The methodology of field investigations in linguistics: (Setting up the problem) [Reprint 2017 ed.] 9783111351452, 9789027930767

Citation preview

JANUA LINGUARUM STUDIA

MEMORIAE

N I C O L A I VAN W I J K D E D I C A T A edenda curat

C. H. VAN SCHOONEVELD Indiana

University

Series Minor,

142

THE METHODOLOGY OF FIELD INVESTIGATIONS IN LINGUISTICS (Setting up the Problem)

by

A. E. KIBRIK

1977 MOUTON THE HAGUE - PARIS

© Copyright 1977 Mouton & Co. B.V., Publishers, The Hague No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.

ISBN 9 0 279 3 0 7 6 7

Translated f r o m Russian. Original title:

Metodika polevyx issledovanij

Printed in the Netherlands

FOREWORD

This study 1* is devoted to one of the most traditional and at the same time one of the most neglected branches of linguistics-field linguistics. In field linguistics the linguist is in the position to describe a language which, at least to begin with, he does not speak and to observe the speech of native speakers in what is usually a natural speech environment. The tradition of field investigations is especially strong in American linguistics; practically all the theoreticians of American linguistics have studied l e s s e r ( o r as they a r e sometimes unjustifiably called "exotic" o r "primitive") languages: F. Boa, L. Bloomfield, E . S a p i r , B. L.Whorf, G. T r a g e r , B. Bloch, M. Swadesh, K.Pike, E.Nida, H. Gleason, J. Greenberg. Many Russian linguists have also made contributions to_field linguistics: M. A. Castr6n, P. K. Uslar, V. G. Bogoraz, L. Ja. Sternberg, J . Baudouin-de-Courtenay, N. A.^Jakovlev, E. D. Polivanov, A. N. Genko, 1.1. Zarubin, L. I. Zirkov, L. V. Scerba, P. S. Kuznecov and others. In fact practically all investigations of l e s s e r unwritten languages and dialects as well a r e c a r r i e d out under field conditions. On the basis of field investigations in the 1920's and 30's alphabets w e r e created f o r m o r e than forty nationalities of the Northern USSR, Central Asia, and the Caucases and voluminous g r a m m a r s a r e being written for the unwritten and newly created literary languages of Soviet nationalities. In a d dition there have been a number of detailed examinations examining p a r ticular questions of insufficiently studied languages. The scope of dialectological investigations, especially of Russian, but of other languages a s well, is broad. The comparative-historical and typological study of languages without a written tradition is also carried out with the help of field work. In spite of the large amount of accumulated experience in practical field work there is really nothing in the Soviet literature devoted to the technique of field investigation. Likewise almost no general works of this kind have been written abroad. The present work was to a certain extent written to help fill this gap. But the author by no means claims to p r e s e n t a complete and comprehensive description of all problems of field linguistics. This is impossible due to the volume of work and the insufficiency of the author's experience. This work is basically a generalizing survey of the author's personal observations made during several linguistic expeditions on languages spoken in the USSR: Lak (1967), Archi (1968 , 1971), Shughni (1969), Xinalug (1970, 1971) Koryak (1971). The aim of the paper is f i r s t to define the place of field linguistics in the general linguistic pattern, after it has been shown that field linguistics is closely related with many other disciplines and to what extent it needs the help of these other disciplines and to what extent the results of field

vi investigations are important for general theoretical investigations (Chapter 1); secondly, to sketch the contours of the problems which arise before the field investigation has been started and to show, in a fragmentary manner, how they may be solved (Chapter 2); finally, to describe the components of field work in itself; its basic steps, interaction with native speakers, and the collection of linguistic facts (Chapter 3). Certain complicated and undoubtedly important questions which have do with the final description of the language based on the results of the field investigations were deliberately left out, since these questions go beyond the bounds of field work itself since they define the general linguistic world view of the investigator. I would like to take this opportunity to thank the heads of the Faculty of Philology of Moscow State University and the Departments of Structural and Applied Linguistics who sponsored the field investigations in which I took part; and to all my numerous colleagues on these expeditions- co-workers and students of the Departments of Structural and Applied Linguistics of the Faculty of Philology at Moscow State University, without whose creative assistance the present work would have been impossible; and also to all the informants with whom we worked in gaining a practical understanding of the nature and difficulty of field work. I wish also to express my gratitude to all those who have read the manuscript and have helped me to eliminate a number of shortcomings, in particular B. Ju. Gorodeckij, V. A. Zvegincev, V. V. Ivanov, S. D. Kacnel'son, G. A. Klimov, S. V. Kodzasov, and V. I. Cincius. NOTE 1* "This work originally appeared as Monograph 10, Publications of the Department of Structural and Applied Linguistics, Moscow State University, 1972 (under the general direction of V. A. Zvegincev). "

TABLE OF CONTENTS

Foreword 1. The role of field linguistics in theoretical and applied linguistics 1. The problem of investigation of insufficiently studied languages 2. Field linguistics as a branch of descriptive linguistics . 3. Methods of field linguistics and its relation to other linguistic disciplines

V 1 1 2 5

2. Preliminary stage ( "zero cycle") in field work 1. The necessity for questionnaires and tests 2. Phonetics 2.1. Conditions on Universal Phonetic Transcription . 2. 2. Phonetic transcription 2. 3. Use of transcription 3. Morphology 3.1. A method of discovery of morphological entities . 3.2. A sample of metalinguistic calculus of grammatical meanings (for the case of spatial meanings) . 3. 3. A sample of questionnaires of diagnostical utterances in the mediator language (meanings in noun phrases, except spatial meanings) 4. Syntax 5. Lexicon

19 19 20 21 25 27 28 28

3. Organization of field work 1. Basic stages of field work 1.1. Ultimate aim 1.2. First encounter with a language 1. 3. Discovery of basic grammatical categories . . . 1.4. Interrogation according to the following system: hypothesis^—data^—hypothesis^—data^—..

45 45 45 47 50

— e l i c i t e d knowledge 1. 5. On practical learning of the language 2. Work with an informant 2.1. The "human factor" in field work 2. 2. Eliciting of linguistic data 3. Collection of linguistic data 3.1. Corpus of sentences 3. 2. Corpus of texts

30 31 32 35

51 52 53 53 56 59 59 61

viii 3. 3. Corpus of lexemes 3 . 4 . Paradigms Conclusion

63 66 71

Appendix 1: Universal Phonetic Transcription based on Latin (variant of the IPA) 73 Appendix 2: A sample of questionnaires in the form of acalculus (spatial grammatical meanings) 81 Appendix 3: A sample of questionnaires in the form of a set of diagnostic utterances (non-spatial meaning in a noun phrase) 89 Appendix 4: An approximate list of sentences for the first encounter with the lexicon and grammar of the t a r get language 97 Appendix 5: Minimum dictionary (an approximate list of words (senses) for the collection of vocabulary . . . . 101 Appendix 6: A sample of grammatical classification of sentences 125

1

THE BOLE OF FIELD LINGUISTICS IN THEORETICAL AND APPLIED LINGUISTICS

1. The Problem of Investigation of Insufficiently Studied Languages. The number of languages spoken in the world is fairly great. There a r e no exact figures but one finds approcimations of between 2. 5 and 7 thousand 1* in the literature (within the boundaries of the USSR alone m o r e than 100 have been recorded 2*). However the number of languages spoken by more then a million persons is only around 150 3*, and half the e a r t h ' s population speaks one of the eight most common languages 4*, of which three 5* a r e the dominant languages f o r m o r e than half of humanity. To these languages are opposed the f a r m o r e impressive class of languages which serve small groups of people. But the m a j o r efforts of linguists have been directed toward the study of languages which have historical, political, and social prestige; not only that, all general linguistic conceptions a r e as a rule based on facts gathered from this handful of privileged languages ( m o s t of whom belong to the Indo-European language family), and the sea of linguistic facts abounding in the r e m a i n der of the world's languages has either gone entirely unrecorded or f o r all practical purposes has not been taken into account in the theory of language. The state of the study of these languages is such that they a r e fortunate to have had even a single investigator. It is not without interest to note that within the USSR alone, despite the great amount of work mentioned in the Foreword, there remain about 30 languages 6* which have so f a r not been the subject of even a single monographical description. F o r a whole variety of social causes " l e s s e r " languages a r e constantly being shoved aside by " g r e a t e r " languages, which leads to increased destruction of the p r i m a r y system of the " l e s s e r " language under the influence of the socially m o r e powerful language in contact with it ( c o m p a r e the influence of Azerbaijani on Xinalug, Tajik on Shughni), and several " l e s s e r " languages are on the brink of extinction ( f o r example, Livian, Itel'men, Kerek, Aleut, and several dialects of Eskimo). If from a general social point of view the disappearance of " l e s s e r " languages is a positive phenomenon (linguistic fragmentation hinders communication and thus hinders the uniting of people into large social groupings - and here we constantly encounter the question of the feasibility of a single world language 7*), then f r o m the linguistic point of view it is a disaster, since the disappearance of undescribed or insufficiently described languages is a loss of linguistic facts which might turn out to be p r i c e l e s s for future generations of linguists and cannot be made up. In addition, description of the present state of unwritten insufficiently studied languages which are not threatened with immediate extinction is also very valuable f o r linguistics; the fixation of the present state of these languages will be

2 important for succeeding investigators because of the unavoidable linguistic change that will take place. 8* Finally, the study of unwritten languages is important not only for the future of linguistics but, as we shall attempt to show, for its present as well. As a rule unwritten languages do not have their own linguistic tradition and at the present time may be described only by linguists who are not native speakers of these languages. 2. Field Linguistics as a Branch of Descriptive Linguistics. The complex of linguistic methods which is directed toward the independent creative ( a s opposed to the pedagogical - with grammars and textbooks) study and description of a living language which is not native for the investigator has been named field linguistics. Such a study may pursue very different aims: the language may be studied as a whole or as a fragment, theoretically or practically, by itself or comparatively, synchronically or diachronically, etc. No matter what the concrete aim the empirical basis of any field investigation is the synchronic state of the language. Therefore we may say that field linguistics by the subject of its investigation and often by the aim can be related to descriptive synchronic linguistics (not to be identified with Bloomfieldian taxonomic theory!), directed toward the study and description of concrete languages, that is, to the kind of linguistics which establishes a correspondence between the real language (or a part of it) and the grammar (model of the language) which is ascribed to it. 9* In any such investigation the language itself is not given to the investigator for direct observation, it is an ideal object of study. However, the concrete speech utterances which represent the realization of the linguistic competence of speakers and listeners who know the language can be observed. The basic objects of any conceivable linguistic descriptive activity are: the subject of the investigation (the language), which we will call the target language, the object of observation (texts or, as we will call them, data 10*), and the product of the investigation (the model of the target language, which is usually called grammar). The different types of descriptive synchronic linguistics differ by the character of the interrelations of the investigator with the target language as shown in Figure 1. In Figure 1, 1A represents the situation where a type of "direct communication" between the investigator and the target language is possible, and the data as it were are absent. In fact what we have in this case is the "introspective" method of investigation based exclusively on self-observation. The target language is familiar to the investigator (it relates to his competence) who himself acts as a "generator" of data on the given language and constructs a linguistic description based on data so obtained. This method of investigation has often been attacked on methodological grounds especially under the prevailing influence of the descriptive approach to the study of language which did not recognize the objective value of linguistic self-observations. 11* The most traditional method of descriptive linguistics is represented by IB, where the investigator who speaks toe target language has a

1A Introspective Method

I B Analytical Method

1C Experimental Method

ID Experimental Method Model ( of a fragment) of language

Investigator

Interpreter

Inform- —- Data Language ant (idiolect)

Fig. 1 Types of Descriptive Synchronic Linguistics

4 certain corpus of independently received data which can be subjected to study (This we will call the analytical method). In Soviet linguistics this is a common method, for example, in studies which use literary works as a basis for the analysis of a language. Bloomfieldian linguistics also postulated the possibility of studying a language unfamiliar to the investigator using only an accidental collage of texts, although we are not aware of any succesful examples of this type of investigation and there are grave theoretical doubts as to its feasibility. 12* One should keep in mind that in the case where the investigator is a native speaker of the target language he may combine the analytical method with the introspective, which is what is usually done, although as a rule no distinction is made between the facts drawn from a corpus of data and the facts obtained by introspection. Among field linguists one encounters the opinion that a language can only be well described if the investigator knows the language well enough to be his own informant, in other words only a combination of methods 1A and I B is regarded as a truly scientific method of studying the language (see the following for more details). In 1C a native speaker (the informant) who is not the investigator is used as a "generator" of data on the target language. The informant serves as the basic investigative "tool" of the linguist, and when used wisely will provide the linguist with the facts about the target language that interest him. Here the informant is considered to be dependent on the investigator and rather than being an unregulated "generator" of data, he is a means of eliciting the kind of information the investigator has requested. This method might be called "experimental". 13* Within the framework of the experimental method as within the analytical method (but not introspectively!) two possibilities exist: 1. the target language is familiar to the investigator, and 2. unfamiliar. In the first case the linguist, as in IB, can also use the introspective method as a supplement. The experimental method for studying a native language is well developed in psycholinguistics; moreover, if the self-observation is a precise and operationally formulated principle of investigation then such an introspective methodology can also rightly be termed experimental: the fact that the informant and the investigator are the same person is external to the nature of the experimental method itself. In the case when the investigator is unfamiliar with the target language beforehand, we have the kind of situation involved in field work. In some cases there is a third person - the interpreter who stands between the investigator and the informant, which is represented by ID. This is necessary if there is no common language between informant and investigator by means of which the investigator can control and stimulate the data he is eliciting. In this case the interpreter is a translator between investigator and informant. The specific nature of field work demands a differentiation of the language in which the informant speaks and which is undergoing investigation (the target language) from the language of communication between the informant, the investigator, and the interpreter (the mediator language) and from the language which is used to describe the target language (the metalanguage). ( One should remember that although the native language of the investigator usually serves as the metalanguage this is not always the case, especially if the investigator chooses to use

5

several metalanguages). The different possible combinations of using languages in various functions in field conditions a r e represented in Figure 2 (using examples of the differing use of Russian, Avar, and Archi). Thus the emergence of field linguistics as a method of investigation is due not to internal (depending on the conditions which exist in the investigative situation). Briefly, we understand field linguistics-in the wide sense of the t e r m to be a specific investigative situation where the investigator is not linguistically competent in the target language 14* and the only source of information about the target language is a native speaker of the language. The t e r m "field linguistics" may also be n a r rowly interpreted: the t e r m arose due to the fact that such linguistic activity usually takes place in the field f a r from the linguist's normal place of work and what is more important, in the natural linguistic e n vironment where the given target language is used as the basic means of communication between people. Henceforth we will regard the n a r r o w e r interpretation of field linguistics as the m o r e valuable (although we do not in any sense repudiate the wider interpretation): everything that is applicable to the wide understanding of field linguistics is also applicable to the narrow; the r e v e r s e is not true. 3. Methods of Field Linguistics and Its Relation to Other Linguistic Disciplines. Thus field linguistics owes its existence to the fact that there are situations where a language is not described by a native speaker but by an investigator who, as a rule, in the early stages of his acquaintance with the language knows little or nothing of its characteristics. In the linguistic literature distressingly little attention is paid to field linguistics even though it is one of the oldest and generally speaking one of the most e s sential branches of linguistics. The investigator who decides to study a language he does not know in o r d e r to describe it scientifically runs into an unexpected problem - the lack of clearly formulated principles of how to do so. In addition to this the investigative situation where the linguist does not know the target language and the only source of information about the language a r e native speakers is a very specific one. The situation specifies at least five clearly formulated basic principles to be n e c e s s a r y for a complete investigation: (1) (2) (3) (4) (5)

Methods f o r describing the target language. Methods f o r discovering grammatical facts. Hypotheses about the properties of language in general. Methods f o r eliciting linguistic information from an informant. Methods f o r the practical learning of the target language.

Though these questions ought to be subjected to independent study and rethought in t e r m s of the particular problems of field linguistics, it is c l e a r that they a r e common to the problems faced in neighbouring linguistic domains and f o r this reason field linguistics should take a d vantage of the results of research in these domains. We have represented

6 Grammar 2A

Investigator Russian MTL

Grammar

PI lMj

Informant

Investigator Russian MTL

Grammar

Data

Russian ML

f

Informant Avar ML

Investigator Russian MTL

Archi TL

Data Russ. - MTL

,

Archi, TL

Interpreter Russian ML

MTL ML Archi - T L Rusa.

Informant Avar ML

Data Archi TL i

•5ST (mtl Avar - ML Archi- TL Grammar vy/z/A t/s/s/A

Investigator

Interpreter Russian ML

Russian IYTTT i

Informant Archi ML

Data Archi TL

Kwv/

I MTL I ML ÌML ' Archi T L Russ.

Grammar 2E

yfflyA V////A

Investigator Russian MTT,

Interpreter Avar ML

Informant Archi ML

Data Archi TT \ y / / / / \

Russ. - MTL Avar - ML ÂrchT

Fig. 2. Interrelations of Mediator Language ( M L ) , Target Language ( T L ) , and Metalanguage (MTL) in Field Work

|M L TL

7

the relationship (and it is possible that our representation is f a r f r o m complete) of field linguistics with other linguistic disciplines in Figure 3 (Note that the relationships between disciplines a r e very close in c e r t a i n cases and this we were not able to show on our diagram). As we see, field linguistics intersects many of the disciplines which are in principle important f o r the modern theory of language but does not coincide in its aims or methods with any of them. It is revealing that almost all the linguistic disciplines which share s i m i l a r methods with field linguistics were given scientific status only quite recently (pedagogy and the theory of g r a m m a r a r e exceptions, although the latter is undergoing very serious changes). Thus it is not surprising that the theoretical foundation of field linguistics has received so little attention: although field linguistics and all its methodological questions have existed for a long time, these questions have not only not been resolved, they have not even been clearly stated, since the other linguistic disciplines with a similar problematic had not yet been reinforced and formulated nor could they have been formulated f r o m the general theoretical point of view. Let us briefly examine the similarity in methods between field linguistics and other a r e a s of linguistics. ( 1 ) Methods of describing language. The point of studying an unf a m i l i a r target language, naturally, is to describe it scientifically. The methodology of language description is the prerogative of the theory of g r a m m a r 15*, and one must admit that the state of grammatical theory has had a very telling effect on the results of field investigations. But one may also a s s e r t the contrary - that is, that field linguistics has in various periods in the development of linguistics stimulated a radical review of linguistic theory. A striking example of this is the influence of the work on Indian languages on the formation of a fruitful structural trenddescriptive linguistics. 16* We could also note the outstanding, but unfortunately too little known pioneering works of other field investigators who w e r e forced by the practice of t h e i r field work to depart from the accepted views on language which were held by the linguistic theory of their own day and create their own descriptions based on a different theoretical platform which corresponded more closely to the problems at hand. Among such outstanding field investigators was P. K. Uslar, who provided brilliant descriptions of a number of Caucasian languages 17* which a f t e r almost a hundred years have not only not lost their scientific value, but have not been superceded in detail, precision, and completeness by anything that has come since. The close relationship between field linguistics and the theory of g r a m m a r comes about because the main aim of the theory of g r a m m a r is the formulation of demands applicable to the g r a m m a r of any language and the creation of a descriptive apparatus which will fulfill these demands. 18* Field linguistics offers a clear opportunity to test such an apparatus since the linguist, who is not a native speaker of the language he is investigating, cannot construct a description based on his own linguistic intuition and is forced into a s t r i c t e r application of the c r i t e r i a of c o r r e spondence between the linguistic objects in the language which a r e under investigation and the concepts which belong to the theory. Thus all the f o r m a l weaknesses of the theory soon become evident. Consequently

8

Fig. 3. The Relationship of Field Linguistics to Other Branches of Linguistic Theory

field linguistics in its turn presents demands to the theory of g r a m m a r which should aid the development of the theory. "Almost every new language which is carefully analyzed t u r n s up some phenomena which cannot be adequately treated in t e r m s of existing concepts and which thus reveal flaws o r weak points and prompt f u r t h e r refinements in theory. " 1 9 * " . . . the most crucial problem f o r linguistic theory seems to be to abstract statements and generalizations from particular descriptively adequate g r a m m a r s and, wherever possible, to attribute them to a gene r a l theory of linguistic structure, thus enriching this theory and i m posing more structure on the schema f o r grammatical description. "20* One may suppose that the methods of field investigation and of constructing a g r a m m a r will a s a m a t t e r of course be enriched by modern linguistic theories whose effectiveness will in turn be checked and evaluated when applied to f r e s h linguistic material. (2) Methods f o r the discovery of grammatical facts. In the initial stage of the field investigation of a target language the utterances of that language a r e incomprehensible and its g r a m m a r is unfamiliar. This situation in field investigation is different in principle f r o m the situation

9 where a theory of grammar is applied to material that is already familiar ( f o r example, to the facts of Russian or English which have been described many times). In this situation it is not only necessary to construct.a good grammatical description, but to reveal all the necessary linguistic facts in order to write the description that is, it is also necessary to "discover" the elements of the target language's grammar. This is why an investigator involved with a language he is not familiar with represents a particular case of decoding. The primary condition for any decoding is the presence of texts in the language being studied. The methods and capabilities of decoding are further determined by the additional information available to the investigator. The most important kinds of information for decoding are: ( a ) is the type of elementary entity used on the substantive level of expression of the target language known, i. e. , is it known which are the most basic entities and how the text should be divided in the target language: the sounds of the human voice, letter-phonemes, letter-syllables, letter-words, etc.). ( b ) is there a mediator language into which texts in the target language have been or could be translated (this language may in some cases coincide with the language of the investigator or the target language). ( c ) is there a given non-linguistic situation which is being described by the text (on a somewhat more abstract level a text in a mediator language can also be considered a non-linguistic situation, 21* but the specificity of such a mediator language is too obvious). These types of information are muttually independent and therefore it is possible to classify the types of decoding by their presence or absence to the investigator (at the initial stage of the decoding). (See Figure 4). Of course the investigator may learn during the investigation and then the type of decoding will change (For example, the minuses in the first column may be replaced by pluses at some stage in the investigation). 22*

Field linguistics in the form now given it by practical requirements belongs to the first and more rarely to the second type of decoding ( b i lingual and monolingual approaches). Despite the fairly wide spectrum of decoding types represented in Figure 4 (see types 1 and 8) and the presence of specific techniques for each of them, it is natural to expect common points connected with the grammar discovery procedures of the target language. It makes sense to think that by using the results of decoding theory field linguistics will enrich the theory itself with its own devices. 23* It is necessary to distinguish between the stage of discovering grammatical facts and the stage of describing them. Although on the one hand it is impossible to construct a description of the target language without having discovered a certain aggregate of facts, and on the other hand an exhaustive discovery of all the facts from a grammatical point of view is generally speaking impossible before a grammar of the target language has been begun (see Chapter 3, 1.4), a collection of grammatical facts is still not a grammatical description (different descriptions may correspond to the same aggregate of facts 24*) although in practice the

10

No.

The type of elementary entities of the substance A Mediator of expression language is known is present

The Nonlinguistic situation is known

Examples Studying the language in oral form with the help of a bilingual informant in his native environment. Studying the language in oral form with the help of a monolingual informant in his native environment. Decoding a text in a known system of writing but from an unfamiliar civilization with the help of a bilingual. (1) Decoding a text in a known system of writing but from an unfamiliar civilization without the help of a bilingual. (2) Studying a language (in an unfamiliar substance) with the help of a bilingual informant in his native environment. (3) Decoding a text in an unknown system of writing from a unknown civilization with the help of a bilingual. (4)

Language learning by the child. (5) Studying animal language, the language of inhabitants of other worlds. ( 6) Fig. 4. Types of Decoding

11

Fig. 4. Types of Decoding (Notes) (1) A common type of so-called deciphering linguistic conundrums ( see: B. Ju. Gorodeckij, A._E. Kibrik, A. K. Polivanova, V. V. Raskin, "Ob odnom klasse lingvistiCeskix zadaC" [ One Class of Linguistic conundrums ], Problemy prikladnoj lingvistiki. Tezisi mezvuzovskoj konferencii 16-19 dekabrja 1969 g. [ Problems of Applied Linguistics. Theses of an Inter-University Conference 16-19 December, 1969 ], M., 1969, 107-111; 200 zadac po jazykovedeniju i matematike [200 Conundrums in Linguistics and Mathematics ], Ivd-vo MGU, M., 1972. i(2) The problem of decoding Etruscan texts: J . Friedrich, Entzifferung verschollenner Schriften und Sprachen, Berlin-GOttingen-Heidelberg, 1954, 113-118. The decoding of Carian inscriptions belongs to this type also. See: V. V. Servoroskin, Issledovanija po desifrovke karijskix nadpisej [Investigations in Decoding Carian Inscriptions ], M., 1965. (3) The study of the sign language of deaf and dumb m e m b e r s of the Urubu tribe. See: J. Kakumasu, "Urubu Sign Language", UAL, v. 34, N 4, 1968; also A. L. Kroeber, "Sign Language Inquiry", IJAL, v. 24, N 1, 1958, 1-19. (4) On the decoding of Old JCgyptian hieroglyphic writing by Champollion see: E.Dobl'xofer, Znaki i cudesa [ Signs and Miracles ], M., 1963. (5) It is not important that the child is an unconscious decoder - a decoder "against his will". (6) F o r example, studying the language of dolphins. In view of the impossibility of obtaining a bilingual as is usual with other types of decoding, it is n e c e s s a r y to breach this s e r i e s of minuses by artificially a signing a non-linguistic situation and observing the communicative behaviour of the dolphins. A peculiar non-linguistic situation is assigned in various projects f o r a cosmic language which would be most easily decoded by intelligent beings. See: H.Freudenthal, LINCOS. Design of a Language f o r Cosmic Intercourse, P a r t 1, Amsterdam, 1960.

