Generic counterpart, and i n others (including lem) there appear to be i n c i p i e n t Generic groupings. In these i n c i p i e n t Generic cases, for example klpon, " c y l i n d e r " developing a sense of " a l l o p t i c a l instruments" for urban speakers, but "(tobacco) smoking pipes" for r u r a l speakers, the p o t e n t i a l Generic c r i t e r i o n is. ' s t i l l c l e a r l y bound to the Perceptual c r i t e r i o n . Whether we look at sense as nounor at c r i t e r i o n of c l a s s i f i c a t i o n , i t i s c l e a r that among these words r e f e r r i n g to concrete objects ( d i r e c t l y as nouns, i n d i r e c t l y as c l a s s i f i e r s ) the perceptual c r i t e r i o n i s more widespread and provides the semantic basis from which the Generic values could e a s i l y have developed. In fact the picture presented by Table 6 c a l l s into question the necessity for p o s i t i n g a genuine dichotomy of Generic versus Perceptual . . . 11 c r i t e r i a . i n n o case except that of the c l e a r l y nominal anomaly khon, can Generic c r i t e r i a be s a i d to have exceeded the parameters set by the basic,, presumably o r i g i n a l Shape c r i t e r i a . As an example, i f for the c l a s s i f i e r tua we accept an extended sense of "body-shaped" as including not only prototypi-c a l four-legged animals, but also the common limbless animals f i s h and snake, we f i n d we have covered' almost a l l e n t i t i e s we are w i l l i n g to c a l l animals by a Generic c r i t e r i o n . Perhaps the extension o f the c l a s s i f i e r to microsco-pic creatures marks the f i r s t t r u l y Generic case i n t h i s l i n e of extension, but even t h i s extreme case i s not abso lute ly d i s t i n c t . The other l i n e o f extension i n the range o f a p p l i c a b i l i t y of tua seems to be based upon a c r i t e r i o n o f s e l f - m o t i v a t i o n , both l i t e r a l l y and f i g u r a -t i v e l y . At leas t the l i t e r a l l y ac t ive e n t i t i e s are arguably motivated by a v i s i b l e c r i t e r i o n of percept ib le motion. The f i g u r a t i v e extensions ( e . g . to lead ing ac tor s , chemical transformers , e t c . ) are r e s t r i c t e d to use of tua as c lass term (head of a compound noun). Thus the c l a s s i f i e r funct ion i t s e l f remains c l o s e l y r e l a t e d to the perceptual foundation. Again there i s a gray area and a d e f i n i t e l y Generic a p p l i c a t i o n i s most c l e a r l y apparent i n cases of modern technology or other h i g h l y t e c h n i c a l contexts . The case o f ton as c l a s s i f i e r for a l l plants i s n o t . f a r extended from the sense o f " r o o t , o r i g i n , source" c l o s e l y t i e d to the SID v e r t i c a l prototype. Again , almost every p l a n t , to be a p l a n t , must have some sort o f stem or source r i s i n g v e r t i c a l l y from w i t h i n the ground. This i s t rue even of a creeper. As for the c l a s s i f i e r luug , I present arguments below which show that t h i s word i s a l so anomalous, probably a noun imported in to the c l a s s i f i e r system to replace a former c l a s s i f i e r for f r u i t . Even i n such a case, the l i m i t a t i o n of the Generic c r i t e r i o n o f luug to a l l f r u i t except small s i z e d f r u i t such as grapes, b e r r i e s , e t c . , shows that the shape c r i t e r i o n o f S3D has not been overridden at a l l . The c l a s s i f i e r baj has a Generic c r i t e r i o n o f a p p l i c a t i o n to a l l con-ta iners regardless o f shape. It i s a fact that the not ion o f container i s i n e x t r i c a b l e from a sense o f t h i n f l a t mater i a l despite the wide range o f shapes , . s ty le s and mater ia ls for conta iners . Thus t h i s c l a s s i f i e r ' s Generic c r i t e r i o n too i s l i m i t e d by. the Shape c r i t e r i o n from whence i t emerged. kk The c l a s s i f i e r lam appl ies to a l l boats and a i rp l ane s , and again the shape s i m i l a r i t y i s cons i s tent . This c l a s s i f i e r may apply to boats or a i r -planes which are NOT long and c y l i n d r i c a l (or h a l f - c y l i n d r i c a l ) , but such ve-h i c l e s are s t i l l so anomalous that I cannot offhand t h i n k o f a s ing le example. Perhaps the c l a s s i f i e r khan " long handle; paddy dyke" has achieved a c l ea re r severance of Generic c r i t e r i o n from Shape c r i t e r i o n . It probably appl ied to non-native l and veh ic le s which had a tongue or two long handles protruding from the front for h i t c h i n g draught animals or for people to g r ip (as i n the case o f palanquins and r i c k s h a s ) . With the predominance o f the automobile i n modern t imes , the " long handle" SID connection i s a l l but l o s t . This case demonstrates the importance o f t echno