12

latter is often substituted for by the former. (3) Hypotheses on the general properties of language. The investigator' s ideas about the general properties of language significantly influence his study of an unfamiliar language (and this is by no means always a conscious phenomenon). While the classical theory of grammar based on the Greek and Roman models was still in force investigators who turned to the "exotic" languages were shocked because classical grammatical concepts were not appliable to the new linguistic facts. This led to the idea of "linguistic relativity" 25* (not to be confused with the Sapir-Whorf hypothesis!) which postulated an infinite and unpredictable variety of languages; however, in spite of the thesis of "linguistic relativity" acquaintace with more and more new languages continues to surprise investigators: a new and equally obvious fact is coming to light - that languages are surprisingly alike. "Underlying the endless and fascinating idiosyncracies of the world's languages there are uniformities of universal scope. Amid infinite diversity, all languages are, as it were, out from the same pattern." 26* As Chomsky indicates, the notion of the variety of languages is greatly exaggerated and in any case has to do with the variety of the surface structures and not the deep structures. 27* It is for this reason that field linguistics requires some notion of the properties of language in general. This notion, however, must not be reduced to the properties of, at worst, the investigator's native language, or, at best, to those of the languages he is familiar with. This problem of field linguistics is related to typology and its newest trend-universalism. As is the case with relationships between other linguistic domains, this relationship is mutual" it is hard to create a scientific theory of field linguistics without having formulated substantive hypotheses about language in general, just as it is utterly impossible to construct a typology or even more so a theory of universals without a sufficiently strict, complete and exhaustive knowledge of many concrete languages (and it is this knowledge which field linguistics provides, since it is hardly possible to wait until each insufficiently studied language can be studied by a native speaker).28* Hypotheses about the properties of language in general are especially important for the preparatory stage of field work; armed with these hypotheses one may conceive ahead of time which elements of language structure should be investigated. This allows the investigator to draw up preliminary tests, questionnaires, and interrogation schemes for use with informants. This in turn makes the field investigator's work a good deal more effective (This part of his activity may be called the "zero cycle"). However, linguistics provides too few ready recipes for this stage of the work (it is enough to point out how unfavorable the situation stands with the problem of a single phonetic alphabet, which is one of the most highly developed parts of general linguistics 29*) and in this area practically everything remains to be done. (4) Methods for eliciting linguistic information. Unlike the theory of grammar or decoding in its narrow sense, field linguistics is inextricably linked with the native speaker who is to provide all the information about the target language. Since the intermediary between the investigator and the language is another human being, psychology and especially psycho-

13

linguistics are very important in developing methods of eliciting linguistic information. The basic tool of field linguistics and psycholinguistics is the experiment; the ability to correctly set up experiments therefore considerably affects the results of the work. 30* It is in this respect that the specific nature of field linguistics is revealed especially clearly; it is here that special techniques and methods of investigation are required. The investigator is faced with such methodological questions as working out a way to collect data from an informant (along with the passive methods used for working with texts there are special methods for eliciting from the informant not merely random texts, but texts which contain the sought for linguistic information), working out ways of interrogating an informant, ways of verifying data, ways of processing data, etc. In the course of the work there inevitably arise such methodological questions as the problem of grammaticality, the problem of contradictory data, the correlation of usage and the norm, of language and idiolect, colloquial and full style, the influence of the investigator and informant's perception of the target language on the data, and so on; all these questions have been raised only fairly recently in theoretical linguistics, but without taking them into account it is hardly possible to gain correct results from field work. An essential factor in eliciting procedure is that the native speaker is capable of linguistic activity in the language but does not know how the language is constructed and cannot answer point blank questions about the language. The native speaker can only produce utterances which are grammatical from the standpoint of the grammar of the language and are meaningful from the standpoint of the non-linguistic situation. The task of field linguistics is to develop ways of actively but indirectly regulating the informant's ability to speak the language which will elicit the information the investigator desires. 31* It should be noted that in this respect field linguistics is closely related to field anthropology and ethnography (in the American tradition linguistics is generally considered one of the anthropological disciplines), and that this relationship is not merely external (a similarity in methodology) but internal as well since language can be seen as a part of ethnological phenomena in general, as one of the most important manifestations of intellectual life. 32* The relationship between field linguistics and psycholinguistics is further shown when the investigator takes into account to a certain extent the "linguistic intuition" of the informant - the informant's psychological perception of the facts of his language, before making any decisions about its grammar. The interrelation of grammar and its conscious or unconscious perception is very complicated and polysemantic. For example, the well known thesis about the phonological perception of speech by a listener (the listener hears in the speech signal only phonologically meaningful differences) has received extensive experimental documentation. But at the same time this fact cannot be used dogmatically in constructing the phonological system of a target language by assuming that the sounds which are distinguished (heard) by the native speaker represent individual phonemes or the converse, that indistinguishable (unheard) sounds do not represent individual phonemes. Thus the sound [ t ] in Russian is perceived by the native speaker to be clearly distinct

14 from [ i ] , but in the great majority of phonological descriptions there is considered to be only one phoneme / i / . On the other hand, in Alutor (dialect of Koryak) the sound [3 ] is not perceived as a separate sound at all, and the sequence [ ( 3 ] is perceived as a pure consonant. However, there are a number of reasons for considering this sound phonologically meaningful and correlating it with an individual phoneme / a / . But neither does it follow from this that the facts of how the informant perceives his language are not essential for the investigator. The linguistic intuition of the informant is a great help in field work, but in order to utilize it most effectively it would be highly desirable to discover the correlations between linguistic competence and one1 s perception of it in the process of language use. Unfortunately, psycholinguistics has not done much of anything positive in this direction, but one may expect that the contact between field linguistics and psycholinguistics will be of great help in solving this interesting problem. ( 5) Methods for the practical studying/learning of the language. A practical knowledge of the target language is one of the methods which helps in eliciting information about it. There are various view points on the importance of learning the target language in order to describe it; some investigators feel that it is blasphemous to try to describe a language without being able to speak it, while others reduce the importance of this factor to a minimum; without going to either extreme one may conclude that a knowledge of the language is very useful for the field linguist for a number of reasons (See Chapter 3, 1 . 5 . ) . In this respect field linguistics comes close to the traditional domain of applied linguistics - pedagogy (the teaching of a foreign language)33* and the newest, and perhaps not even independent discipline - the theory of child speech. 34* Hopefully the discovery of the unique and universal human mechanism for learning the language that one hears spoken around one as a child and for learning it with amazing speed and without any planned course of instruction will help to optimize the learning of a language in field conditions. However this area of linguistics is presently far behind the demands of field work. This field linguistics is closely related to various branches of linguistics. Field linguistics is the experimental polygon for various linguistic theories the supplier of new and needed material for linguistics about languages and their properties and the potential customer for the solutions to numerous problems still being worked on by theoretical linguistics. 35* At the same time, field linguistics is an independent applied branch of linguistics with a complicated complex of problems and methods and depends on modern accomplishments in the theory of language for its theoretical foundation. It appears that the weakly developed theory of field linguistics is a temporary phenomenon. In the last decade of modern linguistics there has been an especial amount of attention given to problems of the synchronic description of language, and this has been accompanied by the demand for a universal linguistic theory (applicable not to concrete language X or Y, but to any natural language, including languages X or Y). The theoretical developments in designing descriptive apparatus

15

for many of the components of language are already significantly in advance of the pace of their application to the description of concrete languages and such a gap, at least for the theory, is highly undesirable. For this reason it is most likely that in the next few years the center of gravity of linguistic investigations will shift to the study of languages which have not previously attracted the general and serious attention of linguists. And if this happens then field linguistics will take its rightful place among the methods of linguistic investigation. NOTES 1* W. Samarin, Field Linguistics. A Guide to Linguistic Field Work, Holt, Rinehart, and Winston, 1967, 3. 2* See: M. I. Isaev, Sto tridcat' ravnopravnyx [ One Hundred and Thirty Equals ], M., 1970. 127 Languages are described in the five volume series Jazyki narodov SSSR [ Languages of the Peoples of the USSR ] (M. , 1966-68). 3* In the USSR - 17, in Western Europe - 26, in Asia - around 70, in Africa - around 30, in America - only two non-European languages; Aymard and Quechua. (Source - the reference book Naselenie zemnogo sara [ The Earth's Population], M. , 1965). 4* These languages are (in order of the number of speakers): Chinese, English, Spanish Hindi, Russian, Japanese, Portugese, German; after these eight come: Arabic, French, Italian, Javanese. " 5* Chinese, English, Hindi. 6* Based on the period 1966-1968 when the five volume work Jazyki narodov SSSR was published. With each year the number of such languages decreases. 7* See for example: "E. Svadost, Kak vozniknet vseobgCij jazyk? [ How Will a Universal Language Come About? ] M. , 1968. 8* For example, we know how valuable any surviving monuments of dead languages are for comparative-historical linguistics and how valuable synchronic slices of the earlier states of presently existing languages are. 9* In this opposition of language and grammar language is understood as being a certain real existing system independent of the will and consciousness of human beings which is located in the brain of speakers allowing them to produce and understand utterances which are defined by this system, i. e., language is, in Chomsky's terminology, the competence of speakers and listeners. Grammar is an artificially constructed system which is in certain respects an analogue of the language. 10* "We understand a text (with the exception of Chapter 3, 3 . 2 . ) as a speech utterance of indefinite length, a word, a sentence, a whole narration. Texts-data exist independently of the investigator. Collections of data specially gathered by the investigator for the purpose of scientific examination we will call the corpus (of data). 11* See also the heated arguments about the mentalism of generative grammar: N. Chomsky, Aspects of the Theory of Syntax, Massachusetts, 1965, 193. 12* Generally speaking there are individual attempts at fragmentary

16 description for example: B. Bowman, "An Attempt at an Analysis of Modern Yucatec from a Small Corpus of Recorded Speech", Anthropological Linguistics, 1 (4), 1959, 43-86, but in this work there is no confirmation that the analysis is correct and the author himself admits to hoping for only a partial coincidence with the true facts of the language. " 13* See an analogous classification in Ju. D. Apresjan's work: Eksperimental'noe issledovanie semantiki russkogo glagola t An Experimental Investigation of the Semantics of the Russian Verb ], M. , 1967, 20-23. 14* From this it follows that, if we include investigations where the investigator speaks the target language in field linguistics, then field linguistics, practically speaking, merges with the experimental method. It is also obvious that the investigator who speaks the target language will use a number of analogous techniques and rescources. See, for example: R. W. Shuy, W. A. Wolfram, W. K. Riley, Field Techniques in an Urban Language Study, Washington, 1968. 15* This term is used in the wide sense and is applicable to all levels of the description of language, in particular semantics which occupies an honoured place in the system of the description of language in modern linguistic theories. 16* H. Hoijer, "Anthropological Linguistics", Trends in European and American Linguistics 1930-1960, Utrecht/Antwerp, 1961, 110-112. 17* See, for example: P. K. Uslar, "Avarskij jazyk" [ The Avar Language ] in Etnografija Kavkaza. Jazykoznanie [ Ethnography of the Caucasus. Linguistics],^. Ill, Tiflis, 1889; P. K. Uslar, "Lakskij jazyk" [ The Lak Language ] in Etnografija Kavkaza. Jazykoznanie, v. IV, Tiflis, 1890; P. K. Uslar, "Xjurkilinskij jazyk" [The Dargin Language] in Etnografija Kavkaza. Jazykoznanie, v. V, Tiflis, 1892. 18* H.A.Gleason, An Introduction to Descriptive Linguistics, N. Y . , 1955, Preface. 19* F. Lounsbury, "Field Methods and Techniques in Linguistics", in A. L. Kroeber et. al. Anthropology Today, Chicago, 1953, 408. 20* N. Chomsky, A s p e c t s . . . , 46. 21* Such an expanded concept of the text is contained, for example, in K. L. Pike, Language in Relation to a Unified Theory of Human Behaviour, The Hague, 1967.! 22* Thus the problem of decoding Creto-Mycenean linear writing B belonged to type 8 after Evans' discovery of clay tablets (all three minuses); after Alice Kober's investigations which established the syllabic nature of the writing - to type 4, and after Ventris had the hypothesis that the texts written in this language were related to Greek - to type 3, after which the texts were decoded. 23* There is an attempt to base "discovery procedures" on tagmemic theory: R. E. Longacre, Grammar Discovery Procedures: Field Manual, The Hague, 1964. The problem of discovery procedures also takes a significant place in P. Garvin's works. See P. L. Garvin, On Linguistic Method, The Hague, 1972. 24* See: A. E. Kibrik, S. V. Kodzasov, I. P. Olovjannikova, Fragmenty grammatiki xinalugskogo jazyka [ Fragments of a Grammar of Xinalug ], M., izd-vo MGU, 1972, Conclusion. 25* See: M. Swadesh, "Language Universals and Research Efficiency in Descriptive Linguistics", The Canadian Journal of Linguistics, v. 10:

17 2,3, 1965, 148. 26* J. H. Greenberg, Ch. E.Osgood, J.J.Jenkins, "Memorandum Concerning Language Universals" in Universals of Language, Report of a Conference Held at Dobbs Ferry, New York, April 13-15, 1961, ed. by J. H. Greenberg, The M. I. T. Press, Cambridge, /Mass/, 1963, p. 255. 27* N. Chomsky, Aspects 118. N. Chomsky, Language and Mind, 1972, Chapter 3. One may seek an explanation for this fact not only in language learning conditions but in the anthropological theory of the "widespread monocentrism" of human origins: "the unity of the mankind now alive rests not only on the regularities of social development, but on the kinship of all its constituent races" (Ja. Ja. Roginskij, Problemy antropogeneza [ Problems of Anthropogenesis ], M., 1969, 123ff.). 28* See B. A. Uspenskij, "Jazykovye universalii i aktual'nye problemy tipologiceskogo opisanija jazyka" I Language Universals and Current Problems of the Typological Description of Language] , Jazykovye universalii i lingvisticeskaja tipologija [ Language Universals and Linguistic Typology ], M., 1969, 12-14. 29* See: A. E. Kibrik, S. V. Kodzasov, "Principy foneticeskoj transkripcii i transkripcionnaja sistema dlja kavkazskix jazykov" [ Principles of Phonetic Transcription and a Transcription System for Caucasian Languages], VJa, 6, 1970. 30* See: A. E. Kibrik, "Psixolingvistiöeskij öksperiment v polevoj lingvistike [ A Psycholinguistic Experiment in Field Linguistics] in Materialy tret'ego vsesojuznogo simpoziuma po psixolingvistike [ Materials of the Third All -Union Symposium on Psycholinguistics] , Moscow, 1970, pp. 160-161. 31* W. Samarin's book (see note 1) was one of the first to break the conspiracy of silence around this problematic. See also, S. C. Gudschinsky, How to Learn an Unwritten Language, N. Y . , 1967. 32* F. Boas, introduction to A Handbook of American Indian Languages, University of Nebraska Press, Lincoln, 1966, 55-59. 33* Among the classical works on language teaching methodology we may indicate: H. Sweet, The Practical Study of Languages: A Guide for Teachers and Learners, L . , 1964 (reprint of the 1899 ed.); H.E.Palmer, The Scientific Study and Teaching of Languages: A Review of the Factors and Problems Connected with an Analysis of Various Methods Which May Be Adopted in Order to Attain Satisfactory Results, L., 1917; for later works see: W. A. Benneth, Analysis of Language and Language Teaching, 1968; R.L.Politzer, Foreign Language Learning, a Linguistic Introduction, 1970; J. O. Robinson, An Annotated Bibliography of Modern Language Teaching, L . , 1969. 34* On the importance of such a theory for formulating the foundations of linguistics see: N. Chomsky, Aspects . . . , 30. On the description of the genesis of child speech see: A. N. Gvozdev, Voprosy izucenija detskoj reci [ Problems of the Study of Child Speech] , M . , 1961. See also: A. Bar-Adon, W. F. Leopold (eds.), Child Language. A Book of Readings, 1971. 35* See the similar point of view: P . Garvin, "American Indian Languages - a Laboratory for Linguistic Methodology", Foundations of Language, v. 3, No. 3, 1967.

PRELIMINARY STAGE ( THE "ZERO CYCLE") IN FIELD WORK

1. The Necessity for Questionnaires and Tests The advance preparation for field work depends on the aims the investigator wishes to pursue with his description. In view of the unlimited number of concrete problems which may be raised in the field investigation of a target language these programs may be quite varied. However, in order to limit the scope of our investigation we are going to concentrate on one fairly large task of any field investigation: the initial (that is, being done for the first time) discovery of the grammar of a language whose genetic relations and typological properties are totally unknown (though it is true that the investigator usually has some rudimentary information about the target language, it is more interesting to start from the extreme case since the whole diapason of grammatical hypotheses is minimally limited). Further, we will consider the theory which held advance preparation for field investigation as useless and even harmful because it supposedly prevented an objective and unbiased analysis of the language free of all apriori grammatical dogmas. It has outlived its usefulness and is no longer a theoretically applicable approach. It is obvious that no grammatical analysis is free from some preconceived concepts, but partiallarly undesirable is subconscious dogmatism hiding behind the mask of unbiased objectivity. If, however, all of our underlying assumptions are fixed and explicitly stated, it is not difficult to exchange them for other concepts should the need arise in order to more adequately describe the object. 1* The basic tasks of the "zero cycle" are first a maximally complete and exhaustive explication of the kinds of specific grammatical facts the investigator is likely to encounter in field work with the target language, and which of these facts are not particular to the target language but are language universals, and secondly to work out linguistic means for discovering the fundamental grammatical facts of the target language in an optimum amount of time. Let us examine the first problem. It amounts to applying the data which has been accumulated by modern theoretical linguistics to the needs of field work. If we begin from the state of present knowledge it is apparent that this is an ideal and one which practically speaking can only be approached at a very distant remove. Nevertheless, a certain systematization of the information accumulated about world languages and a generalization from the data of typology and universalism is possible at the present time. The "substantive" universals according to which entities of any sort whatsoever in any language should be characterizable in terms of a

20

previously fixed class of entities a r e especially important f o r the "zero cycle". One does not usually encounter any specific fact in a list, but r a t h e r a s a c l a s s of facts which a r e alloted definite features; a given fact belongs to its class because of the meanings of these features. 2* The universal inventory of phonetic features in whose t e r m s any phonologically useful sound produced by the human speech organs can be described with sufficient precision is an example of this kind of "substantive" universal. 3* By making use of this universal and very limited inventory the investigator is able to grasp the unlimited variety of sounds which he may encounter. The second task is partly a result of the practical application of the results of the f i r s t , and partly an attempt to compensate for deficiencies in modern linguistic knowledge. That is, without conducting an exhaustive typological analysis it is possible to devise f r o m the mediator language a m o r e o r less effective empirical test which will give us sufficiently complete information on the presence o r absence of any grammatical categories in the target language, without requiring a calculus of the categories chosen by the test. A list of the diagnostic noun p h r a s e s of the mediator language ( Russian, English, French, etc.) compiled in o r d e r to choose the nominal word inflexional categories of the target language is an example of this kind of test ( s e e Chapter 2, 3 . 3 . ) . One should keep in mind that it is impossible to devise a test o r questionnaire sufficiently adequate to elicit all the facts essential to the c r e ation of a model of the target language. Any test constructed on apriori universal principles is self-restricting and cannot envisage all the peculiarities of the target language. The point of such methods is rather m o r e simple: to reveal the basic properties of the target language and to allow the investigator to narrow the class of hypotheses so that the further analysis of the language can be conducted with a concrete specification of the language already at hand. Thus, using the system of universal phonetic f e a t u r e s one can very quickly establish which of these features a r e differential in the target language. This in turn allows the investigator to choose the basic phonemic ensemble of the language, although the knowledge of these features is in itself not sufficient to conduct a full phonological analysis. The development of the "zero cycle" is least of all intended to turn the investigative p r o c e s s into an algorithm and to eliminate its heuristic human origin; it is used to ease, wherever p o s s ible, and to stimulate the heuristic side of the investigator's activity. The specific nature of field linguistics is related primarily to the beginning step of the investigation when the investigator has no real knowledge about the target language. This is why the point of the "zero cycle" is to provide a means of bringing the investigatory linguistic apparatus into play. The difference between field and the usual experimental descriptive linguistics disappears the m o r e the target language has been studied and the greater the body of data that has been collected on it. Thus the techniques of field linguistics a r e the " s t a r t e r " which sets the linguistic engine in motion. Since oral utterances a r e the only initial material in the field study of

21 2. Phonetics Since oral utterances are the only initial material in the field study of an unfamiliar language the first thing one must do is identify the sounds of that language, and record them in writing. The ability to hear the sound of a foreign language is not innate and can be acquired through special general phonetic training, which should not be conducted unmethodically but on the basis of certain classificatory principles. As we noted in 1, the limitless variety of speech may be reduced to a small list of phonetic features which can combine to make up any sound. The ability, therefore, to identify not only sounds but their features as well is of special importance. Thus it is not enough for the field investigator that he has the theoretical knowledge of which phonetic features can be realized in the sounds of different languages but he must also be able to identify them. The means of identification may be: accoustic analysis, the imitation of a sound one has heard, the visual observation of the speaker's speech organs, the native speaker's observations about the position of his articulatoiy organs, and the kinesthetic self-observations of the investigator. The means of transcribing speech is also a very important question. There are a great many different transcription systems applicable to different languages and several so-called universal systems of notation. 4* For field work, of course, a universal system is required, since the specifics of the sounds of the target language are not already known. We will examine the conditions which a universal system of phonetic transcription should satisfy. 5* 2.1. Conditions on Universal Phonetic Transcription First one must examine the question of the character of the phonetic elements which are to be used to describe the text and also the question of the inventory of such elements. It is clear that the first transcription of a text in an unfamiliar language must proceed in terms of phonetic features and their complexes - phones, rather than in terms of distinctive features and their complexes - phonemes. And if it is preferable to describe distinctive features as binary (the distinctive feature indicates whether a phoneme does or does not belong to a given class 6*) then phonetic features may be binary or they may have a greater number of gradations. 7* There are two ways of representing utterances phonetically (articulatory and acoustic) which can be subjected to objective (instrumental) analysis and allow us a physical expression of the phonetic features. Apparently there are two subjective forms which correspond to these objective forms: the phonic and the kinesthetic. In these forms of expression phonetic features exist as kinds of models or standards which are stored in the memory. Both the objective and subjective features related to various forms of the phonetic representation of speech are used in phonetic descriptions. In traditional descriptions articulatory features are usually used, sometimes with the names of subjective qualities of sounds. Lately it has been the trend to describe speech by acoustic features. Obviously a description based on subjective features is practically impossible. Although objective neuro-physiological correlates for acous-

22

tic and motor sensations in principle do exist, so little is known about them and they are so difficult to observe that any translation of sensations into these objective correlates remains only theoretically possible. How is it possible then to record the phonic and kinesthetic standards of a person? It can be done with the aid of some impressionistic and haphazard terms which are kinds of labels for the sensations of the native speaker and the investigator: if we choose to describe by subjective features without correlating our impressionistic terms with objective features, then every linguistic description will be written in its own special language and will be incomparable to any other description. However, impressionistic t e r m s for features are usually kept in linguistic descriptions only in those places where their articulatory and acoustic properties are clearly known (voiced-voiceless, hard-soft, etc.). Other instances of impressionistic naming lead to confusion and the impossibility of understanding what is being described. The word "guttural" is an example of this. Sometimes this word is used for an indiscriminate designation of the acoustic quality of several consonants which include those articulated in the larynx as well as different pharyngeal and uvular sounds which seem strange to a European ear. In a number of cases impressionistic terms may be more closely r e lated to the phonological conceptions of the linguist than to the real phonetic qualities of the sound. The use of the words "strong-weak" in the description of the Caucasian languages is an example. These terms r e f e r to features which have different objective expressions: the intensity and duration of the fricatives, the duration of the closure of plosives and affricates, the unaspiratedness of the plosives. The common term for these features portrays a functional unity (and apparently a certain kinesthetic similarity as well) but does not reveal their phonetic nature in a concrete language and phonetic position. This is why it is desirable not to use impressionistic terms for phonic and kinesthetic images in linguistic descriptions. At the same time these images a r e clearly the reality with which the linguist must deal with first, especially in field conditions. It is therefore desirable to combine the subjective qualities of sounds with their associated objective qualities and refer to the sounds by their objective qualities. The advantage of using articulatory features ( and terms) in field work is obvious. First of all several articulatory features are easily observable 8* and do not require instrumental analysis, for example, labial articulations and articulations which involve the raising (glottalized, pharyngealized) or the lowering (implosive) of the pharynx, and several articulations which are characterized by tongue position. Secondly, training allows one to go from unformed and undifferentiated kinesthetic sensations to the arbitrary control of many articulatory parameters. The trained phoneticist, having sought out by imitation a kinesthetic image which corresponds to the acoustic image, may take a description in precise articulatory terms. At the same time it is not expedient to describe oral speech in terms of acoustic features without using instrumental methods ( spectrography, oscillography, etc.) which are hardly possible in field work. To use the names of acoustic features without the support of instrumental methods really just amounts to referring to the phonic and kinesthetic standards

23

of the investigator, which are in no way related to the objective observation of features, in acoustic terms. There is one more important consideration in favor of the use of a r ticulatory description for purposes other than field work: according to the motor theory of perception, which has recently found wide acceptance, 9* a person usually analyzes an acoustical signal by reestablishing the underlying articulator^ movements which produced the signal. This is due to the close relationship between the phonic images which are created by the acoustic signal and the kinesthetic image which correspond to the articulatory states that caused the acoustic signal. Thus the articulatoiy and kinesthetic forms of representing phonetic features seem to be the underlying forms, while the acoustic and phonic forms are derived from them. All these considerations force us to prefer a phonetic transcription based on articulatoiy features. Sometimes it is permissible to use the familiar impressionistic terms whose precise objective correlates are clear (for example, we will use the terms "voiced-voiceless", "noisy-sonorant", "stressed-unstressed"). In addition it is possible that in describing an unfamiliar language the investigator will hear a special quality of the sound but will be unable to understand the articulatory mechanism, in which case he will be forced to refer to the feature by an impressionistic term which hopefully can be replaced by a precise articulatory term after an instrumental investigation. It is also true that some articulatory terms are fairly impressionistic and more closely reflect undifferentiated kinesthetic sensations than a precise description of the articulatory movements (for example, the name of the zones and rises of vowels). This, however, does not negate the value of the principle of defining sounds in articulatory terms, it just means that these particular terms should be made more precise. The question of the inventory of phonetic elements which is to be used to transcribe the text boils down to this: how detailed sould the initial phonetic transcription be, or in other words, how large should the inventory of phonetic features be? If we are going by the potential capacity of speaker to realize the finest (within physiological limits) articulatory control, and also to differentiate and identify any sound features (within psychoacoustic limits), then the inventory of phonetic features for the initial description of a text might well be enormous. However, we know that human perception is categorical in nature: we do not perceive sound differences very well which are not essential from the point of view of the system of sound oppositions we have learned. Of course one could suppose that the linguist as a result of a long period of training could learn to identify the finest sound differences, which correspond to the slightest differences of a r ticulatory positions, and would be able to make an initial transcription of speech utterances in an unfamiliar language in terms of such a detailed inventory of features. It is also evident that such a transcription would be very cumbersome and in many instances redundant. The natural alternative to such a "purely phonetic" transcription is one whose detail is defined by several functional criteria: it is reasonable to transcribe only those phonetic features which can presumably be used to express linguistic meanings in natural languages.