log i ca l change and c u l t u r a l s a l i e n c e . Presumably the c r u c i a l stage was incept ion of the use of the horse-le s s carr iage whi le r e t a i n i n g the c l a s s i f i e r for horse-(or human-) drawn car-riages . The case o f lem seems s i m i l a r , to the extent that lem can be considered c l a s s i f i e r for books or t o o l s regardless o f shape. The case i s l e s s c l e a r with lem, however, probably because t e c h n o l o g i c a l change, as documented i n Chapter S i x , has been l e s s r a d i c a l than i n the case of the veh ic le s c l a s s i f i -ed by khan. That i s , books are s t i l l predominantly v i s u a l l y f l a t and contain b l a d e - l i k e pages. Even a c r i t e r i o n of length i s r e s i d u a l l y present i n that the p r o t o t y p i c a l modern book, which, i s now almost un iver sa l i n Tha i l and , i s longer (along the spine) than i t i s wide. These perceptual c r i t e r i a (SID, sharpness, and secondary :S2D, plus a probable f ac tor o f binding) are argued> below (Chapter Seven) to be the o r i g i n a l Shape c r i t e r i a for lem. The re su l t i s that books are too uniform i n shape over t ime, and too c lose i n shape to the ear ly books for an independent Generic c r i t e r i o n to have emerged (the only major change i s loss o f the SID c r i t e r i o n ) . S imi l a r arguments apply to the p o t e n t i a l Generic c r i t e r i o n of " c l a s s i f i e r for a l l cu t t ing instruments regardless o f shape". At t h i s point we may re turn to the question o f whether ST lem represents a genuine or p s y c h o l o g i c a l l y r e a l category or not . From Table 6 we have seen that there appears to be a process at work -'whereby Generic c r i t e r i a of c l a s -s i f i c a t i o n develop from Perceptual c r i t e r i a , p r i m a r i l y shape. I draw t h i s conclusion because o f the r e g u l a r i t y o f Perceptual c r i t e r i a - i n every case- in Table 6 except the case of khon, which i s demonstrably anomalous i n almost every way. Generic c r i t e r i a appear i n only some cases, and i n each case they are c l o s e l y associated with the perceptual foundation. In some cases such as with tua , l u u g , khan, klgon,, e t c . , we can see the developmental process at work. I suggest that the Generic c r i t e r i a do i n fact cons t i tu te true cate-gories but the Perceptual c r i t e r i a do not . They merely l a b e l , as asserted above, an a t t r i b u t e o f the noun. To state that Perceptual c r i t e r i a do not cons t i tu te t rue categories r e -quires some q u a l i f i c a t i o n , however. They do not l a b e l .a generic concept as nouns do; that i s , not l e x i c a l l y . L e x i c a l l y they l a b e l a t t r i b u t e s . These a t t r ibu te s i n turn are c h a r a c t e r i s t i c of a v a r i e t y o f objects or e n t i t i e s i n the environment, and s ince there i s a common fac tor among these objects or e n t i t i e s they cons t i tu te a category at l ea s t conceptually., ( th i s pragmatic or rea l -wor ld common fac tor i s p a r a l l e l e d by the syntact ic common factor of co-occurrence with the same c l a s s i f i e r ) . Rosch et a l . (1976:383) provide a minimal d e f i n i t i o n o f a category as any group of items t rea ted i n a s i m i l a r way. In t h i s secondary, ana ly t i c -deduct ive way, perceptual c r i t e r i a can be sa id to give r i s e to ca tegor ies , but l i k e the example above of the category of a l l red t h i n g s , the categories are not relevant to r e a l language use, and only occur when we stop to th ink about i t (analyse i t ) and DEDUCE the common fac tor : kn ives , books and fans a l l take the same c l a s s i f i e r , therefore they must l o g i c a l l y belong to the same category. 1+6 In sum, Perceptual c r i t e r i a merely l a b e l a t t r ibute s such as S I D , S2D, e t c . , for a s p e c i a l purpose i n t h e i r l e x i c a l funct ion . Outside o f that func-t i o n , i n the r e a l wor ld , we observe that they apply to a range o f objects (which are long t h i n g s ) . This i s a pragmatic matter which gives r i s e to the m e t a - l i n g u i s t i c deduct ive-ana lyt ic concept o f a c l a s s i f i e r ' s range of ap-p l i c a t i o n , i t s category. This gives r i s e to notions l i k e " c l a s s i f i e r for long t h i n g s " . In contra s t , the development of a Generic c r i t e r i o n for a c l a s -s i f i e r . may be seen as the incorpora t ion o f the deductive category into the l e x i c a l func t ion . But i n every case the generic c r i t e r i o n i s a subset of the perceptual ly-based group, presinmably se lec ted by some form of c u l t u r a l s a l i -ence. For example S2D, a Perceptual c r i t e r i o n , has given r i s e to the Generic c r i t e r i o n "containers i n ge