24

All phonetic features can be divided into two classes: (1) Features whose different meanings serve as correlates for the different meanings of corresponding distinctive features in concrete natural languages. (2) Features which are not registered in described languages as expressing distinctive features. All the features in the first class answer our functional criterion and are included in the inventory. Features belonging to the second class, in turn, are divided into two sub-classes. (2a) Phonetic features which are registered for a language as specific peculiarities of its phonetic system (for example, the feature of the alveolarity of front stops in English is not phonologically opposed to the feature of dentality, however, English front stops differ phonetically from the corresponding front stops of French which are characterized as dentals). (2b) Non-specific features whose occurence in the speech chain is conditioned by universal coarticulatory effects. These features forcibly appear in any language in the same phonetic conditions. Coarticulation may be simultaneous (for example, the palatalization of a consonant is always accompanied by a certain change in the place of articulation), or successive (for example, consonants are always labialized before labialized vowels). 10* The features in (2a) should, in accordance with our functional c r i terion, be included in the inventory, since in principle they may be used to ejgaress phonological oppositions and the fact that they are not included in Class (1) may be a result of our incomplete phonological-typological knowledge (If we constructed a system of transcription based on knowledge of the phonetics of English and French alone, then in accordance with our principle, the feature of alveolarity-dentality would be introduced for plosives, although phonologically this principle is not "at work" in these languages; there actually are languages where this opposition is relevant, as in the Australian language Arapana. 11* The features in class (2b) should not be included in a universal inventory of phonetic features since in principle they cannot serve to express phonological oppositions. With this functional approach phonological-phonetic considerations of a typological nature define the choice of an inventory of phonetic features. The conditions for the system of transcription are not limited to the question of the nature and inventory of its elements (content demands) but also involve the alphabet of transcription signs which serves as the graphic form for expressing the inventory of phonetic features and phones (formal demands). In principle there are two extremes possible in constructing an alphabet of transcription signs. The first is to refer to every phonetic feature by a special sign, and every phone (complex of phonetic features) by a certain ordered series of such signs (the so-called "analytical method" 12*). In the second method each phone is referred to by a special grapheme which cannot be further reduced (the so-called "synthetic method" 13*). Both methods have their advantages and disadvantages: with the first the inventory of signs is minimal, but the phonetic

25 transcription of the text is cumbersome; with the second the inventory of signs is great and therefore the transcription of the text is compact. Usually a mixed method of transcription is chosen: definite complexes of features are referred to by letters and features not belonging to them by complementary letters or diacritical marks. It would seem that such a mixed system of transcription is more convenient. With such a system it is desirable to observe the demand that correlative features if possible be referred to by complementary signs. This insures flexibility in the transcription system: the discovery of phones in a given language which represent combinations of correlated features which have not yet been registered with known complexes of features does not require the invention of new signs. Of course deviations from this principle are possible when dictated by tradition or the need for graphic convenience (for example, it is justifiable te refer to voiced and voiceless obstruents with special letters). Graphic simplicity and convenience is an important condition for an alphabet of transcription signs: it is undesirable to have an overloading of diacritics, to have graphically complicated signs, or to use signs which are graphically similar. One must keep in mind that the principle of simplicity and convenience is modified significantly depending on the means by which the texts are to be written: (a) a handwritten transcription; (b) a typewritten transcription; (c) a typographical transcription. What is convenient for one type of transcription is not always convenient for another. Thus for handwritten and typographical transcriptions it is convenient to use an inverted comma as a diacritical mark, while with a typewritten transcription this diacritical mark complicates the print. Of course a graphic sign system which satisfied the needs of all types of transcription s y s tems would be optimal, though with such a transcription we may have to omit the references which are most convenient for the commonest form of transcription and replace them with less convenient references. It is also necessary, as much as possible to have a monosemantic use of signs and a monosemantic reference to features (or phones) ( Each sign has one strictly defined phonetic meaning, each feature (or phone) has only one reference). If the system is to have different graphic modifications it will be necessary to indicate which modification is being used. In choosing a system of complementary signs it is necessary to observe the condition that the text be reducable monosemantically into its constituent phones that is, that the boundaries between aggregates of signs which refer to the same phone be drawn monosemantically. 14* 2. 2. Phonetic Transcription We have already established the principles on which a universal phonetic transcription should be based. A fairly large number of conditions have be placed on the transcription system, and in some instances they are contradictory. Therefore it is useful to remember Sweet's remark written at the end of the last century: "We must not forget, moreover, that all alphabets - even the most scientific - are intended to serve practical purposes. Practice implies compromise. Hence every alphabet must in some respects be a compromise between opposite principles." 15*

26

Thus the basic components of a universal phonetic notation are: (a) the elements belonging to the notation; (b) the signs which are the symbols of the elements; ( c) a syntax of the signs (rules for their juxtaposition). The completeness of the inventory of elements which covers the entire variety of sounds in world languages insures the universality of the transcription. Once such an inventory exists it will not be theoretically difficult to choose some means of representing it graphically. In Appendix 1 a transcription system is proposed. It is a variant of the IPA with slight changes and additions. The graphic basis for the transcription is the differentiation of basic graphemes from diacritical graphemes. Basic graphemes refer to "cardinal" sounds, which are characterized by a basic mode and place of articulation. Let us examine the ways of referring to consonants. The following places of articulation are considered "cardinal" for consonants 16*: bilabial, labiodental interdental, dental (alveolar), palatoalveolar, prepalatal, palatal, velar uvular, pharyngeal, laryngeal. In addition the following modes of consonant articulation are considered "cardinal": obstruent plosives, obstruent fricatives, sonorous nasals, sonorous non-nasals (laterals, vibrants, glides). The combinations of "cardinal" places and modes of articulation gives us a table of "cardinal" sounds. To represent the voiced and voiceless variants of each "cardinal" sound, that is, the correlated feature voiced-voiceless, by a special sign is inconsistent; they could be represented by diacritical marks. However the Latin alphabet, which is the graphic basis for the notation already has references for many of the voiced-voiceless correlates and it would hardly be sensible not to use them. On the other hand, alveolar plosives are not given their own special signs and are noted by diacritical marks where necessary despite the principle of choosing "cardinal" places of articulation. "Cardinal" vowels are characterized, first of all, by tongue height and zone, and secondly by roundedness. Five height levels are distinguished: low, middle-low, middle, mid-high, and high; and three zones front, central, and back. In principle the feature of labialization which is a correlative feature would be more consistantly referred to with a diacritical mark (as with the consonants), but on the whole the signs for vowels in the Latin alphabet are distributed so that back vowels are labialized and front vowels are non-labialized (and these are the kind of vowels we find in many languages) so that the feature of labialization belongs to the meaning of the basic sounds. Since we don't wish to change the meaning of these signs it is necessary to add new signs for corresponding labialized front vowels and non-labialized back vowels. Diacritical marks which combine with the basic signs by means of elementary syntactic rules are also used in the notation: they may be above, below, or to the right of the line. The latter may be, in addition, above, below or on the line. The meaning of the diacritic is determined from the sign and the position in which it is used, since in various positions the same diacritical mark refers to different features. If there is more than one diacritic in one position, their order is meaningless: thus [H] and [ ci] both refer to a long nasalized [ a ] . Two letter references are used for affricates and bifocal sounds.

27

2.3. Use of Transcription Correlative features are referred to in such a way that each is distinguished from all the others. The striving to refer to each "cardinal" sound and each correlative feature by a special mark leads to references which are not always convenient for typewritten or typographical transcriptions. Therefore in principle it is allowable to replace a less convenient sign with a more convenient one if this does not lead to polysemy within the text in the target language, and under the condition that the deviation is specified. Thus the sign x (Latin) can refer to an uvular fricative [x] (Greek) if there is no corresponding back sound in the target language, pharyngealization can be referred to by the sign " (the quotation mark is convenient for typewriting), and not c (for example, a", q") if there are no implosives in the target language, etc. Furthermore, one may reject Latin letters altogether and replace them, for example, with Russian letters (here one must resort to some graphic "sleight of hand" since Russian equivalents cannot be found for all "cardinal" sounds), if practical considerations require it. On acquaintance with the target language it may turn out that the means of the given system of notation are insufficient, i. e . , there is a "cardinal" sound in that language which does not correspond to any sign in the table of basic signs, nor to any feature contained in the list of correlative features (for example, it might be necessary to introduce the feature of vibrantization). Nevertheless the logic of a given notation is clear enough to show what sort of sign should be added in a given situation. So far we have been talking only about phonetic transcription. However it is clear that this is not an end in itself but an imposed intermediate stage of notation where nothing is known about the phonological system of the target language and any phonetic feature (or phone) could be differential (or a phoneme). A phonetic transcription is inconvenient for grammatical description because it is cumbersome and redundant and also because it hides many of the regular grammatical correspondences between wordforms which become evident on a higher (phonological or morphophonological) level. Therefore the true aim of a notation is the creation of a phonological or morphophonological system of transcription (unless of course we are conducting a special phonetic investigation). Therefore it is appropriate to indicate briefly the way in which the inventory of distinctive features and their complexes (phonemes or morphonemes) which is derived from the initial phonetic transcription of texts is related to the inventory of phonetic entities, and how the graphic equivalents of these inventories are correlated. After it has been established which phonetic features are contextual realizations of the same differential feature, one of the corresponding phonetic realizations is usually accepted as the "basic form" of the distinctive feature and provides the name and the transcriptive reference for it. Other phonetic realizations of this distinctive feature are not included in the inventory of distinctive features unless they coincide with the "basic form" of another distinctive feature, and this inventory of distinctive features is abbreviated in comparison with the inventory of phonetic features. Thus in the majority of cases the inventory of names for distinctive features is a subset of the names for phonetic features, and the alphabet of signs for a phonological or morphophono-

28 logical transcription of a text is a subset of the alphabet of signs for its phonetic transcription. 17* Only in the rare cases where it is not clear what the "basic form" of the differential feature or phoneme (morphoneme) is, special signs not included in the phonetic alphabet will be introduced for purposes of description on a higher level. 18* Sometimes when it does not lead to confusing phonemes or morphonemes within the target language, it is convenient to replace more "complex" references with simpler ones, having indicated the phonetic meaning of each sign to be used in the new function (Thus in the notation for Archi it is possible not to refer to the aspiration of plosives, since they are not opposed by this feature: and write / t / , / P / in place of [ f ^ ] > [ pk ] . the voiceless uvular fricative [x] (Greek) can be referred to phonologically as / x / (Latin), since in Archi there is no corresponding velar). Thus from a good phonetic alphabet there is an easy transition to a phonological alphabet (Here we have in mind the technical and not the theoretical side of the question). Problems having to do with the final grammatical description of the target language are beyond the scope of the present work, and for that reason we have not touched upon the complex question of defining the phonemic ensemble and choosing a final system of transcription (See however, Chapter 3, 3.). We note only that the resolution of these questions is usually closely related with higher levels of structure, especially with morphophonology and cannot be accomplished independently of them. 3. Morphology 3.1. A Method of Discovery of Morphological Entities The purpose of the "zero cycle" in preparing for the study of an unfamiliar target language is to create a program of investigation which will help to reveal the ensemble and meaning of accessory morphemes in the minimum amount of time, taking for granted that we are able, first to divide the text into words, and second to choose the root and accessory parts of these words (both problems are far from trivial in the sense of solving them for all cases 19*, but are fully solvable if we are willing to be satisfied with what is as a rule a high percentage of obvious cases). Segmenting the text in the unknown language into morphemes is conveniently done by simultaneously controlling the difference in forms and meanings in similar utterances and observing which changes in meaning lead to which changes in form and vice versa. We cannot arbitrarily change the form of an utterance in an unfamiliar target language since the probability of producing a grammatically correct utterance this way is infintesimally small. But we can change the meaning of the utterance if we are able to segment it into its constituent components. For example, if an utterance in the target language can be translated in English as "I live in the country", we can get other sentences similar to the first by changing the components of meaning of the first sentence: "I lived in the country", "I will live in the country", "I often live in the country", "I always live in the country", "I don't live in the country", "He lives in the country", "I live in town", "He lives in town", etc.

29

The translation of these sentences into the target language helps u s carrying out their segmentation into certain of their constituent parts, correlating these p a r t s with certain complexes of meanings. This is the technique used to solve linguistic conundrum-bilinguals which a r e constructed to contain all the information needed f o r establishing the r e lations between the f o r m s and meanings of the morphemes in an unfamiliar language. 20* For this reason it seems more pertinent to use the "zero cycle" to investigate the types of grammatical meanings ( f i r s t of all lexical meanings since their change is especially easy to represent in the mediator language) and to create questionnaires which will allow us to discover those meanings in the target language that have become g r a m matical obligatory. In other words we need to have a typology of g r a m matical meanings, a discovery of the substantive universals on the semantic level analogous to the phonetic substantive universals. The same functional criterion is used in constructing a system of semantic substantive universals a s was used to construct the system of phonetic substantive universals. A meaning is considered grammatical if has been registered in some language as grammatical or, if following deductively from certain premises, it could be grammatical in some as yet unfamiliar language. Having a system of such semantic substantive universals would mean having an inventory of features whose t e r m s could be used to describe and discover the real meanings which a r e grammatically expressed in an a r b i t r a r y target language. Clearly this is not a problem which is going to be solved in the near future. However, we can work towards developing an intermediate solution which will provide us with a known positive result. Before all else we need the most general kind of c l a s s i fication of grammatical meanings and we need to c r e a t e questionnaires for their separate types. (Among these meanings a r e the attributive, the object, the adverbial, the spatial, the temporal, the aspectual, the modal, the pronomial, etc.). The most explicit task of a system of meanings is to serve as a kind of calculus, enumerating all the allowable combinations of the elementary meanings of a certain type and to reflect the c h a r a c t e r of the relations between elementary meanings. This c a l culus is a metalanguage f r o m which the needed meanings can be t r a n s fered to the mediator language. An example of this kind of calculus based on spatial relations is analyzed in Chapter 2, 3. 2. However, it isn't always possible to construct such a system. In this case a questionnaire which can be given directly to the informant in the mediator language is useful (in the form of diagnostic utterances so chosen that their constituent elements introduce some meaning capable of being grammaticalized into the utterance). A model of such a q u e s tionnaire f o r analyzing the prepositional-postpositional-case system (excluding spatial relations) in the target language will be examined in this Chapter, 3.3. It is also possible to construct multiaspectual questionnaires which combine different properties and approach the same goal f r o m different directions. The questionnaire for discovering the categories of grammatical gender and case, which is unfortunately a bibliographical rarity, is an example of this type. 22*

3 . 2 . A Sample of Metalinguistic Calculus of Grammatical Meanings ( F o r the Case of Spatial Meanings) Spatial meanings p r i m a r i l y define the location of the subject o r the object of the utterance (henceforth r e f e r r e d to an S) in relationship to some object which serves as the orienting point (Or) whose position in space is given: "I am walking in the field" - the location of the subject of the sentence "I" is oriented in relation to the object "field", "I am putting the book on the table" - the location of the object of the utterance "book" is oriented relative to the object "table". In principle t h e r e may be several orienting points: the sentence "The chair is standing to the right of the table" usually supposes that the chair is oriented not only relative to the table but to the speaker as well, whose position d e t e r mines the concept "to be on the right of the table" and "to be on the left of the table": :The forest is located up river beyond the village" - the orientation is relative to the village the speaker, and the river. We will call such spatial relations orienting. An orienting meaning assigns a spatial region (Sr) which in turn characterizes the position of S relative to Sr. Meanings which characterize the movement (which can also be exp r e s s e d by r e s t - a lack of movement) of S relative to Sr we will call motive. In many languages it is characteristic f o r orienting and motive meanings to be formally delimited one from the other (to be expressed by different morphological means). They a r e as a rule related to the inflexion of a noun ( a case change in the noun or in the prepositiona-postpositional system), though this is not mandatory ( f o r example, motive meanings may be expressed in the verb - either lexically or morphologically) . Let us f i r s t examine orienting meanings (See Appendix 2); Some features ( 'outside Or' vs 'inside Or', 'the relevance of the top o r bottom f o r Or' vs 'the irrelevance of the top or bottom for Or') may be either marked (plus o r minus) o r unmarked, which gives u s a t e r t i a r y rather than a binary segmentation of meaning. Some features ( ' t h e center of Or' vs 'the edge of Or', 'contact with Or' vs 'non-contact with Or', etc.) a r e applicable to many combinations of elementary meanings and in such c a s e s in o r d e r to p r e s e r v e the compactness of the diagram such a feature is noted once with an indication (dots) that all possible chains of elementary orienting meanings belong to it. Each chain on the branches of the t r e e corresponds to a certain complex orienting meaning. Let us now turn to motive meanings. The opposition 'approaching Sr' vs 'moving away f r o m Sr' indicates the direction of movement ( o r a lack of direction) relative to Sr. The opposition 'contact with Sr' vs 'non-contact with Sr' indicates the possibility/impossibility of the positions of S and Sr coinciding (We note that for motive meanings 'contact' has a different meaning than it does for orienting meanings: in the f i r s t case contact with Sr is supposed, in the second contact with Or; thus both meanings are independent), the opposition 'bounded' vs 'unbounded' indicates that the movement is bounded at a certain movement by the Sr. All three oppositions a r e mutually independent and yield 27 meanings. Moreover, the motive meaning ( along with Sr) may be further complicated by an additional orienting point AdOr)which may be subjective ( ' t h e

31 orientation of the movement relative to the speaker') and objective ('the orientation of the movement in relation to a typical object of the external world - the earth's surface, mountains, rivers lakes, etc.). Vertical movement relative to the earth's surface is an example of an objective AdOr ('up' - 'down'; it is interesting that the zero meaning of this feature is equivalent to the absence of AdOr, which is not the case for a subjective AdOr). In Appendix 2 examples of interpretations of complex orienting and motive meanings are given (Numbering begins at the top of the tree; the branches of each node have been assigned the following values; upper 1, lower 2, middle, if there is one, 0). Any of the orienting meanings may combine with any of the motive meanings. In this calculus 130 orienting and 162 motive meanings are enumerated, thus the calculus gives 130 x 162 = 21, 060 spatial meanings. And this system exhausts neither the orienting nor the motive meanings. This, of course, doesn't mean that in order to discover how these meanings are grammaticalized in the target language it is necessary to ask all 21. 060 questions. It follows from the universal principle of correlatedness (the assumption that linguistic objects are constructed from standard compositions of elementary entities in the given target language) that in field work we are not searching for the presence of grammaticalized complex meanings, but for the presence of grammaticalized elementary meanings and the rules for combining them, which greatly reduces our search. Only an insignificant fraction of the potential spatial meanings is grammatically realized in a concrete target language, 23* the rest are related to optional meanings expressed with the help of periphrasis. Some meanings are so foreign to the usage of speakers of a given target language that it would be hard to find a natural linguistic equivalent for them (this doesn't mean that in another target language these meanings might not only be natural, but obligatory: for example, the category of 'article' in English, French, and German, the category of 'contrast' in Nanaj 24*, the category of 'logical stress' in Yukagir 25*, etc.). In view of the fact that this calculus was constructed for a concrete type of meaning certain elementary meanings which are not strictly speaking spatial (for example, 'subjectivity' vs 'objectivity') could be used to define it, and thus it is not necessarily the case that when the types of meanings are expanded the inventory of elementary meanings will grow proportionally. 3. 3. A Sample of Questionnaires of Diagnostical Utterances in the Mediator Language (Meanings in Noun Phrases, except Spatial Meanings) The questionnaire of diagnostic utterances to be given in the mediator language is compiled inductively by the following method. First we study the available means: (1) the segmentation of relations between objects existing in the language taken as the starting point of the analysis (in constructing questionnaires of this kind for this purpose we would use Russian, therefore we will refer to Russian as the starting language); (2) the correspondence between the segmentation of Russian and the segmentation of other languages which comprise a more or less representative aggregate of languages of different types. The means of study are the following. For point ( 1 ) : (a) the segmentation represented by the case system of Russian; (b) the segmentation represented by the

32 prepositional system of Russian; (c) the more minute segmentation registered in defining dictionaries; (d) the more minute segmentation registered in traditional grammars; (e) the more minute segmentation studied with the help of transformational methods. For point ( 2 ) : (a) the segmentation that is represented by the case and prepositionalpostpositional system of the language being compared with Russian (by grammars); (b) the comparison of this segmentation (2a) with the segmentation of Russian ( l a - e ) (by Russian-foreign and foreign-Russian language dictionaries). We can reasonably expect that with an increase in the amount of material there will also be an increase in the minuteness of the segmentation; at each stage the total segmentation will consist of the segmentations that have been made at the previous stage, combined with the segmentation of the language being studied. We can also expect that at some stage the segmentation will cease to become significantly more minute with each new comparison (the pro-ductiveness of comparing Russian with other languages will be strongly reduced). This will serve as a signal to finish the investigation. Clearly the unlimited introduction of new languages can expand our notions about the possible means of the linguistic segmentation of the real world, but for the purposes of the "zero cycle" these may be ignored since the text is not intended to collect a complete and exhaustive body of data, but is called upon to clarify the general structural tendencies of the language under study. Thus one of the tasks of investigation is the formulation of a set of the classes of utterances which hypothetically might (but not necessarily will) be reflected in the different grammatical structures of the language under investigation. Thus having chosen classes of utterances which do not differ on the level of surface syntactic structures in Russian: (a) "the book of the sister"; "the house of the father"; (b) "the smoke of the fire"; "a ray of sunlight"; (c) "a piece of bread"; "a glass of milk"; (d) "the warmth of a body"; "the freshness of a face"; (e) "aunt of the father"; "mother of the niece" etc., we hypothetically suppose that in the language under investigation (i. e. the target language) all or some of the classes (which in Russian correspond to one type of noun phrase) will correspond to various types of noun phrases. 4. Syntax Syntax is one of the most formal parts of grammar. It is not surprising therefore that it is on the syntactic level that the greatest similarity of languages can be observed. The modern syntactic conceptions of the theory of generative grammars offer a sufficiently developed formal apparatus for describing the deep syntactic structure. 27* It has been noted that the basis of the syntactic component in many languages coincides. The greatest syntactic variety is observable on the level of surface syntactic structures, described in terms of types of word combinations, types of sentences and types of syntactic bonds. As a rule syntax is given much less attention than phonetics of mor-

33

phology in field investigations which is understandable since understanding the phonetic and morphological organization of the language is a conditio sine qua non for any description of the language. As f a r a s syntax is concerned it is usually the syntactic problems of morphology which take precedence, for example, questions about the syntactic meanings of cases, prepositional and postpositional f o r m s the type of predicative syntagm (with the presence of non-nominative constructions, voice f o r m s of the verb etc.), the concord of tenses patterns of a g r e e ment (with the presence of the category of gender or a nominal c l a s s ) . These questions clearly do not exhaust the whole of syntax although they a r e of indisputable importance f o r certain languages. However, this situation is brought about largely because of the insufficient treatment of the typology of surface syntactic structures. Let us examine, as an example, the widely used method f o r the universal representation of a surface syntactic structure by a t r e e of dependences. 28* One of the most difficult problems in constructing a t r e e f o r a sentence in the target language is establishing rules to govern the placement of arrows. The traditional definition of subordination runs as follows: "the subordinating word connection is the formal grammatical dependency of one word upon another, which implies that the presence of the subordinate word (or words) in a particular f o r m (or in particular f o r m s ) is predetermined by categorial qualities of the head-word". 27* This definition, however, runs into a number of difficulties since it does not take into account several f o r m s of subordination that are poorly developed • f — '

in Russian. Thus, we find, the "inverse" subordination of the type X Y-a, where X and Y are p a r t s of the syntagm, a = a morpheme of word Y, subordinated by the word X, the dotted arrow - the direction of formal subordination, and the solid arrow - the direction of "natural" (from other Z

I

standpoints) subordination ("devuska uexal - a" "(The) girl l e f t " ) . ! t Inverse subordination characterizes the postpositional attributive coni—;— i i—!— ^ struction in Tajik ("duxtar-i xuSruj" "(The) pretty girl", "seb-^i duxtar" "(The) apple of (the) g i r l " ) , which is singled out due to its supposed syntactic unusuallness. A formal bond can also be established between words which do not enter into a relationship of "natural" subordination ( They do not depend on each other either directly or mediately). This type of bond we will r e f e r to as "indirect" subordination. It may occur in structures such as: * I * il n I if Z (in the sentence "eS scitajut boln-oj" "(They) her consider sick" the word "ee" indirectly subordinates the gender of the word "bolnoj". A formal bond may also be "double" when some word is f o r mally subordinated by two o r m o r e words (in the preceeding example we also have double subordination: the case of the word "bolnoj" is also subordinated by the word "scitajut"; double subordination is also found in sentences such as "dom, v kotor-om ja zivu" "( The) house in which l I live"). — ----------v.:.•» In some languages (Caucasian, for example) these f o r m s of subor-

34

dination are very widespread. There a r e numerous examples in Archi of: a) inverse subordination: :dija w - e r k u r s i w - i " "Father falling is" i_ ~ " ! the subject subordinates the class m a r k e r s in the predicate; b) indirect I 1 1 I subordination: "d-ez buwa k'an" "I (my) mother love" - the object "buwa" " m o t h e r " subordinates the c l a s s m a r k e r in the subject "dez" I 4H-l . "I"; c) double subordination: "dija buwa d - i r k u r s i w-i" "Father mother searching f o r i s " - the subject and object subordinate class m a r k e r s in -F 1f r the predicate; "mahra e-b-xuqi tu-t i icj" m a r r i a g e of conclusion day" (= "The day when the mariage will be concluded") - the object of the participle and the noun on which the participles depends subordinate the c l a s s m a r k e r s in the participle. It would seem that in deciding debatable c a s e s of arrow placement the history of the generation of the sentence should be given the most weight, i. e. , considerations of theoretical convenience which have as their aim the creation of the simplest and most uniform set of generative rules. The servant in a syntagm is considered to be the part which is maximally independent of the kind of construction in which the given syntagm occurs in a sentence, as opposed to the m a s t e r in a syntagm who, being dependent on some third word may be formally subordinated by that word ( o r may itself subordinate that word in the case of inverse subordination); the m a s t e r in such a construction may stand as a representative of the entire noun phrase. This substantiates the thesis (not always clearly formulated) of the necessity of delimiting the concept of the content syntactic bond from the formal means used to express it. Generally speaking these two a r e independent on one another. It appears that the most appropriate classificatory apparatus for the f i r s t stage of a syntactic analysis is the theory of the parts of the sentence which is fairly close to the tagmemic theory developed by several American field investigators. This theory segments the sentence in t e r m s of universal syntactic meanings without drawing any fundamental distinction between simple and complex sentences ( a distinction which again is related to the formal means of expressing syntactic bonds): these a r e meanings of subject, verb, object, attribute, adverbial modif i e r of time, place, goal, cause, etc. Such a classification f i r s t allows u s to choose the definite elements in the sentence and then to investigate the means of filling the associated positions with word classes, and also to choose the means by which each syntactic position can be formally expressed. This is the type of classification given in Appendix 6. Of course a classification in t e r m s of p a r t s of the sentence (or tagmemes) is generalized and does not yet give u s the means to discover all the nuances of syntactic relations (one position of attributive adjunct alone yields many concrete realizations, depending on the meaning of the modified and the modifier: of membership real o r abstract, full or partial, alienation or non-alienation, of quality, degree, quantity, partitivity,

35

of part or whole, of source, origin, authorship, cause, tool, means, and so on, in which case there may be formal means for choosing each of these concrete meanings in the target language). The types of syntactic meanings realized in noun phrases are contained in Appendix 3 30* which makes it usable for syntactic as well as morphological purposes although we cannot really say that the question of how to relate syntactic and morphological meanings has been finally solved. Perhaps it would be more consistent to describe these meanings in semantics (many concrete meanings are defined by the meaning of the words entering into a syntactic relationship), and though this question has hardly even been posed for insufficiently studied languages it would be undesirable to ignore the grammatical facts which are defined by such types of meanings even in the early stages of the description of the language. It is hoped that these grammatical facts can be described on a surface syntactic level in its own terms. At a certain point in the investigation it is necessary to shift to a deeper conception of syntactic structure since we need not only to inventory the facts, but to calculate them, that is, to propose a general procedure which will elicit any of the observable or potentially observable facts. As we mentioned earlier the theory of generative grammar, which has developed fruitfully in the past few years, gives us a means to describe deep syntactic structures; however in this short treatment we cannot devote further attention to the principles of generative grammar and must limit ourselves to this brief reference to their usefulness. 5. Lexicon In the very first stages of his acquaintance with the target language, beginning with the study of individual sounds, the investigator is inevitably going to come in contact with its words even if he is not interested in compiling a lexicon. Usually one can abandon oneself to chance and not bother to purposefully gather a collection of "useful" words. However, in the complex process of studying the target language, especially in the beginning, it is desirable to receive a maximum of effective information about the language. This applies to the lexicon just as much as it does to phonetics, morphology, or syntax. Thus the question arises of creating a minimum dictionary which will include the most important semantic entities, the ones which need to be discovered first. What principles should we be governed by in selecting the words which are most important in the early stages of studying the target language? Here are a few of them: (1) The usage of the word in the target language. The laws of the statistical distribution of words in a text, which have been developed from various material, convincingly show that the frequency of different dictionary entities fluctuates within a very wide range and that in any language a relatively small set of words makes up the largest part of any text in that language. Thus, for example, in Standard American English with a sample of one million word usages and a dictionary of 50, 000 words, the first most common 135 words make up 50% of the text, the first 510 - 60%, 1120 - 70%. 31* Therefore it is desirable first

36 to include the most frequent words in one's dictionary thereby making the investigation of the target language easier. The greater the number of familiar words in an arbitrary utterance, the more easily and effectively the utterance can be analyzed. Also loading one's mind with little used words in the beginning stages of the investigation only delays that pleasurable moment when the investigator begins to partially understand utterances in the target language. How can one define apriori the frequency of unknown words in an unknown language? Isn't this an absurd task? Not really, since the frequencies of words are essentially the same as the frequency of meanings and meanings are not a strictly linguistic phenomenon, but are a r e flection of the frequency of various situations in the communicative process. In order to be expressed in a text these meanings (situations) require words. To the degree that cultures are similar, the frequency of the meanings expressed in their languages will also be similar. In creating our minimum dictionary (see Appendix 5) we distorted reality somewhat by supposing that the culture served by Russian was representative, and used Russian minimum dictionaries 32*, making adjustments in obvious cases. ( 2) Ease of translation into the target language. In view of the well known facts that words are polysemantic and that the volumes of word meanings in different languages do not correspond, we can expect that the translation from the mediator language to the target language will cause the informant numerous difficulties whose character cannot always be predicted. This consideration is responsible for significant adjustments in the results received from ( 1 ) . In particular this principle provides a justification for the natural desire to include as concrete a lexicon in the minimum dictionary as possible which can be correlated with objects of the external world. Thus some investigators recommend the use of terms which represent cultural artifacts (man made objects) and parts of the body. 33* (3) The syntactic combinability of the words. It is important to choose words for the minimum dictionary which can be used to construct a fairly large number of sentences of varying difficulty. This means introducing words into the dictionary belonging to different semantic classes and capable of filling different positions in sentences. These two considerations argue against putting the minimum dictionary into list form. The most convenient form for a minimum dictionary is a composite of sentences - syntactically simple, clear in meaning, and contextually neutral, in the mediator language. This form is also preferable for morphological-syntactic reasons - the translation of a small list of sentences into the target language also gives rich primary information about its morphological categories and syntactic constructions. (4) Phonetic value. It is obviously a good idea to become familiar with the phonetics and the minimum dictionary of the language at the same time. Therefore the minimum dictionary should include a sufficient number of words in order to make the discovery of the great majority of the phoneme realizations in the target language possible. On the basis of average statistical data it is supposed that 200 arbitrary words (averaging 5 phonemes in length) in the target language will be enough for a first exposition of its ensemble of differential features and phonemes.

37 Thus in a Russian text, for example, of 1000 letters (we are disregarding the distinction between a letter and a phoneme) each letter will occur at least twice. 34* Even though this selection may be clearly more than adequate for discovering the phonetic features wich are contenders for differential status, it is no guarantee that the ensemble of phones and phonemes will be fully described, unfortunately this is something which simply cannot be guaranteed by any mechanical procedure. 35* Thus we can draw the expected conclusion about the independence of the results of the initial phonetic analysis from the dictionary. The fourth (4) principle while predetermining the approximate size of the dictionary places no limitations on the selection of the concrete meanings, while the first three principles (1-3) are content filters through which the selection must pass. 36* At various times for various reasons glossaries of words which have been considered important from one or another standpoint have been compiled. Already in the 18th century lists of sentences had been c r e ated for the initial study of the Caucasian languages. 37* In the last century Steinthal compiled a list of words and phrases useful for the study of an unfamiliar language, divided into thematic rubrics. A detailed thematic glossary, not oriented along the principles enumerated here, has been compiled under the direction of Cohen. 38* A thematic glossary of about 600 words has also been proposed by Whither 39* for the study of African languages with the note that a similar glossary had been compiled by Greenberg. 40* Swadesh used a list of approximately 1000 words in a comparative lexical study of 20 Indian languages. 41* In addition Swadesh has proposed a list of the 200 most stable (historically) words for lexico-statistical investigations. 42* In addition to glossaries there are interesting dictionaries of most important words which are recommended by methodologists for the teaching of foreign languages. I. A. Richards' well known textbooks contain a glossary of 500 words 43* and a vocabulary of Basic English ( 800 entries). 44* Also interesting is his list of the 100 words most useful for describing other words and covering basic ideas. 45* Palmer's glossary of 1000 words would be a good basis for a metalanguage if English is to be used as the mediator language. 46* At least two stages are conveniently separated in the collection of the lexicon in the field: the initial stage when the work begins "from scratch" and the advanced stage when the nucleus of the dictionary and several contours of the grammar are already visible. There will be a different minimum dictionary corresponding to each of these stages. In the initial stage it is desirable to work with words included in an elementary context. Appendix 4 contains such contexts (sentences) which include about 150 words, chosen in accordance with the principles (1-4) already outlined. Presumably this text will not be used dogmatically, but with certain paradigmatic substitutions of lexemes and grammatical categories so that after the investigation is finished up to 200 words will have been collected in the target language along with a variety of grammatical information for which this text serves only as a starting point. The informant first translates the entire sentence and then the indi-

38

vidual new words which are usually not difficult for him to extract from the sentence that he has just heard spoken. Note that in the initial stage of the work it is irrelevant what wordform of a given word the informant provides - the one he encountered in the text of the canonic form ( such as the nominative case of the noun or the infinitive of the verb in Russian). For the further collection of the lexicon the minimum dictionary in Appendix 5 can be used, which due to its size gives words out of context. 47* In this dictionary the entire lexicon is broken down into thematic groups which makes it easier to study the lexicon of the target language systematically (a thematic classification gives us a natural basis for systematically expanding the dictionary although the representation of each group in the dictionary is far from complete), and also reveals more precisely the semantic paradigmatic oppositions in the lexicons of the mediator languages ( This is important because the interpretations of the meanings of individual words are not given in the minimum dictionary: the thematic organization of words demands that the investigator has a more precise understanding of their meanings; see, for example, the group of words under the heading 'Weather' (Appendix 5, 3 . 2 . ) : inclement weather, blizzard, snowstorm, ground wind; under the heading 'Modality of Actions' (Appendix 5, 2. 6): obviously, apparently, possibly, perhaps, probably, it seems that). The entire lexicon is divided into four basic classes: (1) actions or states, (2) objects, ( 3 ) features, and (4) replacives for objects and actions. Actions-states are usually associated with the whole situation which may include objects or features, and in such cases it is convenient to group the objects and features along with the associated action-state. Thus objects and features which are comparatively independent of a situation are grouped separately in the minimum dictionary (Sections 1 and 2) and those defined by a situation are grouped with the reference to the associated action-state ( Sometimes, on the other hand, it turns out that an object or feature is in the lexicon while the action-state is not: under the heading 'Intellectual Impact' there is no verb corresponding to the noun truth, under the heading 'Physical Impact' there is no verb for the features straight, tough, etc.). Under each heading there is a more detailed classification (for example, actionsstates belonging to the heading of situations are classified as 'Being-Existence', 'Movement', 'Physiological Actions and States', 'Physical Impact', 'Work Activity', 'Sense Perception', etc.). The minimum dictionary contains about 1500 words, of which about 600 refer to actions-states, 500 to objects, 300 to features, and 100 to replacives. In using the minimum dictionary it should be kept in mind that it is impossible to give apriori dictionary definition of layers of the lexicon which refer to concrete realia, material and spiritual culture, and the specifics of geographical and climactic conditions. Therefore these categories are only indicated and illustrated by several words in the minimum dictionary; the concrete lexical content of the associated semantic fields is defined by the target language. In working with the minimum dictionary one should not pedantically strive to discover words in the target language which correspond to words in the mediator language since in many cases they don't exist. In this

39

connection we refer to the most common interrelations between words in two languages. (1) The basic (English) word may be split into several words in the target language ( f o r example, there into there above, there below, hand into hand from shoulder to elbow, hand from elbow to wrist, hand from elbow to the tips of the fingers, etc.). ( 2 ) The basic (English) word may be expressed grammatically in the target language ( f o r example, the Russian word mog1 'to be able' can be expressed by the potential mood, the Russian modal word pust' 'let' by the imperative, nel'zja 'must not' - by the prohibitive mood, the Russian words vsegda 'always' and obycno 'usually' - by aspectual affixes, the Russian words opredelSnnyj, neopredelSnmyi 'definite and indefinite' - by articles, etc.). ( 3) In the target language there may not be words ( o r word classes) characteristic of the mediator language ( f o r example, words referring to artifacts: stul 'chair', vilka 'fork', rubanok 'plan e', polka 'shelf', and so on, or words expressing situations which are untypical for the target culture: prodavat' 'to sell', krast' 'to steal', plyt' 'to swim', gresti 'to row', etc.). (4) There may be words which are typical for the target language and not for the mediator language (in such cases the translation is done periphrastically). ( 5) Several words in the mediator language may correspond to one word in the target language (for example, English as the target language may lack the different references to inclement weather which are found in Russian. ( 6 ) The most natural case of semantic correspondence in two languages is an incomplete coincidence of meaning areas. Therefore in collecting lexical items, especially if they are not denotative, it is desirable to have as many illustrative examples to the use of each word as possible: burja' 'metel', bar an 'snowstorm', v'juga, purga ('blizzard'), different types of prominences: xolm, vozvySennost', sopka ('hill').

NOTES 1* On the necessity of having a program for eliciting data see: S. M. Sapon, "A Note on the Gathering and Exchange of Linguistic Data", Orbis, Vm, No. 1, 1958, 83-85. 2* N. Chomsky, Aspects . . . , 28. 3* R. Jakobson, C. G. Fant, M.Halle, Preliminaries to Speech Analysis. The Descriptive Features and Their Correlates. Technical Report N 13, June 1955. 4* We may mention the following systems of notation laying claim to universality: the International Phonetic Alphabet: The Principles of the International Phonetic Association Being a Description of the International Phonetic Alphabet, L . , 1966; R. W. Albright, "The International Phonetic Alphabet: Its Backgrounds and Development", IJAL, v. 24, No. 1, part M, 1958; the generalization of the system accepted in American descriptive linguistics: B. Bloch, C.L. Trager, Outline of Linguistic Analysis, Baltimore, Md. VJ.942; L. V. Scerba's system ( f o r more detail see L. R. Zinder, Obscaja fonetika [ General Phonetics ], L . , 1960, 151) and those of Voegelin ( see F. M. Voegelin, "Guide for Transcribing Unwritten Languages in Field Work", Anthropological Linguistics, v. 1, No. 6, 1959, 1-28), Cohen (M. S. Cohen, Questionnaire

40

Linguistique, Nimègue, The Netherlands: Comitée International Permanent de Linguistes, Comission d'Enquête Linguistique, 1950-1951, 3356), and Reformatskij (A. A. Reformatskij, Vvedenie v jazykovedenie [Introduction to Linguistics], M., 1967, 173 ff.). 5* The exposition of these conditions is given in the author's joint publications with Kodzasov: A. E. Kibrik, S. V. Kodzasov, "Principy postroenija foneticeskoj transkripcii dlja pervicnoj zapisi tekstov neizvestnogo jazyka" [ Principles for the Construction of a Phonetic Transcription for a First Recording of Texts in an Unfamiliar Language ], Kuznecovskie ctenija, tezisy dokladov [ Conference held in memory of P. S. Kuznecov] M., 1970; A. E. Kibrik, S. V. Kodzasov, "Principy foneticeskoj transkripcii i tranckripcionnaja sistema dlja kavkazskix jazykov", [ Principles for a Phonetic Transcription and a Transcription System for the Caucasian Languages ] VJa, 6, 1970. 6* See A. E. Kibrik, "K voprosu o metode opredelenija differencial' nyxj>riznakov pri spektral'nom analize (na materiale glasnyx novogreceskogo jazyka) " [ On the Question of the Method for Defining Differential Features in the process of Spectral Analysis (on the Material of Vowels in Modem Greek)] , VJa, 5, 1962, 1. 7* See: N. Chomsky, M.Halle, The Sound Pattern of English, N . Y . , L . , 1968, 164. 8* The importance of this factor can be judged by the ability of deaf mutes to read lips. 9* See: R e C . Artikuljacija i vosprijatie [ Speech. Articulation and Perception] , under the direction of V. A. Kozevnikov and L.A. CistoviC, M. - L . , 1965. 10* In general it isn't always clear whether we are dealing with a mandatory coarticulatory effect or with a specific linguistic rule for the realization of a feature. Usually this involves "exotic" features rarely encountered and poorly described. 11* Anthropological Linguistics, v. 8, No. 2, 1966. 12* A subtype of this method is used in the matrix transcription where each feature is given a certain position in a matrix thus making it unnecessary to refer to the feature with a special sign. An example of this type of transcription can be found, for example, in: M.Halle, The Sound Pattern of Russian, 's-Gravenhage, 1959. 13* A number of writing systems based on a phonematic principle (Georgian, for example) use this method (a special grapheme corresponds to each phoneme). 14* This all relates to scientific transcription and does not touch the special problems of practical orthography which are brilliantly discussed in: N. J a . Jakovlev, "MatematiCeskaja formula postroenija alfavita" [A Mathematical Formula for Constructing an Alphabet ] , Kul'tura i pis'mennost' Vostoka [ The Culture and Writing of the East ], M., 1928, 41-64. 15* H. Sweet, The Practical Study of Languages: A Guide for Teachers and Learners, L . , 1964, 24. (reprint of the edition of 1899). 16* Here, in what follows, and in Appendix 1 it is assumed that the phonetic nature of the features we are referring to is known to the reader from courses in general phonetics. 17* The different meaning of identical signs in the phonetic, phonologi-

41

cal, and morphophonological transcriptions is expressed with the help of various brackets: [ . . . ] , / . . . /, and { . . . } . " 18* For example, the sign in R. I. A vane sov's phonological transcription for Russian ( R. I. Avenesov, Fonetika sovremennogo russkogo literaturnogo jazyka [ The Phonetics of the Modern Russian Literary Language], M., 1956). 19* See: P. S. Kuznecov, "Opyt formal'nogo opredelenija slova" [An attempt at the Formal definition of the Word ], VJa, 5, 1964; S. E. Jaxontov, "Metody vydelenija grammatiCeskix edinic" [ Isolation Methods of Grammatical Entities |, Jazykovye universalii i lingvistiCeskaja tipologija [ Language Universals and Linguistic Typology], M., 1969, 203-228. 20* See A. A. Zaliznjak, "LingvistiCeskie zadaCi" [ Linguistic Conundrums ], Issledovajija po strukturnoj tipologii [ Investigations in Structural Typology ], M., 1963; A. E. Kibrik, Izbrann.ye glavy arcinskoj grammatiki v zadacax [ Selected Chapters of the Grammar of Archi in Conundrums ], (in press). 21* On the classification of meanings into lexical and syntactic see: I. A. Mel'cuk, "O nekotoryx tipax jazykovyx znacenij" [ On Several Types of Linguistic Meanings ], O tocnyx metodax issledovanija jazyka [ Exact Methods in Linguistics ], M., 1961. 22* See: O.G.Rezvina, TipologiCe~skaja anketa (2). Kategorija grammatiCeskogo roda i klassa (Ucebnye materialy k speckursu po strukturnoj tipologii jazykov) [Typological Questionnaire ( 2 ) . The Category of Grammatical Gender and Class. (Instructional Materials for a Special Course in the Structural Typology of Languages)], IVJa, Laboratory of Structural Typology and Linguostatistics, M., 1967 (rotaprint). 23* An analysis (and its justification) of the realization of this system of meanings in the Caucasian languages is given in: A. E. Kibrik, "K tipologii prostranstvennyx znaCenij (na materiale padeZnyx sistem dagestanskix jazykov [ Towards a Typology of Spatial Meanings (on the Material of the Case Systems of the Daghestan Group) ], Jazyk i Celovek [ Language and Man] , M., 1970, 110-156. 24* V. A.Avorin, Grammatika nanajskogo jazyka [ A Grammar of Nanaj] , v. 1, M.-L. , 1959, 129. 25* E. A. KrejnoviC, "Sisterna morfologiCeskogo vyraZenija logiCeskogo udarenija v jukagirskom jazyke" [The System of Morphological Expression of Logical Stress in Yukagir ], Doklady i soobSCenija IJa AN SSSR [Reports and Communications, Institute of Linguistics, Soviet Academy of Sciences ] VII, 1955. 26* This list is a revised version of the questionnaire compiled under the direction of the author by M. Alekseev and A. Glozman. For the use of the questionnaire see Chapter 3, 1. 3. 27* See: N. Chomsky, Aspects . . . , Chapter 2 , 3 . 28* On the correlation of the grammar of dependences common to the European tradition with the grammar of immediate constituents characteristic of American linguistics, see E. V. Paduceva, "O sposobax predstavlenija sintaksiceskoj struktuiy predlozenija" [Ways of Representing the Syntactic Structure of the Sentence ], VJa, 2, 1964. 29* Grammatika sovremennogo russkogo literaturnogo jazyka f Gram-

42

m a r of Contemporary literary Russian], M. , 1970, 486. 30* For an attempt to classify surface-syntactic structures s e e : I.A. Mel'cuk, Avtomaticeskij sintaksiceskij analiz [Automatic Syntactic Analysis ] , Kibernetika v monografijax [ Cybernetics in Monographs ] , I.Novosibirsk, 1964, 17-27; A. K. Zolkovskij, Sintaksis somali (poverxnostnye i glubinnye struktury) [ The Syntax of Somali ( Surface and Deep Structures)], M. 1971, Chapter 1; I. A. Mel'cuk, Opyt teorii lingvisticeskix modelej tipa "Smysl-*—"tekst" [An Attempt at a Theory of Linguistic Models of the Type "Meaning-«-»-Text" ] , Chap. 2, M . , 1973. 31* H. KuCera and N. Francis, Computational Analysis of American English, Brown University Press, Providence, Rhode Island, 1967, 300-306. 32* È. A. Stejnfel'dt, Castotnyj slovar' sovremennogo russkogo literaturnogo jazyka [ Frequency Dictionary of Modern Literary Russian ] , Tailing, 1963; Ju. Ovsienko and others, 2380 naibolee upotrebitel'nyx slov russkoj razgovornoj reci [ The 2380 Most Used Words in Conversational Russian], M . , 1968. 33* See: J . H e n r y , "A Method for Learning to Talk Primitive Languages", American Anthropologist, 42, 1940, 637. 34* A . M . Jaglom, I . M . Jaglom, Verojatnost' i informaeija [ Probability and Information ], M . , 1960, 189. 35* See the arguments about the phonemic ensemble of Russian between the Moscow and Leningrad phonological schools which could not be r e solved by any selection (See: A.A. Reformatskij, Iz istorij otecestvennoj fonologii [ From the History of Soviet Phonology ], M., 1970). 36* Similar principles were formulated by Palmer for selecting the first lexicon for the study of a foreign language: H. E. Palmer, The Scientific Study and Learning of Languages, L. , 1917, 132. 37* These sentences are cited in: A. Cikobava, "K voprosu o pervyx tekstovyx zapisjax po gorskim iberijsko-kavikazskim jazykam" [ On the Question of the F i r s t Textual Transcriptions of Mountain Ibero-Caucasian Languages], VJa, 4, 1962, 100. 38* Steinthal: "Linguistik" in von Neymayer, Anleitung zu wissenschaftlichen Beobachtungen auf Reisen, Band n, Hannover, 3rd ed., 1906, 271-278. 39* M.Rioux, J . R o u s s e a u , J . - P . V i n a y , "Instructions pour une Enquete ethno-linguistique", Centre de Recherches d'Anthropologie Amérindienne de l'Université d'Ottawa, 1954. 40* W.Whithey, "Suggestions for Recording a Bantu Language in the Field", Tanganyika Notes and Records, 62, 1964, 1-19. 41* M. Swadesh, "On the Penutian Vocabulary Survey", UAL, v. 20, No. 2, 1954, 123-133. 42* In W . J . Samarin, Field Linguistics. A Guide to Linguistic Field Work, N. Y . , 1967, 220-223. 43* I.A.Richards, English Through Pictures, 1956; I.A.Richards, M. H. Ilsey, Ch. Gibson, French Self-Taught With Pictures, 1950. 44* I.A.Richards, Basic English and Its Uses, L . , 1943. 45* I.A.Richards, How to Read a Page: A Course in Effective Reading with an Introduction to a Hundred Great Words, L. , 1943. 46* H. E . Palmer, A Grammar of English Words: One Thousand English Words and Their Pronunciation Together with Information Concerning

43 the Several Meanings of Each Word, Its Inflexions and Derivatives and the Collocations and Phrases into Which It Enters, L. -N. Y. -Toronto, 1938. 47* This dictionary is a revised version of the minimum dictionary compiled under the author's direction by O. Sundukova.

3

ORGANIZATION OF FIELD WORK

1. Basic Stages of Field Work 1.1. Ultimate Aim As we indicated earlier one may pursue the study of an unfamiliar language for very different reasons. One's aim significantly predetermines the means which are used to optimize the study, the specification of the concrete problems which are to be solved in the process of the investigation, as well as the amount of time needed for it. The investigation of the target language without a clearly formulated aim can hardly be effective. Thus the first thing the investigator must do is visualize the ultimate aim of the work. This aim (which may consist of one or of several problems) should first of all be sufficiently universal so as not to be irrelevant to the study of the target language (for example, studying the case system of nouns In languages of the English type), secondly it should be substantial enough to occupy the entire period of field work, and third, its difficulty should be proportional to the amount of time available to the investigator for its solution. Clearly the formulation of the aim of the work is not an aim in itself but brings in a conscious selection of methods and means to be used in the study of the target language. In the previous chapter we talked about the "zero cycle" as a necessary link in the chain of field work: the preliminary selection of the linguistic material, glossaries, and tests which are to be used in the field. Experience has shown that working with an informant without a preexisting plan, a "scenario", of the conversation, is not very productive; the investigator should always have it in his power to control the conversation and direct it toward the solution of certain definite problems. Not only does one need to know the aim of the investigation, one should also have in mind the final form which the description which results from one's data is going to take, in other words, one should follow a definite linguistic theory in whose terms the target language is to be described. The idea that one can make an "unbiased" observation of linguistic facts and that these facts themselves will suggest a scheme for their own description is a serious error. There is no need to underestimate the influence of the data on the formulation of the theory, but in fact, in order to discover the deep regularities which lie beneath the data it is always necessary to have two things: the data themselves and some standard (theory) by which these data can be evaluated. This does not deny the fact that in the process of cognition, as a rule, a definite discrepancy will emerge between the theory and the facts, which will lead to a r e construction of the theory (this understanding of the process of cognition

46 is based on the heuristic and not the operational nature of cognition 1*). Therefore the denial of an apriori theoretical platform only leads to inconsistency, since such a position is impossible to achieve (the linguistic consciousness of even the most unsophisticated person, not to mention the professional linguist, is shot through with axioms, postulates, and dogmas about language and some theory will, perhaps even unconsciously, be included in his investigation, the difference of course being that since he does not acknowledge its existence the influence of his theory on the data will be impossible to control). Knowledge of the proposed final form for the description has a great influence on the methodology of the work and the type of linguistic data to be sought after. Without knowing the form of description it is unlikely that either the desired data j r the systematicity of these data will be achieved. Thus no matter how paradoxical it may sound, the greater part of field work should be done at home. Due to the unlimited variety of local problems which may be encountered in a field investigation it is hardly possible to propose a universal "recipie" for constructing programs of field investigation: in each concrete case the program is unique and unrepeatable. Nevertheless if we limit ourselves to several large classes of problems (for example, the synchronic description of one or several levels of the target language), then it is possible to choose some of the general properties of the programs for such investigations. A sample of such a program for the intensive study of the target language with the use of a tape recorder to fix all the activities with the informant was proposed by Voegelin and Eobinett. 2* A month's field work was broken down as follows into three parts: the first (7 two hour sessions, 8 hours of tape) - the study of phonetics, the minimum dictionary, the most simple texts, the second (14 two hour sessions, 16 hours of tape) - a paradigmatic interrogation, the choosing of accessory morphemes, the third (7 two hour sessions, 8 hours of tape) - checking the tapes, transcriptions, the organization of word lists texts, and paradigms. A significant part of the work is reserved for after the expedition (decoding the tapes). Generally speaking it is in the beginning stages of the investigation when nothing is generally known about the target language that the specific nature of field work and the universality of the investigator's activity are most evident; thus it is to this stage that we are directing most of our attention. How one prepares the program of investigation and how one conducts it depends a good deal on whether the target language is to be investigated individually or collectively. If the investigation is to be handled individually, all the investigatory functions are concentrated in the hands of one man, if done collectively there is a division of labor which introduces the problems of specialization and coordination. The aims of the investigation must be rationally divided into sub-problems to avoid excessive doubling. The solution to the entire problem should come as a result of solving all the sub-problems. Coordination of the work is very important when a team of investigators is involved since the individual components of the linguistic system are in reality closely related and the understanding of some concepts is often not possible without the knowledge of others and also because the data which are discovered always carry information

47

about many other facts in the language in addition to the information they were intended to provide. The difficulties of the collective method a r e redeemed by its operativeness and effectiveness which cannot be matched by a single investigator working alone. 1. 2. F i r s t Encounter with a Language Since we are dealing with an unwritten language the f i r s t stage of the investigation, naturally, should be devoted to phonetics, regardless of the ultimate aim of the investigation. In the previous Chapter we examined the question of fixing spoken sound in writing: the universal phonetic transcription. However, it is clear that knowledge of a system of r e f e r ences does not guarantee the ability to interpret the unusual sounds of a target language in t e r m s of these references. This ability is developed by the general phonetic training which every field linguist must undergo. How should one become familiar with the sounds and phonetic features of the sounds in the target language? When using the active method of studying the target language is not enough to simply listen to the language being spoken. Rather it is desirable to work purposefully with an i n f o r m ant. To simply ask the informant, however, "What sounds are there in your language?" is f a r f r o m being the best way of discovering those sounds ( m o r e about "direct" questions l a t e r ) . In Chapter 2, 5, we said that in principle any lexical material could be used for the discovery of phonetic features and sounds since in all languages they occur in limited numbers and a r e frequently repeated. We also said that it is desirable to extract the maximum amount of information from the language with each question and to attempt f r o m the very f i r s t phonetic investigations to accumulate as much lexical, morphological, and syntactic information as possible. Thus it is desirable to use a test s i m i l a r to that in Appendix 4 for the f i r s t acquaintance with the lexicon. 3* One need not be alarmed if the very f i r s t sentence causes great complications in differentiating the sounds and in transcribing. However, should one strive for an adequate transcription at every stage? One should strive, naturally, but without going to the absurd principle: not a step further until we are absolutely sure of the phonetic c o r r e c t ness of our transcription of the preceeding utterance. This applies not only to phonetics but to the other aspects of language a s well. The investigator constantly (quite often in the f i r s t stages of the work and then l e s s and l e s s frequently as the work p r o g r e s s e s ) encounters data that he can't possibly explain by any method other than "a stroke of genius" since he doesn't yet know the other facts which a r e required to explain that data. In these cases when all the available means of interpretation have been exhausted one should simply note the doubtful place and move on. It may and often does turn out that other data will present themselves which may be studied in combination with e a r l i e r information as a consequence of which new light will be thrown on the phenomenon which had e a r l i e r been incomprehensible and "strange". This happens particularly in the case where a complicated complex phenomenon will be explained by the discovery of its simpler constituents. In phonetics this dictates the principle of "selective listening" advanced by Nida. 4* F i r s t one should try to hear not words, but features and phones and at each moment listen to only one feature ( o r ensemble of features),

48 secondly, one should progress from simpler (from the standpoint of one's own perception) features and phones to more complex ones. Since an arbitrary word in the target language may have a differing degree of complexity, the degree of precision of the notation of different utterances may at first be uneven; precision will increase as the material and the knowledge of the features and phones increases and as the investigator's hearing and articulator^ apparatus become better trained. There may be two types of difficulty in perceiving: (1) the investigator hears a specific feature of the sound but cannot reproduce the sound; (2) the investigator can neither hear nor articulate the feature. In (1) the phonetic nature of the feature may or may not be known to the investigator. In (2) the investigator will simply not notice the feature until he runs into some difficulty with the notation (the appearance of a minimal pair, the native speaker's indication that the transcription does not correspond to the sound, etc.). The problem of difficulty in perceiving is related to the principally important question: how precise should phonetic "hearing" be? But this question has already been answered in formulating the principles for a phonetic transcription. The preciseness of recognition in the end result is defined by phonological considerations. If we were studying Russian we would encounter the phones [ e ] and [ £ ], which are opposed by the feature of raising, but to "hear" the difference between them is not phonologically mandatory as it would be in French for example. The same applies for the voiceless lateral in the word m8tl 'of brooms' [ m, otl] which we might encounter before a word beginning with a voiceless consonant. Thus having encountered a specific feature or sound in the target language does not mean that one is obliged to register it. During the course of the work it may turn out that some phones have been more and others less recognized than is necessary. This confirms the thesis that it is desirable to shift as quickly as possible to a phonological transcription. In field work the linguist practically never uses a pure phonetic transcription, but begins to 'phonologize' it from the first moment of his acquaintance with the language: "We can realistically expect a linguistic field-worker to have some notion of a phonemic system within fifteen minutes of beginning work and the main essentials within an hour or so, " claims Swadesh. 5* The investigator must decide still another question for his practice: how precise should phonetic reproduction be? It is generally the case that the ability to perceive speech on all levels precedes the ability to reproduce it.(This is especially true of the way children learn language: a one year old child knows several dozen words without being able to pronounce any of them, or is able to pronounce only several syllables which have a meaning equivalent to that of certain word.) Therefore the investigator usually hears more than he can pronounce. At the same time less precision is required to hear an utterance and still understand it than the precision required to pronounce it ( so that the pronunciation of such an utterance doesn't differ from that of a native speaker). The difficulty or even the impossiblity of getting rid of one's "accent" in learning a foreign language is well known. The pronunciation of a language is characterized not only by its useful features which are used as differential representations but by a great number of redundant phonetic features

49 as well. The latter are the features which must be precisely reproduced in order to speak without an accent. Of course, good pronunciation of the target language heightens the investigator's self esteem and his prestige in the eyes of the native speakers, but it is a luxury for interrogative purposes which requires much time and effort to develop. For this reason it is not at all necessary to strive for an "unaccentless" pronunciation. How precise the pronunciation should be is again dictated by phonological considerations: how should the phonetic representation of each phoneme in each position be pronounced so as to distinguish it from the phonetic representations of other phonemes which are opposed in that position? Despite the importance of the initial phonological analysis and the development of phonetic abilities in the first stage of the investigation one should not forget about the other aspects of the investigation of the target language. The expansion of the glossary, the discovery of sharper morphological features, and the basic syntactic constructions of the target language are going on concurrently with the transcribing of utterances. In translating them into the target language certain test sentences may be wholly or partially altered for phonetic or lexical reasons. For example, the presence of some unusual sound in the lexicon of the target language may cause us to seek out other words with the same sound, the segmentation of a word into morphemes which is not clear may bring about the desire to change some category within the word ( " M a t ' prinesla xleb "Mother brought the bread" — » " M a t ' prineset xleb" "Mother will bring the bread" —»-"Mat' nesBt xleb" "Mother is bringing the bread"), if the rules for word order are unclear we may want to rearrange the words in the utterance, if a difficulty in translating a certain word into the target language arises we will need to replace it with another word, etc. These deviations are not only necessary, but desirable. A careful study of the data which should be elicited by our test will require about 10-12 hours of work with an informant. A f t e r this 10-12 hours one should generalize from the accumulated information: establish initial inventories of differential features and phonemes, choose the positions f o r several of them which ought to be investigated, construct a phonemic system and discover the possible omissions in it ( f o r example, if the phonemes/t/, /k/, As/, Af / and the phonemes /t'/, /k'/, /ts'/, are in the language, then it is reasonable to suppose that there may be an unregistered phoneme Af'/), and also to settle debatable and unclear points in the system. We note that even if the framework of the phonemic system appears fairly rapidly, the individual "blank spots" can remain there f o r a long time ( see, f o r example, L. Ja. Sternberg's methodological aside on his study of Gilyak (Nivkh): "If I succeeded in penetrating comparatively easily to the g r a m matical secrets, such was not the case with phonetics" 6*), and the mere accumulation of textual material may not be sufficient to solve them. At the end of the first stage it is necessary to make the transcription simpler and more specific with the discovery of the phonemic ensemble and the relevant oppositions. Morphological and syntactic observations should be generalized as well. It is desirable to memorize the words from the minimum dictionary that have been discovered.

50 1. 3. Discovery of Basic Grammatical Categories Immediately on completion of the first stage of the investigation of the unfamiliar language ( 3-5 days) one can shift to the systematic study of grammatical categories. The morphological questionnaires of the "zero cycle" should help in this study. In Chapter 2 we examine some trial methods in the form of a calculus and in the form of diagnostic utterances ( see Appendices 2 and 3). Each of these methods has its advantages and disadvantages. Working with a questionnaire in calculus form means that the linguist is equipped with a metalinguistic deep representation of the system of meanings which makes his investigation a lot easier - the semantic components of each comples meaning are clear. However, for each meaning it is not always easy to find an equivalent in the mediator language which the informant is able to understand. Working with diagnostic utterances the form of the interrogation is outwardly simple: the informant is presented with diagnostic utterances in the form of easily understood sentences. However, the deep semantic meanings that lie beneath these diagnostic utterances are often not expressed since the classification of meanings which accompanies these utterances is tentative and imprecise. A questionnaire combining the advantages of both types would seem to be the most convenient. But no questionnaire should be used mechanically because when translating into the target language in practice numerous difficulties will arise and thus various inaccuracies may occur. The investigator should be constantly on guard and check each newly elicited utterance with the proceeding one, testing them for mutual non-contradictoriness. No matter how carefully a questionnaire has been compiled it cannot replace the creative element of the investigator in the discovery of the grammar of the target language. Questionnaires are only a helpful tool which increase the effectiveness of the investigation, springboards from which the investigator's imagination must leap; they do not have any meaning in themselves. In the course of the work questionnaires may be adjusted and expanded. Along with the use of questionnaires it is necessary at this stage to collect spontaneously produced utterances and short connected texts which we will discuss in more detail in Chapter 3, 3. The collection of paradigms is also extremely important at this stage. The dominance of the morphological element in the work does not mean that the syntactic and semantic elements should be slighted. On the contrary, at this stage the dictionary continues to grow intensively, the structure of the sentence becomes more expanded, and the syntax more varied. Syntax and semantics may occupy more or less space at this stage depending on the degree of special interest in them. Greater knowledge of the language leads to the discovery of its arsenal of word inflexional resources and then begins the most interesting and painstaking stage in the study of the target language, which, if we ignore the essential experimental nature of field work, differs little from the methods which are used in ordinary descriptive linguistics. The investigator discovers the most difficult points in the grammar of the target language and it is upon these that his attention is directed.

51

1.4. Interrogation According to the Following System: Hypothesis^ —»-Data^—- Hypothesis^ — - Data„—-

— - Elicited Knowledge

Some linguists consider that field work, which is conducted as a rule under strict time limits, requires a limitation in the number of stages for collecting and generalizing material. It seems to us that such a rigid limitation is undesirable and may actually lengthen rather than shorten the length of time spent in the field. The collection of data cannot be regarded as a mechanical process, since what is important is not the quantity, but the quality of the material. 7* By merely statistically compiling sentences and texts we have no guarantee whatsoever that thus we may lay the basis for plausible hypothesis. It is much more productive to work according to the system: hypo thesis^—»-data^—«-hypothesisg—>-data2—»-... —»- elicited knowledge. The first hypothesis about the essential qualities of the target language are the conceptions of substantive and formal universals which are expressed in the program of investigation and the questionnaires. The first investigation narrows the class of possible qualities which had been attributed beforehand to the target language and makes it more relevant. The addition of new data should continually revise the existing conceptions about the target language so that at each stage of the investigation there will be a hypothesis which can help to explain the elicited data. This hypothesis has a definite predictive power: it predicts the construction of new data the investigator has not yet encountered. The further collection of data continues, the predictive power of the hypothesis is taken into account and each time the hypothesis collides with new data it is necessary to check them both against the criterion of truth ( as we shall see, data can also be false). The collection of new data is regulated by the data which have already been collected through the explanatory hypothesis and continues until this hypothesis starts to predict new data accurately or the new data cease to alter the hypothesis. In principle this process is endless since the predictive power of the hypothesis is always increasing. However, it is reasonable to stop (and consider the investigation complete) when the elicitation of a fairly large collection of new texts produces a relatively small change in the preceding predictive hypothesis. 8* Of course in testing the hypothesis one should act with a minimum of passion, neither becoming the slave of the hypothesis nor trying to fit the facts to it. This is the same process the child undergoes in learning a language ( of course the child's hypothesis is constructed unconsiously, nevertheless the presence and change of such hypotheses is clearly manifested in his speech activity). Turning these thoughts on language into practical recommendations means that after each session with the informant the investigator should study the data he has elicited and adjust the plan of further work in accordance with the new data. Due to the increased corpus of data and the complication of the hypothesis as a result of the organization of the target language there is a constant fluctuation in the correlation between the amount of time to be devoted to working with the informant and the amount spent on the analysis of data. If more time is spent in the beginning with the informant, then gradually the proportion shifts to the time spent in ana-

52 lyzing the data ("Reviewing and checking the material require more time than transcribing it" 9*) and as a rule there comes a moment when the amount of data exceeds the amount of time the investigator has r e served for studying it and this is the moment when it is advisable to take a break from field work. After the data elicited in the field have been gone over it usually turns out that certain data necessary for the completion of the final model or the next stage of the model of the target language are lacking. It is thus necessary to return to the informant for help. 10* Data accumulation may proceed at various rates in concrete situations but experience shows that 20-30 days of intensive work are usually satisfactory for a first "satiation" of data. Thus in the process of field work it is highly desirable to construct intermediate models of a fragment of the target language, striving to fix them in written summary form and, if the work is a collective one, discussing them with one's colleagues, since the requirement of formulating one's knowledge in order to convey it to another person helps to make visible logical leaps which might have otherwise gone unnoticed for a long time. 1. 5. On Practical Learning of the Language Field linguists do not agree on the necessity of being able to speak the target language. Some assert that constructing a description of the target language is pointless if one is unable to speak it fluently, since otherwise the linguist is completely at the will of the informant-translator whom he can never completely understand. 11* According to others it is absolutely unnecessary to master the language since this in itself will not guarantee a high quality of scientific description: the native speaker who speaks the language beautifully can hardly be expected to make much of a contribution to its description unless he is equipped with a linguistic theory. Both these view points seem false when expressed in their extreme forms; it is better to specify the positive and negative aspects of mastering the target language than to make categorical statements. A mastery of the target language is useful in the following respects: (1) It allows us to work with the target language as a mediator language, that is, to work with a monolingual informant and also to work with a text without a translator. (2) The investigator can collect data by the selective "passive" method, observing the natural speech behaviour of the speakers and choosing the desired facts. (3 ) The investigator can subject himself to self observation which significantly shortens the investigative process. (4) The practical knowledge of the target language serves as an additional source of self assurance for the investigator that the model he has constructed fits the facts of the target language and also raises the authority of the investigator in the eyes of the native speakers who are not always convinced of the value of the linguist's investigation. Not attempting to gain a practical mastery of the language also has its advantages: (1) Since an explicit knowledge of the grammar does not follow from the knowledge of the language, the investigator need not expend a great deal of effort not directed toward the accomplishment of his direct aim. (2) The peculiarities of the structure of the target lang-

53

uage a r e Immediately evident to the investigator who does not know the language and he often notes facts which the man who has learned the language does not notice since at the level of practical m a s t e r y the rules which define these facts have passed into the investigator's subconsciousness. (3) Mastery of the target language is always relative and if the investigator is not critically aware of the stage of m a s t e r y he is at, he may p r o duce ungrammatical utterances and consider them to be grammatical (the danger of self delusion). In conclusion one may say that some m a s t e r y of the language (gained without special effort), which insures control over the informant's t r a n s lation and which enables the investigatior to learn, although not always fluently, new expressions, and to test these for grammaticality, is very useful in the scientific study of language 12*. However, there is no need to turn the practical m a s t e r y of the language into an end in itself. 2. Work with an Informant 2.1. The "Human F a c t o r " in Field Work As we mentioned in Chapter 1, 2, field linguistics is an experimental method of descriptive linguistics (types C and D in Figure 1), where the linguist is separated from the object of his study (the target language) and the means through which it manifests itself (the data) by a person (the informant) with whose help he elicits all the information about the object of his interest. It is difficult to overrate the fundamental importance of this fact. The only material the linguist has to begin with a r e the informant's grammatical utterances in the target language pronounced arbitrarily in a natural or assigned communicative situation o r stimulated artificially by the investigator. The ultimate aim of field work is to discover the g r a m m a r of the target language; the specific condition f o r accomplishing this aim is that an informant stands between the language and the linguist. However, the informant is not an automoton for the mechanical generation of grammatically c o r r e c t utterances in the target language, but a person who is used to the uncontrolled u s e of language for communicative p u r poses. On the one hand this is very valuable to the linguist (the linguistic behaviour of the informant is not complicated by preconceptions) but on the other hand it is a r e a l hindrance to the work (generating isolated •utterances which a r e directed toward demonstrating some linguistic fact is abnormal f o r an unsophisticated native speaker). The basic interest of the unsophisticated informant is concentrated on the content of the utterance and not its f o r m . This can lead to various difficulties in the work, for example, when the informant r e f u s e s to admit the grammaticallity (oral) of an utterance owing to its absurdity where its meaning is concerned (he fails to distinguish between semantic and g r a m matical correctness). It is an even c r u d e r methodological mistake to think that the informant can conduct a linguistic analysis or comparison of the utterances he has produced. Thus strictly speaking questions beginning with "Why" should not be put to the informant (such as asking the informant "Why in English is t h e r e the pair of wordforms ' shirt' - ' shirts' while there is only one form 'pants' ?", "Why is it possible to say in English "If I knew it now

54 (to-morrow)" but not * "If I knew it then (yesterday) . . . ", "If I had known it now . . . " ? " "Why does the accent in Russian shift to the ending in the indirect c a s e s of the word stol 'table' and does it remain on the stem with the word stul ' c h a i r ' ? " "Why are the words edva 'hardly' and Cut'-Cut' 'hardly' in Russian interchangeable in the utterances "On Cut'-Cut' xromaet, On edva xromaet "He hardly limps at all" but not in the u t t e r ances Qijize_edvajcromaet "He hardly limps at all any more", "On uze cut'-cut' x r o m a e t " ? " 13*, and so on). Although the investigator may frequently be tempted to ask such questions he will always be either d i s appointed or punished as a result. Even answering questions about the existence or the degree of grammaticality of an expression or a general characterization of its meaning (of the sort "Can you say this?", "Is this f o r m used in speech?") is often difficult for the informant. This is not to mention the difficulties created by the mediator language which the participants in the investigation may speak with greatly differing degrees of fluency resulting in the introduction of supplementary i n formational noise into the work. Special difficulties a r e involved in studying a dialect using its mother literary language as the mediator language. 14* The linguistic prestige of the mediator language may influence the i n f o r m ant to tailor his responses to the investigator's speech, especially when direct questions a r e asked (for example, after the question "How do You pronounce the word h e a r t h ? " one can expect the answer "Just like You do. " 15*) even though this may contradict the linguistic reality. One shouldn't forget that a person always remains a person; he gets tired, has unconscious lapses of attention, is psychologically predisposed towards certain kinds of activity, acts emotionally, has social prejudices, etc. , which often have an unforseen influence on the results of field work. But all these difficulties do not mean that one should forgo trying to a c tively influence the informant's speech behaviour; they merely warn the investigator and f o r c e him at all times to take the specific conditions of his work into account, which requires special methods f o r eliciting information and f o r carefully checking it. The linguistic field method is close to its anthropological peer 16*, the field study of ethnic and social groupings, and from this branch of anthropology we can get some ideas on procedural methods for working with an informant and the attendent psychological and ethnic problems. 17* In light of the above it seems unlikely that every native speaker will make a good informant. Just the opposite, a "good" informant is very r a r e l y encountered if we demand a whole complex of qualities f r o m him. In reality different tasks in field work require different qualities in the informant. Here a r e a few of them. (1) Knowledge of the target language. It is not always obvious whether the native speaker knows the target language well, especially if he has lived for a long time in another language environment or knows various dialects of the target language. Thus it is important to complete a q u e s tionnaire for every informant establishing his contacts with other languages and the period of time he has spent outside the given language community. (2) Knowledge of the mediator language. In a bilingual approach the investigator is really observing two kinds of linguistic phenomena:

55

the utterances in the target language and their semantic equivalents in the mediator language. In addition the latter may also be a means of governing the linguistic behaviour of the informant. In all these cases the degree of the informant's knowledge of the mediator language is very important for the effectiveness of the work. (3) Translating talent. Not every informant, even one who knows the target language well, can cope with translating connected texts.' This also applies to translating utterances in order to clarify their semantic specificity, where the linguistic instinct of the informant can make the work a good deal easier. (4) Pronunciation. Native speakers differ considerably in the clearness and quality of their articulation and this is an important quality in phonetic investigations (in both the initial study of the language and in special phonetic investigations) so that it is desirable to have an informant with an "announcer's" pronunciation. ( 5) The associative mobility of thought. When the informant is r e quired to generate a connected text in the target language (arbitrarily, by theme or picture) and also in spontaneous utterances by association it is important that he has a well developed imagination since a dearth of content usually leads to texts that are grammatically trivial. (6) Patience. Certain kinds of work (especially paradigmatic interrogation) and also the insufficiency of the investigator's memory (questions may have to be repeated) demand great patience from the informant. (7) Honesty and a lack of the feeling of "linguistic prestige". The situation often arises in field work (usually through the fault of the investigator) where the informant has trouble answering a question or doesn't understand it, and there is the grave danger that he will misunderstand his role: often the informant may assume that since he knows the language he should be able to answer any question in it, and will thus try to hide his difficulty. This can cause mistakes. (8) Strictness. A condescending attitude on the informant's part can be dangerous when the investigator is testing the correctness of his own use of the target language (pronunciation, producing grammatical forms, utterances, etc.) if it leads him not to find faults or correct the investigator for fear of insulting him. (9) Experience. Many of the above qualities are not unchanging and experience shows that the informant learns from working with the linguist. This may be very valuable. The informant will work more effectively after he knows the investigator's techniques and can understand at once what is required of him than he did at the beginning of the investigation. This has nothing to do with understanding the linguistic problem which lies behind the investigator's questions. The informant merely understands what reaction the investigator wants to elicit with his question. The methodology of questionnaires usually supposes that the informant should be "deceived": his behaviour should be so stimulated so that he does not suspect the true aims of the linguist and pronounces the utterance normally as though he were pursuing some other completely different aim and as though this were the way the investigator intended it. Of course this is not done for the sake of deceiving the informant, but just the opposite, to help him reply to the true sense of the question without his being aware of it.

56 (10) Linguistic unsophistication. Independent of experience is the quality of linguistic unsophistication, which is a v e r y positive phenomenon since the various linguistic dogmas (often f r o m very bad sources) which the informant may hold make the work much m o r e difficult. It is best if the informant holds no apriori classification of facts at all, a l though as the investigator is creating one the informant may perceive this and make the r e s t of the work e a s i e r by supplying facts which c o r respond to the classification ( Thus it is sometimes useful f o r the informant to learn the transcription on which the classification is based, the system of grammatical oppositions, word derivational models, and so on). Depending on the nature of the work to be done different informants may be chosen who a r e "good" f o r that particular purpose. In practice people who would make optimal informants a r e f a r f r o m easily found. One of the important conditions is that the informant has a lot of f r e e t i m e and this is not always the case with the people who have the right properties f o r the investigation. The question a r i s e s in field work of the number of informants that should be used in o r d e r to justify the extrapolation of the data elicited onto the target language as a whole. When working with only one informant one always runs the r i s k of describing an idiolect and not the language as a whole. However, trying to investigate as many informants a s possible is not always justified either, since all a i m s do not require the interrogation of a large number of informants and working with each new informant demands a certain amount of time to reach a mutual understanding, a certain amount of "rubbing together". It is desirable to have varied texts from different native speakers, but it is best to c a r r y out grammatical discovery with a limited group of informants who best meet the demands made of them. It is undesirable to work with two informants at the s a m e time (unless t h e i r functions a r e delimited - f o r example, an informant and an i n t e r p r e t e r ) since this often, if not always leads to the awakening of "linguistic prestige" with all its accompanying and unpleasant consequences (arguments, reticence, tendentious behaviour, and so on). It may turn out in working with two informants that the forms they have given differ in one way or another, which does not always indicate a mistake in the transcription (although it may), but may be the result of age or dialect differences, differences of full or colloquial style of speech, and also of the presence of doublet f o r m s in the language. 18* The reasons for these divergences should be cleared up and then a d e cision can be made a s to whether to include them in the description o r not. 2 . 2 . Eliciting of Linguistic Data Field conditions a r e usually such that the linguist has the chance to work in the natural linguistic environment. A desire to avoid the difficulties enumerated in 2 . 1 . may lead to an acceptance of the unbiased approach where nothing is forced on the native speaker and the linguist merely observes the behaviour of the speakers as scientific. We may call this the "passive" method. The methodological purity of this method is u n arguable. However, it seems to u s that the battle f o r a pure method should be waged in a different direction and that the active method of

57 eliciting data should carry the most importance in the bilingual approach to a target language. The active method broadens the linguist's capabilities and helps making possible a more effective study of the target language than the passive method by which the p a r a m e t e r s of a needed u t t e r ance cannot strictly enough be assigned. The basic difference between the active and passive methods is that with the f o r m e r the investigator consciously directs the speaker's linguistic behaviour. This direction can be accomplished by linguistic and non-linguistic means. Non-linguistic influence on speech behaviour is close in spirit to the passive method (though its non-linguistic c h a r a c t e r is only relative): gestures are used (though here there is always the danger of using a gesture which has a special interpetation in the given culture), assigning a speech situation by one's own behaviour (for example, a demonstration that one has a headache), o r the use of pictures which depict some situation. The last method is widely used by some linguists. Generally speaking this method influences the theme of the informant's utterances, the theme in t u r n defines the aggregate of lexical entries. The picture also acts on the s p e a k e r ' s imagination. For example, Hayes recorded the following spoken reaction to the picture of a dog sitting under a table 19*: "Here, we the old t i m e r s never knew what a table was, to eat on. We the old t i m e r s used bowls that had been made, and when the food was cooked in the middle of the room, we ate there. Now the younger generation's eating is done on a table, so called. Now as f a r a s I can see there, this dog is standing under here. The dog seems to want to eat bread, the old t i m e r ' s bread . . . " The picture method can also be used for eliciting some types of grammatical meanings. 20* It is likely that this method can be f u r t h e r perfected. 21* Nevertheless the linguistic regulation of.the informant's speech behaviour, that is, the use of commands and questions in the mediator language and the testing of the grammaticality of utterances in the target language, remains the most powerful tool. The following methods for working with an informant may be specified: (1) Translation f r o m the mediator language into the target language. This was the basis of the "zero cycle" programs discussed in Chapter 2 and may also be used for the special study of individual fragments of the g r a m m a r . This method assigns the semantic information in the mediator language which is to be expressed in the target language. As we noted earlier, both grammar and semantics can be investigated with this method. It is always desirable when working with this method to ask f o r a controlled retranslation into the mediator language a f t e r eliciting an utterance in the target language, since quite frequently the u t t e r ance-stimulus and the utterance-response do not mean the same thing ( f o r a variety of reasons). "The worst example I know of such an e r r o r was in the case of a native's sentence purporting to mean "Thou shalt not take the name of the Lord in vain", but which when retranslated into the contact language turned out to mean "Thou shalt not pay attention to the words of the Lord. " 22* (2) Paradigmatic interrogation. Where there is sufficiently rich word inflexion ( a s in Alutor where the verb has word f o r m s of the o r d e r 10^) and especially if the word inflexion is accompanied by a variety of d i f f e r ent morphophonological rules, one cannot be satisfied with word f o r m s encountered at random, elicited in sentences and texts. Therefore a f t e r

58

the basic categories of word inflexion have been discovered it is expedient to start an intensive investigation of paradigms, which are elicited, as a rule, out of context. This kind of paradigmatic interrogation is possible after the informant has acquired the habit of separating the needed linguistic entities. ( 3) Syntagmatic interrogation. This method of substitution is widely known in Bloomfieldian linguistics and is used, for example, when the morphological segmentation of an utterance is analyzed: the elements of meaning in a fixed utterance are substituted for in a consistent manner, the change in forms is constantly observed, and a correlation is established between meanings and forms. This isn't always a simple method; thus, for example, it sometimes happens that what appears to be an elementary meaning from the investigator's point of view is a complex meaning in the target language, and this causes significant changes in the utterance in the target language. (4) Cross interrogation. Sometimes in order to eliminate the undesirable relationships that arise between questions (especially where paradigmatic interrogation is concerned) it is desirable to alternate different types of questions. The informant must answer each question independently of the preceeding one. This method is also useful when a mistake is suspected in the informant's answers, since it is highly undesirable to directly expose him making a wrong answer. ( 5) The associative interrogation. This method is used when one works from a text or when the previous utterance one has elicited is not entirely clear. For each such utterance the attempt is made to discover all its elements and their meanings which in turn generates a number of new utterances which in turn have their unclear points and so on until all the unclear utterances which result from the analysis have been exhausted. 23* Associative analysis is always present to a certain degree, but is especially pertinent when certain of the contours of the grammar and the basic lexicon are already known (the desire to understand everything in each utterance in the beginning stage of the work will cause the investigator to lose track of the main direction of the study since there is too much about the target language that is unknown). ( 6) Periphrase. Sometimes when investigating the syntactic transformational capabilities of the target language, the semantic entities, and the grammatical categories, it is desirable to ask the informant to try to express the same idea in a different way. (7) Leading question. It is impossible to ask directly about many of the phenomena in the target language that are of interest to the investigator ( especially rarely used and variant forms) without disrupting the naturalness of the native speaker's spoken response. This disruption distorts the real situation. In such cases leading questions are used. The perfection and effectiveness of this method depends on the investigator's linguistic and psychological instincts and his field experience. 24* ( 8) Eliciting examples. Experience has shown that informants usually learn quickly to think abstractly about linguistic objects and when this is the case they can be asked to give examples of a phenomenon assigned by the investigator: the meaning of a word, a grammatical form, a situation, etc. This method, similar to the leading question (7), yields

59 spontaneous utterances not related to a direct translation from the mediator language. (9) Stimulus with corrections. When the investigator has a hypothesis he can construct several utterances according to it in the target language and test their grammaticality. This applies to the phonetic as well as the grammatical level. All these methods can be combined in various ways; none are mutually exclusive. As we mentioned earlier it is inevitable that one will run into all kinds of mistakes in field work. Either the informant or the investigator can be responsible for these mistakes. When transcribing and translating a text the investigator may be imprecise, may overlook facts because of his insufficient knowledge of the language. A mistake may be due to the way a question was asked, preventing the informant from answering in an non-preconceived manner (for example, the questions: "How do you produce that sound?", "When do you use that verb form?" 25*) and also due to a lack of mutual understanding between investigator and informant. The following factors may contribute to mistakes in working with an informant: (1) The attempt to get a literal translation in the target language which destroys grammatical and semantic naturalness. (2) The informant's adaptation to the "ignorance" of the investigator resulting in the conscious simplification of his speech. (3) The pressure of the paradigmatic interrogation - outside of context, especially when there are variant forms, the informant will produce a form analogous to the preceding one he has given. (4) A mistaken expectation - if the informant expects one question and gets another. ( 5) Non-linguistic factors - the informant, like everyone else, cannot work without respite and will make a certain percentage of errors. What methods do we have for verifying data? There are (a) the cross interrogation ( see 2. 2 . ) ; (b) a controlled check of data by the interrogation of different informants; ( c) a controlled check of the data and description of the same theme as studied by different investigators; (d) the pressure of the system - when contradictions are found in the data starting from a descriptive hypothesis; (e) a deeper knowledge of the language - the more the linguist knows about the language the easier it is for him to notice a mistake in the data and not to make it himself. 3. Collection of Linguistic Data 3.1. Corpus of Sentences The most natural pronounced entity and the one that should be used for analysis is the sentence. Of course the smaller constituents of the sentence will also be analyzed, but as a rule they figure within the framework of the sentence. It is the easiest way to elicit the smaller constituents from the informant in the form of a sentence. If stationery is available when one works with the informant it is convenient to use a large notebook for grammatical transcriptions as this allows the investigator to keep a lot of material in

60 front of him, otherwise a pad is best. During the work various types of sentences come up, stimulated in various ways. In o r d e r to evaluate the grammaticality of a sentence at some future date it is desirable to know how it was elicited and a note to this effect should be made for each sentence. From this point of view there a r e the following types: (1) Spontaneous (isolated utterances, pronounced naturally by the informant, not r e s t r i c t e d by an initial sample in the mediator language). (2) Translated f r o m the mediator language. (3) T r a n s f o r m s (coming as a result of the partial transformation of a sentence or sentences previously elicited). 26* Spontaneous sentences may not depend on the investigator at all (the informant may offer them voluntarily) or they may be either examples of a semantic (the meaning of a word) or grammatical (the use of a certain tense, a phonotactic regularity) fact. Although sentences a r e usually collected on one given linguistic phenomenon, it usually turns out that they are interesting f r o m more than that single point of view. It is, however, not always easy to discover immediately what these other points of view might be. When the material is recorded only in notebooks this makes the subsequent s e a r c h f o r examples of a needed phenomenon very difficult and time consuming. And if one investigator is still able to find his way around in his material (and only within certain limits) then it is extremely impractical in collective work to store the data in notebooks: it is practically impossible to make use of someone e l s e ' s m a t e r i a l s if they a r e in a notebook. Thus a r i s e s the task of creating a single corpus of sentences which can be conveniently sorted and classified many times. It is most practical to store this corpus on punch c a r d s which a r e a convenient technical means f o r marking the features satisfied by each sentence. The features should be sufficiently minute o r general so that each feature defines a relatively small number of sentences in the linguistic corpus (so that they can be sorted quickly by hand) - in any case not m o r e than 100, but not too few; otherwise the number of features will be too great and difficulties will present themselves with marking (this does not apply to those f e a t u r e s which w e r e used apriori to divide the entire corpus into a limited number of c l a s s e s - 2 , 3 , 4 ) . The most convenient type of punch card for storing sentences is the K-6 type, 105-147 m m . , with 74 double perforations. The blank area on these c a r d s is the same as that of an ordinary catalogue card. Each sentence with its translation is written down on an individual punch card and is given an ordinal number so that it can always be r e f e r r e d to and found in the corpus. Then the card is marked with a punch in accordance with the f e a t u r e s relevant for the sentence recorded on it. The great convenience of a corpus of punch cards is that sentences of a certain type can be searched f o r mechanically: a metal needle is placed through the perforation corresponding to the needed feature and all the punched cards containing the feature fall out. Sentences can also be sought by a group of features (i. e. sentences which satisfy a number of f e a t u r e s ) . With a complex (multilevel) investigation of a language where there is a large corpus of data punch cards allow the investigator to consider all the data which has been elicited (this is even m o r e true of a collective work). In Appendix 6 a sample of a punched card developed for u s e with

61

an unfamiliar language of the Caucasian type is given (knowing the language type had some effect on the system of features marked on the card, for example, the choice of the ergative construction, the case system); in addition many of the features treated universally on this card are interpreted differently in each concrete language. The features are chosen so as to make the corpus suitable for both syntactic and morphological analysis. Phonetics and semantics are also presented in a fragmentary manner. These features are examined in Appendix 6 (the majority of the features are given in a straight code, with the exception of the number of the sentence or text which is given in an ordinal selector code). The features chosen in the present system are sometimes partially r e peated (for example, a simple and a complex sentence are marked separately, and in addition different types of complex sentences are separatedly marked: the feature indicating the presence of a dative construction is marked separately from that marking the presence of the dative case, and several others). This can be explained by the fact that the present system was not intended to be logically invulnerable, but practically useful so that the required information could be extracted with a minimum of sorting. Note also the fact that in a number of cases the feature that is marked on the deep perforation is a species (i. e. subfeature) of the feature marked on the shallow perforation, for example, "dialogue-text", "semantic example-general example", "cause, expressed by a complex sentence-cause in general", etc. Sometimes these pairs are logically independent, for example, "genitive case-dative case", "comparison-equativeness", and so on. Such features are combined in pairs out of considerations of economy and if a feature on the shallow perforation interests us we have to first pull all the examples with a deep cut and then all the examples with a shallow cut in order to get to it. 3.2. Corpus of Texts No linguistic investigation can be content with only specially elicited isolated utterances. Natural connected texts are also an important source of information about the target language. These may be canonic texts (tales, parables, anecdotes, legends, and other sorts of folklore) as well as text-improvisations (happenings from real life, biography, stories about professional activity, customs and ceremonies, stories inspired by a picture, etc.) Texts which demonstrate maximally unregulated human speech are interesting as well as those pronounced specially for the investigator. It is these dialogues or conversations between several persons which reveal the use of the target language in speech. It is desirable to record text-improvisations on a tape recorder at their normal speed. Canonic texts can be recorded from speech if the informant is also the translator, otherwise a tape recorder is the handiest way of making a first transcription. Once he has some knowledge of the target language the investigator can transcribe from the tape recorder by himself, however, it is more convenient to use a native speaker who need not speak the mediator language, but understands his function (Children can also serve as assistants, although helping with the transcribing demands great patience from the informant). Usually it is easier to transcribe the text if one understands it, but in principle the process of transcription can be separated

62 from the process of translation, especially if the work is being done by different investigators and/or with different informants. The transcribed texts are then translated. The basic steps of translating are: by morphemes, by words, subscript, and literary, although it is hard to make a strict enough distinction between them: " . . . there are as many degrees of literalness and freedom of translation as there are levels of hierarchical structure in the passage. " 27* Creating a corpus of texts poses two basic problems: the problem of transcription levels and the problem of translation (of course this also applies to the corpus of sentences), each of which has two aspects how to carry it out and how to present it in final form. The transcription process is based on the principles discussed in Chapter 2 , 1 . ) . The stages of the translation process have been described in detail by Voegelin, 28* who envisages an initial segmentation of the sentence, the identification of morphemes, the word for word translation of the text, the elimination of redundant elements in the translation and the addition of new elements where necessary, translating with regard for the word order of the mediator language ( which sometimes requires significant reordering of individual words), the placement of punctuation marks. The final form the translation is to take deserves special attention and has been decided in different ways by different authors. As far as the transcription is concerned it is obvious that the phonetic level is unacceptable since it records an abundant variety of surface structures which makes it difficult to see the underlying common units upon which they are based. There are arguments over whether it is best to use a transcription based on the morphophonological or the phonological level. The phonological level is convenient because the rules for shifting to speech are fairly simple and the reader who is not familiar with the target language can easily learn them. The inconvenience is that, as a rule, it is impossible to segment words into morphemes on the phonological level, since some morpheme junctures cannot be outwardly shown, for example, by a hyphen. In a morphophonological transcription the morphemes are given in a maximally uniform manner and it is not complicated to show the segmentation of the text into morphemes, but in order to read such a transcription the reader must know a lot of rules which are often scattered throughout the entire description. This makes the use of the text more difficult since we know that the perception of a written text is much harder for a reader who is not able to "speak it" to himself. Thus it is a good idea to present the text in both forms, or if this is not possible, in a compromise variant on both levels with an orthography which gives a series of conventional spellings. The problem of translation is even more difficult. Often texts are accompanied only by a so called "literary" translation, which makes the work of the reader who wants to familiarize himself with the texts very difficult even with a dictionary. If we are going to translate for the reader who does not know the target language then we should make it so that he will be able to segment a text in the target language into words and morphemes and to correlate them with the elements of meaning contained in the translation. Obviously there are a number of theoretical reasons why this cannot always be done in a literary translation, since part of the information has been lost and part of it added. Therefore

63

either the intermediate level (or levels) of the translation or commentaries should be given. In view of the relationship which exists between language and culture extralinguistic as well as linguistic commentaries may be required. Compiling a corpus of texts is fairly time consuming. Some of the work, for example, the transcribing (if the transcription rules are properly formulated and the informant knows them well) and the initial translation (which the investigator checks afterward) can be done by the informant. Only a sufficiently educated informant can provide this kind of help. However, even a less qualified informant can help in c r e ating a corpus of texts. For example, there is the technique that uses two tape recorders. The first tape recorder reproduces the basic text which is copied sentence by sentence onto the second tape recorder; after each sentence the informant records the individual words, their translations, and a translation of the whole sentence. 29* This way the investigator has some definite knowledge about the target language and he can process his prepared text at home. The corpus of texts serves as the material to conclusively check the correctness of the hypothesis formulated from the analysis of isolated and usually artificially stimulated sentences. 3. 3. Corpus of Lexemes The most time consuming task is the compilation of a corpus of lexemes the dictionary). This is due to the number of entries which must be collected, the difficulty of eliciting them, and also the variety of information which should be attributed to each entity. Lexemes may be collected "at random" or by a program. The problem of collecting lexical material is resolved in various ways depending on the aim of the lexicological description, but one must not be mislead into thinking that collecting the lexicon is necessary only for semantic investigations. Even a morphophonological description is impossible without a sufficiently large glossary to reveal the variety of positions in which given morphophonological rules appear or are realized. A program of interrogation and methods for eliciting lexemes is e s sential in a special study of the lexicon of the target language but since it is impossible to foresee everything, one must unavoidably add those lexemes that one encounters "at random". The method of "thesauruses" is convenient for collecting the lexicon: by the choice of definite semantic groups of words (which have a common semantic feature) and their systematic elicitation.30* The minimum dictionary (Appendix 5) is constructed according to the method of thesauruses. The usual method is to begin with the mediator language, although it is precisely here that the investigator may expect to be confronted with many surprises: in order to avoid them it is desirable to have a metalanguage for semantic description which might take the form of a defining dictionary of the mediator language. The essential contours of such a dictionary have been sketched by Ju. D. Apresjan. 31* This dictionary would give the meaning (or meanings) of every word in the mediator language in an obvious form and would make the task of correlating a word in the target language and its equivalent in translation a good deal easier (Present defining dictionaries are most uniformative in this respect). As soon as the word has an interpret-

64 ation it will at least become evident what needs to be discovered in a word in the target language in order to place an equal sign between them. However no such dictionary exists and the attempt to construct one on the spot while working on other problems would result in excessive simplification and vulgarization. Thus the field linguist usually tries to collect as many relevant contextual usages of each lexeme as possible in order to clarify its meaning without unnessarily complicating the problem. Clearly this method is far from perfect, but then theoretical linguistics, however surprising it may sound, has given the field investigator a very skimpy scientific apparatus for understanding the semantics of an unfamiliar language. Thus one of the essential needs of Russian field linguistics is the creation of a scientific defining dictionary of Russian. The difficulty of lexical work is that the word is defined not only by its linguistic context, but by its cultural context as well: " . . . the fundamental problem for the field worker is the fact that he must not only translate language but inevitably must translate culture", 32* the relationship of culture to language affects not only lexical but grammatical meanings as well. Nida gives the example of two "possesive" systems in New Caledonia: "intimate" and "non-intimate". 33* The words mother, liver, descendents belong to the first system, the words father, heart, and personal life belong to the second. Such an opposition is unexplainable if we start from the meaning of the given English words and is related to the world view of the people who speak the target language: the society is a traditional matriarchy which regards the liver as a symbol of the whole person and where offspring have a more intimate and lasting relationship to the person than does his own life. The study of the lexical combinability of words makes up an essential part of a lexical investigation. The lexical functions of Mel'cuk and Zolkovskij 34* may serve as a formal apparatus (a language for the description and also the eliciting of data). These functions may be regarded as a special lexicographical questionnaire guaranteeing sufficient fullness and consistency to the description. 35* Working with such a questionnaire requires great skill and caution since obviously one cannot ask the informant to perform Operi from the word step or Funco from rain. It is likely that when the techniques of interrogation corresponding to this method are worked out they will help considerably in the discovery of lexical combinability, which is difficult to do when one works only from a text or when one takes questions at random. It may be that the investigator will run into trouble when trying to describe a certain meaning in terms of lexical functions because he is not sure that he completely understands the meaning. It seems to us that the fundamental value of lexical functions in field work is not so much in ascribing them to word usages in the target language as in discovering what these usages are. Though we did not set out purposely to introduce them, the minimum dictionary (Appendix 5) contains several derivative lexical functions. We list several of them here: S Q : 'derived substantive' - (in the thematic groups'Being, Existence', 'Movement', 'Mutual Action', etc.), S^: 'name of the subject' - (used in the groups 'Intellectual Impact', 'Work Activity'), S_: 'name of the object' - (in the thematic group 'Intellectual

65 Impact'), S i n g t : 'that with which something is done' - (in the groups 'Intellectual Activity', 'Work Activity'), S ^ : 'the place where' - (in the group 'Work Activity'), S r e g : 'result' - (in the groups 'Possession', 'Intellectual Impact'), A^: 'type property of the subject' - (in the groups 'Physical Actions and States', 'Mutual Action', and others), A^: 'type property of the object' - (in the group 'Physical Impact'), Adv: 'derived adverb' - (in the group 'Time'), Caus: 'to cause a situation', 'to create' - (in the groups 'being', 'Movement' and others), Func^: 'typical way of functioning' - (in the group 'Weather'), Anti: 'antonym' - (in many of the groups in the section 'Features'). This process of collecting the lexicon is completed with the compilation of a dictionary. The basic types of dictionaries are: an alphabetical target language-Russian dictionary ( Russian is the equivalent of the metalanguage); an alphabetical Russian-target language dictionary; a thesaurus 36* (the words are organized not alphabetically but by thematic groups 37*; thesauruses can be more or less detailed); a reverse dictionary of the target language (word order is alphabetical but beginning from the end of the canonic word or word form; such dictionaries are convenient for grammatical and phonotactic investigations 38*). The information collected about the lexeme is determined by the form the lexical article in the dictionary is to take. Therefore it is best to take a look at this question. Existent dictionaries are as a rule unsatisfactory from the standpoint of the role they can play in a description which is constructed as an active model. All the information needed for an active model should be collected in the dictionary: morphological, syntactic, and semantic, which is defined by concrete lexical entries, in order that in conjunction with the grammar the reader can construct (for a given meaning) grammatical sentences in the target language. The fact that this ideal is unattainable at the present time does not mean that we should not be working toward it. The form of this dictionary is developed within the framework of the "meaning text" model and can be used for any type of dictionary (including bilingual ones). A lexical article in the dictionary should consist of the following parts: (1) The title word-translation. For each target language one must decide the initial lexical form of the word. This may be some word form of the word, a composite of word forms, a root, or stem of the word. In some cases it may be more practical to introduce an abstract form of the word from which its concrete word forms can be conveniently drawn. The translation of the word into the descriptive language is the interpretation. It is appropriate to note that an idiomatic translation is not always convenient. Let us examine the following Archi utterances with their translations: "gudu laha os bala abcuqi" "That fellow brings only unhappiness", "Qurum abcas" "Tokindle the f i r e " t "k'alumli raq acas" "To draw a line with a pencil", "zon surat arcarsi wi" "I am drawing ajpicture", "naqb acutut harq" "The roof is covered with earth", "zimiz abcutut qal"J'The skin is covered with goose pimples", "zari ak' k'umk'umiak arcarsi i" "I'm putting meat in the pot", "zari ak' k'umk'umlas arcarsi i" "I'm taking meat out of the pot". The same word

66 " a c a s " i s translated by the words to bring, to kindle, to draw, to cover, to put in, to take out. An enumeration of these words as interpretations of the meaning of the word acas will not reveal its semantic nature. We might guess that this word has no independent meaning at all, but realizes a certain lexical function expressed differently in our metalanguage. This function we may define a s Caus Func 0 : 'to cause something to function', f o r example, "To cause meat to be in the pot/out of the pot" - it turns out that the concrete meaning to 'put in-take out' does not belong to acas and the paradox is averted. (2) Morphological zone. Here we find indicated: (a) the word inflexional type ( f o r example, the part of speech, the type of declension, and so on); (b) stems not combined by the rules of word inflexion (if t h e r e a r e any such s t e m s ) ; (c) non-standard indications (suppletism, i r r e g u l a r formation of individual word f o r m s , incomplete paradigms, and so on). (3) The syntactic zone - the active valences of the word, the means of filling positions. Examples of each model of government a r e given. (4) Lexical zone - the lexical combinability of the word (with examples) . ( 5) The thesaurus zone - the semantic classification of the word, generic and several specific words according to different semantic features. (6) The contextual zone - examples of word use not belonging to the lexical o r syntactic zone. (7) Encyclopedic zone - interpretation of the word in its cultural context. Lexical material is conveniently stored on cards arranged in alphabetical o r d e r . This arrangement makes the needed words e a s i e r to find on the one hand and on the other it insures that the dictionary can be constantly expanded without disrupting its internal order. In view of the frequent need to classify the lexical fund by variable f e a t u r e s it is good to use cards with r i d e r s f o r holding the lexical information: r i d e r s (colored flags inserted in special slots) allow the investigator to frequently change the system of features and f r e e him f r o m the need of making cuts in the cards. 3 . 4 . Paradigms P a r a d i g m s a r e an important means of organizing data if we a r e working with a target language which has word inflexional categories. Paradigms s e r v e a s an empirical basis for morphological generalizations. In collecting paradigms one should scrupulously observe the following rules: (1) the paradigms should be complete; (2) the paradigms should contain m a t e r i a l illustrating different types of word inflexion within the given category (that is, all types of declension should be considered f o r nouns and all types of conjugation f o r verbs so that the entire morphology can be extracted f r o m them). This means that even though the p a r a digms a r e being collected in the fairly early stages of the work (that) they a s s u m e their final form only a f t e r final decisions have been made about the system of word inflexion. The collection of paradigms may be a very time consuming task f o r investigator and informant alike in languages which have highly developed

67 word inflexion. Therefore much time and attention must be directed upon this problem from the very beginning. NOTES 1* A.Einstein, "Motiv des Forschens" in Zu Max Plancks 60 Geburtstag (Ansprachen in der Deutschen Physiealischen Gesellsehaft, Karlsruhe, 1918, 2 9 - 3 2 ) ; A. Einstein, "On the Method of Theoretical Physics", Philosophy of Science, v. I, 1934, 162-169. See also C. P e i r c e ; s guessing instinct in examining the nature of knowledge in: N. Chomsky, Language and Mind, 1972, 90-91. 2* "C. M. Voegelin and F. M. Robinett, "Obtaining a Linguistic Sample", IJAL, v. 20, N i l , 1954, 89-100. 3* Bogoraz used a similar phrase book in his field work. See I. S. Vdovin's foreword to V. G. Bogoraz, Materialy po jaz.yku aziatskix ¿skimosov [ Materials on the Language of Asiatic Eskimos ], L . , 1949, 22. 4 * " E . Nida, "Selective Listening", Language Learning, 4, 1952-53, 92-101. 5* M. Swadesh, "Language Universals and Research Efficiency in Descriptive Linguistics", The Canadian Journal of Linguistics, v. 10: 2 , 3 , 1964,^149. 6* L. J a . Sternberg, Materialy po izuceniju giljackogo jazyka i fol'klora [ Materials on the Study of Gilyak Language and Folklore ], Saint-Petersburg, 1908, IX. 7 * "This was also pointed out by one of the leading representatives of the Russian field linguistic tradition - Baudouin-de-Courtenay, who was engaged in many dialectological field investigations: " F o r science a small collection of true and undoubtable facts and the conclusions based on them a r e more desirable and more useful than an enormous pile of hurriedly noted phenomena and the suppositions, often of a highly problematic nature, based on them" (I. A. Bodu6n de Kurten6 [ J . A. Baudouin de Courtenay ] , Opyt fonetiki rez' janskix govorov [ An Experimental Phonetics of the Rez'ja Subdialects], Warsaw-Saint-Petersburg, 1875). " 8* This is just what Bogoraz has in mind when describing the method he used in studying Chukchi: he worked on and off for a year and a half "until forms of unfamiliar origin came to be encountered relatively r a r e ly" (V. G. Bogoraz, Materialy po izuceniju cukotskogo jazyka i fol'klora [ Materials on the Study of the Chukchi Language and Folklore ] , Trudy Jakutskoj ekspedicii, snarjaZennoj na sredstva I. M. Sibirjakova [ Works of the Yakut Expedition, Financed by I. M. Sibirjakov ], Section HI, v. XI, part HI, Saint-Petersburg, 1900, p. II). 9* " P i s ' m o prof. A. I. Belica k I. A. BoduSnu de KurtenS o sobiranii dialektologiceskogo materiala" [A Letter from Prof. A. I. Belie to J . A. Baudouin de Courtenay on the Collection of Dialectological Material ], Izvestija otdelenija russkogo jazyka i slovesnosti imperatorskoj Akademii Nauk [News of the Department of Russian Language and Philology of the Imperial Academy of Sciences] , v . X V m , b o o k l , 1913, 238. 10* Baudouin de Courtenay insisted on the necessity of a repeated check with the informant of the material after its theoretical study: cf. S. K. PoZarickaja, "Metodika dialektologiCeskix issledovanij I.A.BoduSna de

68

KurtenS" [The Methodology of J . A. Baudouin de Courtenay's Dialectological Investigations ], Materialy i issledovanija po russkoj dialektologii [Materials and Investigations in Russian Dialectology], New Series, v.ni, M., 1962, 167. 11* See Uhlenbeck's remark: " . . . linguist should converse with the informant without using an intermediary language" ( E. M. Uhlenbeck, "The Study of the So-Called Exotic Languages and General Linguistics", Lingua, 9, 1960, 443). 12* See: M. Mead, "Native Languages as Field-Work Tools", American Anthropologist, 41, 1939, 189-205. 13* The example was provided by Ju. D. Apresjan. 14* See: A. P. Dul'zon, "Dialekty tatarskix aborijjenov Tomi" [The Dialects of the Tatar Aborigenes of the Tom' ], UcSnye zapiski Tomskogo pedinstituta [Scientific Notes of the Tomsk Pedagogical Institute ], XV, 1956, 320. 15* B. Bloch, "Interviewing for the Linguistic Atlas", American Speech, 10. 1935, 7. 16* See: H. Hoijer, "Anthropological Linguistics", Trends in European and American Linguistics 1930-1960, Antwerp/Utrecht, 1961, 110-112. 17* See, for example, the following articles: R. H. Wax, "Reciprocity in Field Work"; St. A. Richardson, "A Framework for Reporting Field Relations Experiences"; K. W. Black, "The Well Informed Informant"; A. Vidich and J . Barsman, "The Validity of Field Data"; K. H.Wolff, "The Collection and Organization of Field Materials: A Research Report"; H. S. Becker and B. Geer, "Participant Observation: The Analysis of Qualitative Field Data" - in the book: R. N. Adams and J . J . Preis (eds.), Human Organization Research: Field Relations and Techniques, in, The Dorsey Press, Homwood, 1960; see also: J . A. Barnes, "Some Ethical Problems in Modern Fieldwork", British Journal of Sociology, 14, 1963, 118-134; W. J . Hanna, "Image-Making in Field Research: Some Technical and Ethical Problems Arising from Research in Tropical Africa", American Behavioral Scientist, 8 (5), 1965, 15-20. 18* See the notes on differences in forms proposed by different informants in: E. A. Krejnovic, Glagol ketskogo jazyka [ The Ket Verb], L . , 196«, 144, 149, 168, 169, 190, 208, and elsewhere. 19* A. S. Hayes, "Field Procedures While Working with Dieguefio", IJAL, v. 20, 1954, 187. 20* For example, spatial meanings. See: J . Yegerlehner, "A Note on Eliciting Techniques", IJAL, v. 21, 1955, 286-88. 21* Ju. V. Vannikov, A. N. Scukin, Kartinnyj slovar' russkogo jazyka [A Dictionary of Russian in Pictures ], M., 1969, can serve as an aid for such a method (with the exception of certain of the situations of city life and so on unfamiliar to the country dweller). 22* J.Henry, "A Method for Learning to Talk Primitive Languages", American Anthropologist, 42, 1940, 637. 23* This was Bogoraz's basic method: "I took a dictation of fragments of stories from Ajnavanvat and other Chukchas and then made an approximate translation, subjecting each word to a detailed grammatical analysis, clarifying for myself the correlations between words and thereby making my translation more precise" (V. G. Bogoraz, Materialy po izuCeniju cukotskogo jazyka [Materials on the Study of Chukchi ] (see note 96, p. 11).

69

24* See: S. S. Vysotskij, "Metody izucenija dialektov" [Methods of Dialect Study ], Doklady i soobscenija IJa AN SSSR [ Reports and Communications of the Institutute of Linguistics, Soviet Academy of Sciences ], 6, 1954, 137. 25* L. Bloomfield, Outline Guide for the Practical Study of Foreign Languages, Baltimore, Md., 1942, 2. 26* See this kind of note in: A. E. Kibrik, S. V. Kodzasov, I. P . Olovjannikova, Fragmenti grammatiki xinalugskogo jazyka [ Fragments of a G r a m m a r of Xinalug] , Izd-vo MGU, M . , 1972. 27* Ch. Hockett, "Translation via Immediate Constituents", IJAL, v. 20, 1954, 313. 28* C. E. Voegelin, "Multiple Stage Translation", IJAL, v. 20, 1954, 271-80; C. E. Voegelin, "A Modern Method f o r Field Work Treatment of Previously Collected Texts", Journal of American Folklore, 67, 1954, 15-20. The application of Voegelin's method is described in: J . Yegerlehner, "The F i r s t Five Minutes of Shawnee Laws in Multiple Stage Translation", IJAL, v. 20, N 4, 1954, 281-294; M. Davis, "Translation from F L Cuicateco to TL English", IJAL, v. 20, 1954 , 302 -312. 29* See: F. Lounsbuiy, Field Methods . . . , 410. 30* See, f o r example, the attempt to c r e a t e a program of investigation to elicit the words belonging to the semantic group "firewood": D. G. Metzer, G. E. Williams, "Some Procedures and Results in the Study of Native Categories: Tzeltal " F i r e Wood" ", American Anthropologist, v. 6, N 2, p. 1, 1966, 389-407. 31* Ju. D. Apresjan, "Ob Sksperimental'nom tolkovom slovare russkogo jazyka" [ An Experimental Defining Dictionary of Russian] , VJa, 5, 1968, 34^19; Ju. D. Apresjan, "O jazyke dlja opisanija znaCenij slov" [ A Language f o r the Description of Word Meanings ], Izvestija AN SSSR. Serija literatury i jazyka [ News of the Soviet Academy of Sciences. Series on Language and Literature] , v.XXVIII, issue 5, 1969, 415-428; Ju. D. Apresjan, "Tolkovanie leksiceskix znacenij kak problema t e o r e t i ceskoj semantiki" [ The Interpretation of Lexical Meanings a s a Problem of Theoretical Semantics ], Izvestija AN SSSR, Serija literatury i jazyka, v. XXVIII, issue 1, 1969, 11-23; J u . D . Apresjan, I.A.Mel'Cuk, A. K. Zolkovskij, "Semantics and Lexicography: Towards a New Type of Unilingual Dictionary" in F. Kiefer (ed.), Studies in Syntax and Semantics, Dordrecht, The Netherlands, 1969, 1-33; see also the attempt to i n t e r pret the verb to be and several others with a comparison of the interpretations in different languages: F. H. Bendix, "Componential Analysis of General Vocabulary: The Semantic Structure of a Set of Verbs in English, Hindi, and Japanese", IJAL, v. 32, publication 41, 1966. 32* H. P. Phillips, "Problems of Translation and Meaning in Field Work" in R. N. Adams and J . J . P r e i s ( e d s . ) , Human Organization Research . . . , 90-307. 33* E. A. Nida, "Analysis of Meaning and Dictionary Making", IJAL, v. 24, 1958, 282. 34* A. K. Zolkovskij, I. A. Mel'cuk, "O semanticeskom sinteze" [On Semantic Synthesis], Problemy kibernetiki [ Problems of Cybernetics ], issue 19, M., 1967, 137-238 (in French: A. K. Zholkovskij et I.A.Mel' chuk, "Sur la Syntese Semantique", Revue Internationale du Traitement Automatique du Langage, N. 2, 1970); Ju. D. Apresjan, A. K. Zolkovskij,

70 I. A. Mel'cuk, "O sisteme semanticeskogo sinteza. in. Obrazcy slovarnyx statej" [On the System of Semantic Synthesis. HI. Samples of Dictionary .Articles] , NauCno-texniCeskaja informacija [ Scientific-Technical Information] v No. 11, 1968. 35* A. K. Zolkovskij, I. A. Mel'cuk, "K postroeniju dejstvujuscej modeli jazyka "smysl tekst"J' [Towards a Functioning "Meaning Text" Model of Language ], Masinnyj perevod i prikladnaja lingvistika [ Machine Translation and Applied Linguistics ], issue 11, M., 1969, 32 (English translation appeared in Linguistics, N. 57, 1970, 44). 36* On the "thesaurus" type dictionary see: M. V. Arapov, "Nekotorye principy postroenija slovarja "tezaurus" " [Some Principles f o r Constructing a Thesaurus Type Dictionary ], Naucno-texniceskaja informacija ^Scientific-Technical Information], 4, 1964, 40-46; A.I. Mixajlov, A.I. CBrnyj, R. S. Giljarevskij, Osnovy informatiki [ The Bases of Information Theory] , M . , 1968, 413-32; V. V. Morkovkin, Ideografi&eskie slovari [ Ideographical Dictionaries ], M., 1970. The best known thesaurus is Roget's: P. M. Roget, Thesaurus of English Words and P h r a s e s Classified and Arranged so a s to Facilitate Expression of Ideas and to Assist in Literary Composition, L . , 1852. 37* The comparative dictionary ( f o u r Eskimo dialects) in G. A. Menovscikov, Jaz.yk sirenskix éskimosov [ The Language of the Sireniki Eskimos ], M. - L . , 1964, 204-216, may s e r v e a s an example f o r the thematic ordering of a dictionary f o r an insufficiently described language; the thematic organization of the lexicon is presented in: A. L. Grjunberg, Jaz.yk severoazerbajdzanskix tatov [ The Language of the Northern A z e r baijan T a t a r s ] , L . , 1963, 114-122. 38* See, f o r example, the r e v e r s e dictionary of Russian: H. H. Bielfeldt, Rückläufiges Wörterbuch der Russischen Sprache der Gegenwart, Berlin, 1958.

CONCLUSION

Thus we have briefly examined the basic questions encountered by the field investigator: the place of field linguistics in linguistics as a whole, information and materials one should have in hand before beginning the investigation of an unfamiliar language, how the process of field investigation itself should be organized. These questions were dealt with in varying degrees of completeness and detail. It is clear that there r e mains much to be done in understanding and developing the technics of field work. For fairness' sake, because it is a fact too often neglected, we would like to remind the reader that the singular attraction of field work consists in the unique union of general theoretical investigation with the scrupulous but poetic observation of a real functioning living human language and all its human problems; for it is people who occupy the most important place in one's understanding of language. B. Bloch put it well: "Field work is perhaps one of the most exacting forms of research but it certainly is the most exciting and the most varied. The dialect hunter must face bad roads, worse lodging, growling dogs, suspicious farmers, and hour upon hour of routine as he reels off his questions for the ninetieth time by the light of a kerosene lamp. But he learns to know the living speech of the people as no one can who has only read of it in books. He leads a peasant, wandering life on the road, and finds a new world of friends among the kindly men and women he meets in his travels. " 1 *

1*

B. Bloch, "Inverviewing for the Linguistic Atlas", American Speech,

10, 1935, 9.

Appendix 1 A UNIVERSAL PHONETIC TRANSCRIPTION BASED ON LATIN (VARIANT OF THE IPA) [See: Chapter 2, 2 . 2 . ]

74

I

«-v •o N M

Ö

xo

N

Labialized

&

Ü

0

a


-

55

)o

w

8

-O 03 & >o ÜH — ta cd

m

Zone

«

*

Mid-High

ai ni

m

ci 1A — '

S i?

X

Plosives

IQ tì O o

bo

Noisy

§

Ü io a*

¿1 c, t/>c, dz > z, d3>j, q/Oq. kx> k, cf. the phonological translitteration of Archi. 4* x and y - signs, denoting sounds from corresponding places of articulation and, as a rule, the same means of articulation. Concerning halfcircles see Note 3.

76

Sample Transcriptions - Archi (Daghestan Group of the Caucasian Family) Phonetic transcription [ recorded correlative and prosodic features: aspiration, glottalization, labialization, lateralization, affrication, phaiyngealization, duration - strength, stress, and secondary stress ] : 1.

os ébdal ebdat'u xfbuù k^ululò II

2.

do'z éptamxur k'an do'zamu bòli II

3. 4.

zon wirx mus xóuta bòli II w u'qjc'a'l os mat/'élji wfrx^ mus II

5.

iqxit'atu ba|ór ^ydsli // jàmumarak* 1 k^óùfuli joù lo II

6.

dagàwur óuli jémimai hlnumatij hànu in3ft ^ytajou sónatij fat ópkxmus

7.

jou lo t'al óuli jàmu bajórmu wfrx

^

— *

—

—

—

— —

—

mus os Mqac'^mS/i II

jat h bòli jltsj' sob dàpxut'otl bagkówu k h em bòli à'pqx'a't'ou — * h li ^ x àloù buk ém bòli ... w Phonological transcription [the same features are recorded except for aspiration ( =non-duration) and secondary s t r e s s ] : 8.

1. os ébdili ébdit'u Xibuwu kulti lo i 2. doPz ébtimxur k'an doPzummu bòli 3. zon wirx^mus xówti bòli 4. uqlali os mac'éjsi wirx„mus 5. iq it'vtu bosór éwdili jàmummerak kówsuli jow lo 6. dagàwur ówli jémimmaj hànnumitis hànnu in2it éwtisaw sónnitis sat óbkmus 7. jow lo t'al òwli jàmu bosórmu wfrx^mus os biq^masi 8. jat bòli jàtis sob dàbxut'aw bankówu k o en bòli à bqlut'aw x 6 àllowu bukén bòli . . . Translation: 1. Once upon a time lived three orphans. 2. When they had grown up the eldest said: 3. "I'm going to work". 4. He went to a certain place to work. 5. ( There) was a cruel man, whom he ran into. 6. They agreed that the first person to get fed up with the other would have a muscle ( along his backbone) cut out.

77 7. This man sent that fellow to work in a certain place. 8. He said: "Eat the contents of this j a r without taking the lid off, and eat this bread without breaking it" . . . Xinalug (Daghestan Group of the Caucasian Family) Phonological Transcription [recorded correlative and prosodic features: aspiration, glottalization, palatalization, duration-strength, stress, secondary stress, affrication] : 1. s, fr, yr sa k

i ajae

2. d,ae k|*i vtik'ur, i b, i3ib, iq'i 3. hat k h p h / i g, ae./ijae. tpiibmKi 4. k^i IStq'ujae ink, imzttr Ji'k'ui 5. hina d,32. vaz ttivujaeinklaak ui 6. k^ijaen, fqublax 1 , p h j k u j a e . vtik'ui ^

L

1.

7. hie. j « n , Iqublax 1, Sekjp, ijkund, a?q'4lii hin, f lukfi t » m , fs kujlm.ae 8. hae. k, i g, a £ r , & g , ¿en,

vtfk'ura hin, f p', i t, a i m , fs

4nt«r vaxkua 9. h&na d, ae h, in, f k! i r qflvi 10. hlna d,'at hin, f Jax ¿lturq'uj ae. hin, t itj ir It, i murdlrdai rm.at 11. tena. d,ae h x lluki t J ' i t W 1 * h. r m .ha . h 12. h^na d,2R hae lvika k san qormti|k, it arm,aft Translation: 1. There was a fat ram in the herd. 2. They wanted to slaughter that ram. 3. They put the ram on a horse and brought him in. 4. They threw the ram on the ground and tied his legs. 5. Then they took of knife and sharpened it. 6. They turned the ram's head to the south and slaughtered him. 7. If you don't turn the ram's head to the south, the meat won't be "clean". 8. You need to cut the ram on a slope so that all the blood will run out. 9. Having slaughtered the ram, they skinned it. 10. Then, having removed the ram's belly, they throw away all the filth the entrails. 11. from Then they cut the meat in pieces. 12. Then they roast the meat well.

The Alutor Dialect of Koryak ( Palaeo - Asiatic Family) Phonetic transcription [recorded features: duration, accent] : 1.

wutin iti wutiji ?iju

2.

g5mnin turn upturn masinata talàlqivi qaj^swane

3.

wajàlyirjki yàntamijivlin wina

4.

gè tag ¿agviilin talà?ugak wesita

/

5.

ktawt jjalyàlalqivlin mfkanak àno ? Snu tanyàrnikgatkanin

6.

tjaninfalu ?ayi yerjàìin

7.

wajfnvan yalà?ulin namérjqm ¡prnik

8.

òro ^arjujavlin to yamsàturalqivlin

9.

ySrnik àmanrja Jamsàturalqivlin

10. jsmjfSmyatkana to jagàkjita yapkàvlin m f l ^ a r a tanmarjki . . . Translation: 1.

That was last year.

2.

My friend drove through the mountains in a car.

3.

During a snow storm he lost the road.

4.

Then he began to search on foot.

5.

Then suddenly he began to notice someone pursuing him.

6.

There was a heavy fog that day.

7.

Just the same he saw a large animal.

8.

Finally he weakened and began to rest.

9. The animal also began to r e s t . 10. Afraid and because of the fog he was not able to kill ( it) with weapon . . . Fula ( West-Atlantic Group of the Niger-Congo Family) Phonological Transcription [recorded features: implosiveness, duration, nasalization]: 1. hid"on adi wocfema gelai e s a r i eder tali b"aleb"e b"en h a r a j folocTirde telema ka j-"ojre 2. jiade go? gelai e s a r i halduno sud"ocfirgol 3. s a r i r e 3én huci fi jo gelai f a i heb"u no sud"oro 4. gelai gal dari mij-iti ha 5. gal asi gajkun ka hakucté j-ule beru b " adi magai 6. gal iri hore magai ka cfer gajkun aciti cape d"en ka kene

79 7. hid"on adi wodema cape gelal gal no nadi e lekoj jorkoj eder den fitare 8. sarire Elen jahi ari jurni j-"ewi d"ab"itiri no d"ab"itirte non fow 9. de d"ab"iti ha de ronki ka gelal gal sud"ino d"on . . . Translation: 1. As is known, in African tales the partridge and the hare often compete in a battle of wits. 2. One time the partridge and the hare agreed to hide from one another. 3. The hare turned around so that the partridge could hide. 4. The partridge stood for a long time, thinking things over. 5. -6. Then she dug a hole between some stumps the size of her body and stuck her head in, leaving her legs on the outside. 7. Everyone knows that a partridge's legs look like dry branches in the woods. 8. The hare walked and walked, looked and looked, searched everywhere. 9. But no matter how hard he looked - he couldn't find the partridge . . .

Appendix 2 A SAMPLE OF QUESTIONNAIRES IN THE FORM O F A CALCULUS (SPATIAL GRAMMATICAL MEANINGS) [ See: C h a p t e r 2

3.2.]

83

86

Examples of Interpretations of Orienting Meanings 1.1.1. 0.1.1. 1.1.1.1.1.1. 1.1.1.2.1.1. 1.1.1. 0. 2.1. 1 . 1 . 1 . 1 . 2.1. 1.1.1. 2. 2.1. 1.1.1. 0. 2.2.

1. 0 . 1 . 1 . 1 . 1 . 1. 0 . 1 . 2 . 1 . 1 . 1. 0.1. 0.2.1.

beneath Or, touching Or in the center beneath Or, touching Or at the lower edge of Or, touching Or beneath Or, not touching Or in the center beneath Or, not touching Or at the lower edge of Or, not touching Or beneath Or, not touching Or, and located in some spatial relationship to AdOr in the center beneath Or, not touching Or, and located in some spatial relationship with AdOr at the lower edge of Or, not touching Or, and located in some spatial relationship with AdOr on Or in the center on Or at the edge on Or above Or in the center above Or at the edge above Or above Or, and located in some spatial relationship with AdOr in the center above Or, and located in some spatial relationship with AdOr at the edge above Or, and located in some spatial relationship with AdOr in the center on the vertical surface of Or at the edge of Or on the vertical surface of Or around the vertical surface of Or

1. 0.2. 0.1.1.

on the inclined surface of Or

1. 0.2. 0. 2.1. 1.2.1.1.1.1.1. 1.2.1.1.1.2.1. 1.2.1.1. 0.1.1.1. 1. 2.1.1. 0.1. 2.1. 1.2.1.1. 0. 0.1.1. 1. 2.1.1. 0. 0.2.1. 1. 2.1.1. 0.2.1.1. 1.2.1.1. 0.2.2.1. 1.2.1.1.2.1.1. 1.2.1.1.2.2.1. 1. 2 . 1 . 2 . 1 . 1 . 1 . 1.2.1.2.1.2.1.

around the inclined surface of Or immediately in front of Or near the front of Or immediately to the right of Or to the right of Or immediately to the side of Or to the side of Or immediately to the left of Or to the left of Or immediately behind Or behind Or immediately in front (relative to the speaker) of Or near the front (relative to the speaker) of Or

1.2.1.1.1.1.2.

immediately in front of Or, and located in some spatial relation to AdOr

1. 2.1. 2.1.1. 2.

immediately in front (relative to the speaker) of Or and located in some spatial relationship to AdOr

1 . 1 . 1 . 1 . 2.2. 1.1.1. 2. 2. 2. 1.1. 2. 0.1.1. 1.1.2.1.1.1. 1.1.2. 2.1.1. 1.1.2.0.2.1. 1.1.2.1.2.1. 1.1.2. 2. 2.1. 1.1.2. 0. 2.2. 1 . 1 . 2 . 1 . 2.2. 1 . 1 . 2 . 2 . 2.2.

87

1. 2. 0. 2.1. 1.2. 2 . 1 . 1 . 1. 2. 2. 2. 0.1. 1.2. 2 . 2 . 1 . 1 . 1.2.2.2.2.1. 1 . 2 . 2 . 2 . 0.2.

encircling Or immediately near Or outside Or near, alongside of Or far from Or outside Or, and located in some spatial relationship with AdOr

0.1. 2.1.1.1. 2.1.2.1. 2.2.1.1.1.1.

2 . 2 . 1 . 1 . 2.2. 2.2.1. 2.1.1.

in, on, at Or ( indefinite spatial relationship to Or) inside Or in solid space ( e. g. in flour, in sand, etc.) inside Or in empty space ( e. g. in a box, bottle, etc.) between Or, touching some parts of Or (e. g. immediately between the trees) between Or and AdOr, touching them between Or, not touching any parts of Or ( e. g. between the trees) between Or and AdOr, not touching them between Or and the speaker, touching them

2.2.1. 2.1.2.

between Or, the speaker, and AdOr, touching them

2. 2 . 1 . 1 . 1 . 2. 2. 2 . 1 . 1 . 2.1.

2.2.2.1.1. 2 . 2 . 2 . 2.1.

in the midst of Or, mixing up with it ( ink in water) in the midst of Or, not mixing up with it ( a coin in the midst of rubbish) Examples of Interpretations of Motive Meanings 1.1.1.1. to move right up to contact with Sr 1.1. 0.1. to move to Sr 1.1.2.1. to move to Sr and across Sr (touching) 1. 0 . 1 . 1 . to move up to Sr 1. 0. 0.1. to move towards Sr 1. 0 . 2 . 1 . to move to Sr and farther (touching not mandatory) 1.2.1.1. to move toward some neighbourhood of Sr ( not touching it, and going no farther) 1.2. 0.1. to move towards Sr ( without contact) 1. 2. 2.1. to move towards Sr and past Sr ( without touching Sr) 0.1.1.1. to move along Sr, without leaving Sr 0.1. 0.1. to move along Sr; to be in contact with Sr 0.1.2.1. to move along Sr, entering or coming out from Sr 0.0.1.1. to move at an indefinite closeness from Sr, not changing this state 0. 0. 0.1. to move at an indefinite closeness from Sr; being located at an indefinite closeness from Sr 0. 0 . 2 . 1 . to move at an indefinite closeness from Sr entering or coming out from Sr 0.2.1.1. to move alongside Sr, not changing this state 0.2. 0.1. to move along side Sr; to be located alongside Sr 0.2.2.1. to move from a certain moment or until a certain moment alongside Sr 2.1.1.1. to move directly from Sr, which is the starting point of the motion

88

2.1. 0.1 2.1. 2 . 1 2.0. 1.1 2.0. 0.1 2.0. 2 . 1 2.2. 1.1 2.2. 0.1 2.2. 2.1 1.1. 1.2 1. 1. 1.1. 1.2 1. 0. 1.1. 1.2 1. 2. 1.1. 1.2 2. 1. 1.1. 1.2 2. 2.

to move directly from Sr to move through Sr and from Sr (touching) to move from an indefinite neighbourhood of Sr, which is the starting point of the movement to move from Sr to move through òr past Sr and from Sr to move from some neighbourhood of Sr, which is the starting point of the movement to move from some neighbourhood of Sr to pass by Sr (not touching Sr) and f r o m Sr to move towards the speaker right up to contact with Sr to move parallel to the speaker right up to contact with Sr to move from the direction of the speaker right up to contact with Sr to move downwards right up to contact with Sr to move up right up to contact with Sr

Appendix 3 A SAMPLE OF QUESTIONNAIRES IN THE FORM OF A SET OF DIAGNOSTIC UTTERANCES (NON-SPATIAL MEANING IN A NOUN PHRASE) [See: Chapter 2, 3. 3. ]

90

Meaning Connotation

Basic

1. Object

Examples

1.

washing linen to catch fish to hear a song to listen to father to lead the people to engage in sports to believe in the truth to begin work to rely on a friend to bring somebody to trial to talk about war to spy on the enemy to be afraid for one's daughter to laugh about the enemy power over one's son that happened to me a story about old times

2. Nuance of Causality

to receive for work to love for one1 s wit to surprise with one's wit

3. Nuance of Goal

to desire fame to go for water to learn to write

4. Nuance of Intention

feed for livestock

5. Nuance of Directedness

to write a letter to a friend to promise father to come to threaten an enemy with a weapon a letter to one's brother to take by the hand

6. Nuance of distancing

to fear water to be a stranger to envy to rid oneself of disease to come to one's senses from fear

7. Source, not a Person

to find cut from stories

8. Source, Person

to find out from father to take an example from one's brother

91

Meaning Basic

Connotation 9. Nuance of Causation

Examples to sew a d r e s s to turn the desert into a garden to break into pieces

10. Object is Subjected to Action in P a r t s

to drink some water

11. Nuance ( P a r t i a l ) of the Transformation of the Object

to grind grain to make out of wood to paint the roof

12. Object - Part of the to wave one's hands Subject of the Action

1. Object

13. Nuance of Instrumen- to reward with an order tality to cover one's face with a collar to play the guitar to speak one's native language 14. Nuance in Instrumen- to fill with water tality and Spatiality 15. Nuance of Spatialily

to to to to to

run a c r o s s the road lift a stone kiss on the cheek wrap up in a scarf beat on the shore

16. Nuance of Co-occur- to meet with the enemy rence 17. Nuance of C o r r e spondence

to be guilty before one's father

18. Nuance of Limitation full of t e a r s faithfulness to a husband resignation to fate danger f o r a child harmful to health to reflect on one's mood 19. Choosing

the best of all

20. Dividing

to give everyone an apple

92

Meaning Connotation

Basic 1. Object

2. Attributive

Examples

21. The Object Refers to give a lot of toys to a Significant Number of the Possible Set of Predicates 22. Predicative

to consider him a fool to work as a teacher

1.

t e a r s of joy striped t r o u s e r s a f u r coat

2. Nuance of Cause

smoke from the bonfire foot prints

3. Nuance of Cause and Limitation

o r the smell of smoke

4. Nuance of Result

the grief of parting

5. Intention

tea cakes ware-house

6. Intention and Number

a fifty gallon drum

7. Nuance of Number

a t h r e e story house

8. Source

lamp light honey cakes

9. Origin, Subject-Thing wine bottle 10. Origin, Subject-Person

peasant f r o m the country

11. Choice + Source

a page from the book

12. Choice

the best of the athletes

13. Element of a Set

one of his friends

14. Aggregate

peasant detachment

15. Contents

j a r of water

16. P a r t and Whole

roof of the house chicken's head handle of the suitcase

93

Meaning Connotation

Basic

Examples horse's tail girl with pig tails

2. Attributive

17. Feature and Carrier warmth of the body 18. Nuance of Limitation black with a streak of gray

3. Restrictive

4. Belonging

19. Nuance of Spatiality

a noise throughout the entire forest

1.

weak in spirit short for hands laugh until tears to be clever at music relative on the husband's side good at heart a wizard of invention the man behind the wheel business success wide shouldered dangerous in a fight frightful when angered

2. Nuance of Cause

happy with victory

3. Nuance of Goal

convenient for walking

4. Intention

medicine for pain medicine against pain suitable for work

5. Qualitative

swarthy to the point of being black

6. Quantitative

expenses up to one hundred dollars

7. Temporal

to work till morning

8. With a Mandatory Attribute

to shout with a loud voice

9. Correlative

friendly with people

1.

father's book I have friends to leave f o r the buyer

94

Meaning Basic

4. Belonging

Examples

Connotation 2. Belongs to an Author

f a t h e r ' s words

3. Indirect Belonging

b r o t h e r ' s aunt

4. Member and Collective commander of a detachment 1. 5. Quantitative

6. Instrumental 1.

7. Comparative

8. Adverbial

glass of milk to weigh a ton h e ' s over forty to weigh about a kilogram a herd of horses to chop with an axe to shoot with a rifle

1.

to fly like a bird to walk in a c a t ' s fashion divorce Italian style doughlike

1.

to to to to to

2. Nuance of Co-occurrence

to read with glasses to listen with a smile

3. Correspondence

to act according to the law

4. Nuance of opposition

to do against one's will to do in spite of everyone

5. Replacement

to work instead of one's brother

6. Intermediary

to t r a n s m i t through one's sister

speak in a whisper work on an empty stomach live n e a r the estate go under guard stand guard

to shorten by an hour 7. Nuance of Quantity 8. Nuance of Temporality to sleep to music

Meaning Basic

9. Comitative

Connotation

Examples

1. Object-Person

to go without one's brother to live with mother brother and sister a q u a r r e l among f r i e n d s

2. Object-Thing

to go with a rifle to go with a report

3. Property

to speak with e m b a r r a s ment

4. Opposition

one against ten

5. Nuance of Spatiality

to quarrel with people to become e m b a r r a s e d in front of company to say out of jealousy to return because of the rain to cry from happiness a c r y of pain to turn gray with grief to revenge an injury wet with t e a r s

10. Causal

11. Aim

95

1.

to to to to to to

2. Intention

to give something to r e member

work f o r money p r e p a r e for sale say in defense fight f o r freedom go hunting go b e r r y picking

a d r e s s length to p r e p a r e a field f o r cabbage to save up for a rainy day 1. Simultaneity 12. Temporal

to to to to to in

come in the evening break up at midnight set for the morning say at dinner leave at dawn childhood

96

Meaning Basic

Connotation

Examples to get up at daybreak

2. Preceeding

to leave before sundown to return by dinnertime just before evening the calm before the storm

3. Following

to get m a r r i e d in a month's time afternoon

4. Period

to leave as a child to work from Wednesday to . . . to do in a week to learn a f t e r a month the sum f o r the y e a r to work up to and through today between work and leisure

12. Temporal

5. Nuance of Frequenta- to visit at night tiveness to work nights to go f o r a walk every other day 6. Nuance of Quantity

to read for an hour

7. Figurative

to utter a s c r e a m at those words in the moonlight

1. Figurative

to get out of trouble to return f r o m work the c a r e s lay on the mother

2. Plurality of Objects

to go from house to house

1.

the son sleeps the mother's reproach I catch

13. Spatial

14. Subject (agent)

15. Affective, dative

I hear I see I like it to give to one's mother to be rude to one's father something one's brother needs

Appendix 4 AN APPROXIMATE L E T OF SENTENCES FOR THE FIRST ENCOUNT E R WITH THE LEXICON AND GRAMMAR O F THE TARGET LANGUAGE [ See: Chapter 2, 5. ]

99

Diagnostic Sentences

New Words

1. 2. 3.

mother, to bring, water what father, to take, bread, table, black, white

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

Mother brought some water What did mother bring? Father took bread (black, white) from the table. What did your father take? Where did your father take bread from? The wife gave the husband milk Who gave the milk? The girl came home Where did the girl come? The man left the house (in the evening, in the afternoon, in the morning) Where did the man leave from? You (they) work in a field Where do you work? I wash my hands with water (in the morning) The woman is cooking soup (in a kettle) The woman is frying meat (on a fire) The woman eats soup (meat) The woman drinks water (milk) The woman feeds the child with soup (meat) The woman gives the child water (milk) to drink He talks (a lot) about the child Mother loves the child He loves to be silent, he doesn't talk much Why doesn't he talk much? I see the sun (the countryside, a mountain, a river) I look at the sky and see the sun I look at the mountain but see nothing there I listen to my friend's stoiy I listen to their conversation, but don't hear anything I hear the sound of the water The moon (sun) is in the sky I am writing (reading) a book (a letter) The teacher teaches the children to write My father has pretty clothing

where from wife, to give, husband, milk who girl, to come, home where man, house, to leave, in the evening, in the afternoon, in the morning you, they, to work, field where I, to wash, hand woman, to cook, soup, kettle to to to to

fry, meat, fire eat drink feed, child

to give to drink he, to talk, a lot to love to much, to be silent why to see, sun, countryside, mountain, river to look, to see, sky nothing, there but, negation to listen, story, friend to hear, conversation sound moon, to be book, letter, to write, to read teacher, to teach, children to have, my, pretty, men's clothing

100 35. My mother has pretty clothing 36. Clothing ( men's, women's) is hanging on the wall 37. At night the boy lies in his bed 38. I am sitting on a chair (at the window 39. I opened the window (door) 40. I am standing at the window 41. I am sitting under a tree 42. The cow eats grass 43. I am pasturing sheep on the bank of the river 44. Today (yesterday) it is (it was) snowing (raining) 45. When did it snow? 46. Did it snow yesterday?-Yes (no) 47. Is your friend home (or not) ? Yes (no) 48. That old man (old woman, person) died yesterday 49. He (very much) wants to eat 50. He's still young (not old): he should work 51. He lives closeby: he can come here 52. 53. 54. 55. 56. 57. 58. 59.

He rarely comes to (see) me He asked me for some money He tells tales well The grass is green This tree is tall It's a nice day today This man is his brother That's the man whom I saw yesterday 60. Close the door 61. It is a house (man, woman)

women's clothing to hang, wall boy, to lie, bed, at night to sit, chair, window to open, door to stand tree cow, grass, to eat (about animals) to pasture, sheep, bank today, yesterday, to snow, to rain when yes, no your, friend that, old man, old woman, person, to die to want, very much should, to work, still, young, old to be able to, to come, here, to live, closeby rarely to ask for, money well, to tell, tale green this, tall day, nice his, brother whom to close it

Appendix 5 MINIMUM DICTIONARY (AN APPROXIMATE LIST O F WORDS (SENSES) FOR THE COLLECTION OF VOCABULARY) [See: Chapter 2, 5 . ]

103 Explanation of the Minimum Dictionary 1. The words of the minimum dictionary are classified in four basic classes: objects, features, situations (states or actions ± the objects and features related to them) and replacives (of objects, features, and situations). Within each of these classes a more detailed thematic classification is drawn. The order of these thematic groups is as-follows: 1. Objects: 1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 1.8. 1.9. 1.10. 1.11. 1.12. 1.13. 1.14.

Names of persons by age and sex; Kinship; Parts of the human body; Domesticated animals; The animal kingdom (animals, birds, fish, insects); The vegetable kingdom; Heavenly bodies; Inanimate nature; Housing; Everyday household objects; Clothing, footwear, decoration; Food; Geometric concepts, measures; Geographic terminology (+ orientation in space).

2. Features: 2.1. 2.2. 2.3. 2.4. 2.5. 2. 6. 2.7. 2.8.

Color; Size, form, and location in space; Number; Limitation, comparison; Order, classification; Modality of action; Aspect of action (+ time); Evaluation.

3. Situations ( states or actions + objects and features related to them); 3.1. 3.2. 3.3. 3.4. 3. 5. 3.6. 3.7. 3. 8. 3.9. 3.10. 3.11. 3.12.

Time; States of the weather and related objects and natural phenomena; Being, existence (+causation); Possession; Physical Impact; Physiological actions and states (and their causatives); Sense perception; Speech and sound production; Emotional states; Intellectual activity; Intellectual impact (+ speech); Mutual action;

104 3.13. Movement (+ causality); 3.14. Greetings, polite forms; 3.15. Work activity. 4. Replacives: 4.1. 4.2.

Personal, possessive, and reflexive pronouns; Interrogative, demonstrative, quantor, indefinite pronouns and adverbs (this group follows the classification proposed by A. M. Barulin).

2. In many thematic groups (especially the class of "Situations ") morphological (by parts of speech, for example: nouns-verbs) or even smaller segmentations (for example: 'action' - 'feature of the actor'; 'action - 'actor' - 'instrument' - 'place of action') are introduced. This classification is expressed by columns within each thematic group. Each dash in a column represents a word nest with a common lexical meaning. Within each concrete thematic group the organization of the words by lines is more or less motivated by semantic considerations (if possible words having a common part in their interpretations are placed together). 3. Some words are marked with an asterisk, indicating a certain lack of correspondence between that word and the classificatory principle: either to the feature of the semantic class in the title or to the structure of the word nest in which it is found. 1. OBJECTS 1. 1. Names of persons by age and sex [ Nos. 1 - 1 0 ] 1. 2. 3. 4.

man child little girl little boy

5. 6. 7. 8.

boy girl man woman

9. old man 10. old woman

1. 2. Kinship [Nos. 11-45] 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

family relative parents father mother husband wife son daughter brother (elder, younger)

11. sister (elder, younger) 12. father's father 13. mother's father 14. father's mother 15. mother's mother 16. grandson 17. granddaughter 18. husband's father 19. husband's mother

20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

wife's father wife's mother son-in-law daughter-in-law brother-in-law sister-in-law father's brother mother's brother father's sister mother's sister

105

1.2. Kinship [ Nos. 11-45] 30. male cousin (father's 32. female cousin (father's 34. nephew side) side) 35. niece 31. male cousin ( mother's 33. female cousin ( mother's etc. side) side) 1.3. Parts of the human body [ Nos. 46-88] 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

head foce forehead brow eye eyelash pupil nose cheek mouth lip tooth tongue chin beard

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

temple ear nape of the neck hair neck body back chest breast side belly hip shoulder arm elbow

31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

palm finger nail leg knee heel brain heart stomach lung skin bone blood

23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

deer foal dog male dog bitch cat ( male) cat (female) kitten cock hen chick

1.4. Domesticated animals [Nos. 89-120] 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

animal livestock (large) livestock (small) ram sheep (female) lamb bull cow calf heifer billy-goat

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

nanny-goat lid horse stallion mare gelding foal ass she-ass mule deer

1.5. The animal kingdom (animals, birds, fish, insects) [Nos. 121-142 ] 1. 2. 3. 4. 5. 6. 7. 8.

animal (furred) bear wolf fox hare aurochs sable polar fox

9. 10. 11. 12. 13.

bird hoopoe sparrow seagull eagle

14! fish ' 15. pike

17. grayling 18. humpbacked salmon 19! 20. 21. 22.

ny' bee beetle ant

106 1.6. The vegetable kingdom [Nos. 143-167 ] 1. 2. 3. 4. 5. 6. 7. 8. 9. 10i.

tree shrub trunk branch leaf bark root birch aspen fir

11. cedar 12. grass 13. flower • • a . • 14. berry 15. strawberry 16. cloudberry 17. cowberry

18. 19. 20. 21.

fruit apple pear apricot

22. 23. 24. 25.

vegetable onion carrot potato

5.

star

13. 14. 15. 16. 17.

iron gold silver bronze wood

13. 14. 15. 16.

stove, hes stairs balcony porch

, , ...

1. 7. Heavenly bodies f Nos. 168-172 1 1. *sky 2. sun

3. full moon 4. moon

1. 8. Inanimate nature [ Nos. 173-189 ] 1. 2. 3. 4. 5. 6.

soil sand clay turf dust dirt

7. 8. 9. 10. 11. 12.

stone pebbles gravel water air fire

1. 9. Housing f Nos. 190-205 | 1. 2. 3. 4. 5. 6.

house courtyard gates fence room roof

7. 8. 9. 10. 11. 12.

wall floor corner door lock window

1. 10. Everyday household objects f Nos. 206-229] 1. •thing 2. table 3. chair 4. bed 5. rug 6. skin 7. shelf 8. mirror 9. lamp

10. 11. 12. 13. 14. 15. 16. 17. 18.

drawer chest dishes spoon fork knife mug cup bowl

19. 20. 21. 22. 23. 24.

plate basin kettle pan pot pitcher

. . . • •

107

1.11. Clothing, footwear, decoration [ Nos. 230-253 ] 1. 2. 3. 4. 5. 6. 7. 8.

clothing footwear underwear [men's top clothing ] [women's top clothing ] headgear handkerchief shirt

9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

dress skirt pants belt button buttonhole buckle pocket stockings cuffs

19. boots 20. shoes 21. felt boots 22. ring 23. bracelet 24. earring

1.12. Food [Nos. 254-269] 1. 2. 3. 4. 5. 6.

meat fish bread flour sugar salt

7. 8. 9. 10. 11. 12.

onion garlic egg milk cheese butter

13. 14. 15. 16. , ,

sour cream soup groats tea

...

1.13. Geometric concepts, measures [ Nos. 270-282 ] 1. 2. 3. 4. 5.

part piece half quarter third

6. 7. 8. 9. 10.

middle side edge center circle

11. sphere 12. line 13. stripe

1.14. Geographic terminology (+ orientation in space) [Nos.283-335 ] Geographic objects (+ absolute orientors) 1. 2. 3. 4. 5. 6. 7.

earth dry land mountain knoll hill billock top of the mountain 8. base of the mountain 9. slope 10. pass 11. canyon 12. cliff

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

precipice river stream estuary lake swamp sea bay strait gulf shore spit

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36.

forest grove valley field road bridge town village •north •west •east •south

108 Relative orientors (relacives) 1. 2. 3. 4. 5. 6.

up down in front of behind around from one side

7. 8. 9. 10. 11. 12.

here on the spot there (come) here (go) there far

13. 14. 15. 16. 17.

near everywhere forward backward back

9. 10. 11. 12.

green yellow red brown

2. FEATURES 2 . 1 . Color [Nos. 336 347] 1. 2. 3. 4.

black white grey light

5. 6. 7. 8.

dark bright dark blue light blue

2 . 2 . Size, form, position in space f Nos. 348-369 1 1. 2. 3. 4. 5. 6. 7. 8.

large small high low upper lower wide narrow

2. 3. Number [ 1. 2. 3. 4. 5. 6. 7. 8. 9.

one two three four five six seven eight nine

9. 10. 11. 12. 13. 14. 15. 16.

thick thin long short deep shallow straight curved

17. 18. 19. 20. 21. 22.

near distant right left round acute

19. 20. 21. 22.

nineteen twenty thirty forty

1.370-393] 10. 11. 12. 13. 14. 15. 16. 17. 18.

ten eleven twelve thirteen fourteen fifteen sixteen seventeen eighteen

23. one hundred 24. one thousand

2 . 4 . Limitation, comparison f Nos. 394-403] 1. 2. 3. 4.

only almost too much nonetheless

5. 6. 7. 8.

except also even like

9. as 10. equivalent

109 2 . 5 . Order, classification [ Nos. 404-415 ] 1. 2. 3. 4.

order turn aspect first

5. 6. 7. 8.

9. 10. 11. 12.

second other following latter

extreme last remaining different

2 . 6 . Modality of action [Nos.416-4351 1. 2. 3. 4. 5. 6. 7.

to attempt to be able to one may one may not let should without fail

8. 9. 10. 11. 12. 13. 14.

of course it is obvious that it is true that actually obviously it turns out that apparently

15. possibly, it is possible that 16. perhaps 17. probably 18. in my opinion 19. it seems that 20. as though

2 . 7 . Aspect of action (+ time) [ Nos.436-450 ] 1. 2. 3. 4. 5.

again often rarely once usually

6. 7. 8. 9. 10.

sometimes always never already soon

11. 12. 13. 14. 15.

a long time at once early late over again

2.8. Evaluation [ Nos. 451-520 1 Features of objects 1. 2. 3. 4. 5. 6. 7.

the (definite) a ( indefinite) only, sole beautiful good bad correct

8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

slow fast rich poor easy difficult heavy strong soft hard

Features of actions and features

1. 2. 3. 4. 5. 6. 7. 8.

beautifully well bad(ly) correctly little much slowly quickly

9. 10. 11. 12.

easily difficultly heavily strongly

110 Features of objects

Features of actions and features

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

13. simply

simple complicated real new old usual dangerous useful basic main important special great pleasant strange convenient hot cold warm tasty sweet bitter salty sour fresh

14. 15. 16. 17.

really newly as of old usually

18. usefully

19. strangely

20. 21. 22. 23. 24. 25. 26. 27. 28.

very much enough quite approximately completely totally fully hardly barely

The columns in this and following groups are arranged so that lexemes which are related by a certain intellective relationship are grouped together in each line (see explanation 2 ) . 3. SITUATIONS ( states or actions + objects and features related to them) : 3 . 1 . Tune [Nos. 521-569] Names of intervals of time

Temporal features of actions

1. 2.

1. *this year

time year

Ill Names of intervals of time

Temporal features of actions

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

2. 3.

last year next year

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

in the spring in the summer in the fall in the winter in the morning in the afternoon in the evening at night at sunrise at sunset at dusk

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

yesterday the day before yesterday today torn morrow the day after tomorrow earlier then a long time ago recently now at this moment afterwards then, following

the last year the next year month week twenty four hours hour minute spring summer fall winter morning afternoon evening night sunrise sunset dusk [ the days of the week ] [ months of the year ]

3.2. States of the weather and related objects and natural phenomena [ Nos. 570-614 ] States, objects, phenomena 1. 2. 3. 4. 5. 6. 7. 8. 9.

weather inclement weather fine weather hot weather drought frost cold hard frost rain

Typical means of functioning 1.

to be stormy

2.

to rain, to pour

112 States, objects, phenomena 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

rainy season thunderstorm thunder lightning hail thunderhead cloud snow snowfall ground wind snowstorm blizzard ice snowdrift wind whirlwind tornado hurricane tsunami fog dew rainbow northern lights mirage

Typical means of functioning

3. 4. 5.

to thunder flashes beatsdown

6.

falls

7.

blows

8.

rages

9. hangs 10. clings 11. shine 12. shimmers

3. 3. Being, existence (+ causative) [ Nos. 615-669 ] Verbs

Nouns

1. to be, to exist 2. to be located 3. to become 4. to create 5. to spoil (trans, and intrans.) 6. to perish 7. to rot 1. 8. to melt 9. to kindle 1 0 . - 1 . to crumble (trans, and intrans.) 1 2 . - 3 . to spill (trans, and intrans.) 14. to live 2. 3. 15. -6. to be bom, to bear 17. to grow 4. 18. to die 5. 19. to kill 6. 2 0 . - 1 . to begin (trans, and 7. intrans.)

rot

life birth growth death murder beginning

113

Verbs

Nouns

22.- •3. to end (trans, and

8.

end

9.

continuation

intrans.) 24. •-5. to continue (trans, and intrans.) 26.--7. to ready (trans, and intrans.) 28.-•9. to repeat (trans, and intrans.) 30. to happen 31. to produce 32. to seem 33. to turn out 34. to be present 35. to be absent 36. to disappear 37. -8. to hide (trans, and intrans.) 39. to tidy up 40. -1. to preserve, to be preserved

10. readying 11. repetition 12. happening

13. presence 14. absence

3.4. Possession [ Nos. 760-708 ] Verbs (actions or states) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

to to to to to to to to to to to to to to to to to to to to to to

have obtain take accept take away receive borrow lead collect grab catch choose steal find search for lose leave give hand over transmit return give a gift

Names of actions and objects which are the result of actions -

1.

collection -

2. 3. 4. 5. 6. 7.

catching choice theft find search loss -

8.

transmiss

9.

gift

-

114

Verbs ( actions or states)

Names of actions and objects which are the result of actions

22. 23. 24. 25. 26.

10. proposition

to propose to change to sell to buy *to cost

11. sale 12. purchase 13. * price

3 . 5 . Physical Impact T Nos. 709-785 ] Actions 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

to to to to to to to to

rock wave shake push tighten twist tie fold butt bite tickle scratch crack scrape rub clean dirty

18. to sharpen 19. to bend 20. to crush 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

to to to to to to to to to to to to to

squeeze divide break chop crumble cut break up destroy beat mix glue on pin on add to

Features of the object of the action

1. 2. 3. 4.

tightened twisted tied folded

5.

bitten

6. 7. 8. 9. 10. 11. 12. 13. 14.

clean dirty easily soiled sharp curved straight * soft tough résiliant

15. fragile 16. 17. 18. 19.

beaten mixed glued pinned

115

Actions

Features of the object of the action

34. 35. 36. 37. 38.

to to to to to

39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

to spread to open to close to measure to repair to touch to paw to touch (intrans.) to smooth to kiss

20. 21. 22. 23. 24. 25. 26. 27. 28.

grease cover wet dry fill

-

greased covered wet dry full empty spread open closed -

29. •in good -

3. 6. Physiological actions and states (and their causatives) [ Nos. 786-862 ] Actions, states

Features of the actor

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

1. 2.

eating feeding

3.

chewing

4.

drinking

to eat to feed to breast feed to chew to swallow to suck to drink to give to drink to chomp to hiccup to smack one's lips to sneeze to cough to sniffle to breathe to sniff to go blind to go deaf to urinate to defecate to freeze to shiver to sweat to swell to fall asleep to put to sleep

5. *to have a cold 6.

sniffling

7. 8.

blind deaf

9. sweaty 10. swollen 11. sleepy

116 Actions

Features of the object of the action

27. 28. 29. 30.

12. sleeping

to sleep to awake to waken -1. to wash (trans, and intrans.) 32. -3. to comb (out) (oneself) (trans, and intrans.) 34. -5. to shave (trans, and intrans.) 3 6 . - 7 . to dress (trans, and intrans.) 38. -9. to undress (trans, and intrans.) 40. to smoke 41. to be sick 42. disease 43. health 44. to mature 45. youth 46. old age 47. to get fat 48. to lose weight 49. strength 50. weakness

13. *waking, awake 14. washed 15. dishevelled 16. shaven 17. dressed 18. *naked 19. sick 20. 21. 22. 23. 24. 25. 26. 27.

healthy grown up young old fat skinny strong weak

3.7. Sense perception [ Nos. 863-877 ] Verbs 1.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

to feel to listen to hear to look to see to look around to observe to notice *to discover *to find out *to clarify

Nouns 1.

feeling

2. 3. 4.

hearing look vision

-

-

-

3. 8. Speech and sound production [ Nos. 878-910 ] Verbs

Nouns

1.

1. 2.

to speak

speech *word

117 Verbs

Nouns

2. tell 3. converse 4. communicate 5. name 6. shout 7. be silent 8. to sing 9. to whisper 10. to whistle 11. to wheeze 12. to howl 13. to greet 14. to say goodbye 15. to low 16. to roar 17. to mew 18. to squeak 19. to chirp etc.

3. 4.

story conversation

5. 6. 7. 8. 9. 10. 11. 12.

*name shout silence singing whisper whistle wheeze howl

13. squeak 14. chirp

3.9. Emotional states [ Nos. 911-984 ] Actions and states Verbs 1. 2. 3. 4. 5. 6. 7.

to to to to -

to to to -

8. to 9. to 10. to 11. to 12. to 13. to -

Nouns be pleased with 1. love be attracted by hate 2. 3. be upset 4. fear 5. be frightened 6. 7. frighten suffer 8. grieve 9. 10. cry 11. mourn be sad 12.

14. to become furious 13. 15. to become angry 14. 15. 16. to envy 16. 17. to smile 17. 18. to laugh 18.

-

love -

hate indifference agitation fear fright cowardice -

suffering grief pain cry -

sadness -

fury anger goodness envy smile laughter

Features of the actor -

1. *loved -

2. upset -

3. timid 4 . cowardly 5. fearful -

-

6. sad 7. gloomy 8. furious 9. angry 10. good -

11. funny

118 Verbs

Nouns

19. to enjoy oneself 20. to be glad

19. 20. 21. 22. 23. 24. 25. 26.

21. 22. 23. 24. 25.

-

to to to to to

be satisfied be surprised hope desire be proud

-

enjoyment joy happiness satisfaction surprise hope desire pride -

-

26. to be lazy 27. to try

-

27. laziness

Features of the 12. 13. 14. 15. 16.

merry joyous happy satisfied surprised -

17. 18. 19. 20.

-

proud calm serious lazy -

-

3.10 Intellectual activity f Nos. 985-1014 ] Actions (verbs)

Actions or their result (nouns)

Features of the actor

1. 2. 3. 4. 5.

to to to to to

1.

thoughtful

6. 7. 8. 9. 10. 11. 12. 13.

to to to to to to to to

1. 2. 3. 4. 5. 6. 7.

2. 3. 4. 5.

dreamy believing intelligent stupid

6.

forgetful

-

think consider dream believe know

thought opinion dream faith knowledge stupidity memory

remember forget understand study think up 8. mistake be mistaken doubt 9. doubt be interested in 10. interest

-

-

-

7.

-

interested

3.11. Intellectual impact (+ speech) [ Nos. 1015-1107 ] Actions (verbs)

Actions or their result (nouns)

Actors, objects, and their features

1. to educate 2. to teach

1. education 2. studies

1. educated 2. teacher 3. pupil

3. 4. 5. 6. 7. 8.

to to to to to to

explain convince persuade advise show promise

3.

advice

4.

promise

advisor

119 Actions (verbs) 9.

to deceive -

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to

ask plead demand permit forbid threaten revenge curse swear complain praise insult care upset calm help hinder bore torture feel sorry for ask answer agree submit depend on argue refuse punish forgive joke send dispatch let go

Actions or their result (nouns)

Actors, objects, and their features

5. 6. 7.

lie truth request

5.- 6. lying, liar 7. honest 8. petitioner

8. 9.

demand permission

-

-

9.

-

-

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

threat revenge curse swearing complaint praise insult care upset peace help

10. threatening 11. revengeful 12. accursed -

13. plaintive -

14. insulting 15. careful -

16. helper

-

-

-

21. 22. 23. 24.

-

torture pity question answer

17. torturer 18. pitiful -

-

-

19. submissive

-

25. 26. 27. 28. 29.

demanding

argument refusal punishment forgiveness joke -

-

20. arguer -

21. joking 22. messenger

-

-

-

-

3.12. Mutual action [ Nos. 1108-1154 ] Actions or states Verbs

Nouns

Actors

1.

1. acquaintance

1.

acquaintance

2. 3. 4.

2.

one whom one meets

2.

to get acquainted to meet

3.-4. to kiss (trans. and intrans.)

meeting appointment kiss

120 Actions or states Verbs 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Nouns

to get married to marry (man) to marry (woman) to be married 6. to divorce 7. 8. to reach an agreement to quarrel 9. to fight 10. to battle 11. 12. war 13. to make peace 14. to make friends 15. *to be a guest

marriage matrimony divorce agreement quarrel fight battle enemy war peace friendship

Actors 3. 4. 5. 6. 7.

newly-weds groom bride husband, wife divorcée -

8. quarreler 9. fighter 10. battler, warrior 11. opponent 12. enemy -

13. friend 14. host 15. guest

3.13. Movement ( + causality) [ Nos. 1155-1257 ] Actions, states Verbs

Nouns

2. 3.

walking riding, trip

1. to stir (trans.) 2. to move ( trans. ) 3. *to carry 4. *to drag 5. *to pull 6. to lead 7. to convey

4. 5.

walk ride

8.

6.

flight

1. 2.

to stir (intrans.) to move (intrans.) 1.

3. 4.

to go (on foot) to go (by conveyance) to go for a walk to go for a ride

5. 6.

7. to fly 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Movement + causality

to swim 7. to flow to drip 8. to sprinkle (intrans.) to crawl to climb to jump 9. to dance 10. to run 11. *to hurry(intrans.)12.

movement

swim drop

jump dance run hurry

to take someone for a ride 9. *to throw 10. to get somewhere, to hit 11. to pour 12. to drop 13. to sprinkle (trans.)

14. *to chase 15. to hurry (trans.)

Actions, states Verbs 18. to attack 19. to spin round 20. to approach 21. to arrive (on foot) 22. to arrive (by conveyance) 23. to move away from 24. to go away (on foot) 25. to leave (by conveyance) 26. to enter 27. to exit 28. to pass by 29. to split up (go different directions) 30. to rise 31. to descend 32. to walk across 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46.

Nouns

Movement + causality

13. spinning

16. 17. 18. 19.

-

14. arrival 15. arrival -

16. departure -

17. rising 18. descent 19. crossing

to jump across to follow to pass to fall behind to step on to lie (-movement) to lie down to sit (-movement) to sit down to stand ( - m o v e ment) to get up (from sitting or lying position) to stop 20. stop to fall 21. fall to hang ( - m o v e ment) —

-

to to to to

spin (trans.) bring to lead to convey to

20. to carry away 21. to lead away 22. to convey away 23. 24. 25. 26.

to to to to

lead in lead out lead by split up ( trans. )

27. to raise (trans.) 28. to lower 29. to lead across to carry across -

30. to put down -

31. to seat (trans.) -

32. to stand (trans.) -

33. to stop ( trans. ) 34. to drop 35. to hang (trans.) 36. to take down

3.14. Greetings, polite forms [ Nos. 1258-1267 ] 1. 2. 3. 4.

greetings hello good morning good evening

5. 6. 7. 8.

good day goodbye farewell thanks

9. please 10. don't mention it 11. no trouble etc.

122 3 . 1 5 . Work activity [ Nos. 1268-1414 ] Action

Actor

1. 2. 3. 4.

to to to to

work do dig plow

5. 6. 7.

to sow to plant to weed

8. 9. 10. 11. 12. 13. 14.

to to to to to to to

furrow reap mow pasture cut hair milk gather

-

2. 3.

ditchdigger 1. 2. plowman

-

-

reaper mower shepherd barber milkmaid -

16. to hunt 17. to spin

10. hunter 11. spinner

-

21. to tin 22. to pot 23. 24. 25. 26. 27. 28.

fisherman

3. -4. hoe, chopper

-

garden -

-

sickle scythe whip scissors milk pail - 1 . basket, bag 12. -3. net, pole

5. 6. 7. 8. 9. 10.

3.

-

earth field

-

4. 5.

meadow pasture

6.

cow shed

-

-

7.

fishing . .

14. rifle 15. -6. spinning wheel, spin die 12. weaver 17. loom 13. seamstress 18. -9. needle, thread 14. shoemaker 20. awl 15. blacksmith 21. -2. hammer , 8. forge anvil 16. tinsmith 23. potter's 17. potter wheel 18. painter 24. brush 25. saw 19. woodsman 26. axe 20. carpenter 27. plane 21. planer 28. nail

to paint to saw to chop to hew to plane to nail together 29. to construct 22. construction worker 30. to bathe 31. to launder 23. laundress 32. to sweep

shovel 1. plow (wood- 2. en plow)

-

-

4. 5. 6. 7. 8.

-

-

9.

20. to forge

Place of action

worker

15. to fish

18. to weave 19. to sew

Tool

9. 29. brush . . . 30. -1. soap, bucket . . . 32. -3. besom, broom

construction site

10. bath house 11. laundry

123 Action

Actor

Tool

33. to cook 34. to fry

24. cook

34. pot . . . 12. kitchen 35. -6. frying pan, spit 37. -8. kettle, tea p o t . . . 39. match 40. -1. camp fire, stove

-

35. to boil

-

36. to light 37. to burn

-

38. to extinguish 39. to wipe 40. to write 41. to read 42. to draw 43. to drive

-

25. reader 26. artist 27. driver .

44. to harness 45. to saddle 46. to row 47. 48. 49. 50. 51. 52.

*to *to *to *to *to *to to

Place of action

-

_ -

_ rest play 28. player get tired wait be late have time

42. rag 43. -5. pencil, pen, * letter 46. book 47. colors . . . 48. -9. wagon, sleigh 50. -1. yoke, shaft 52. saddle . . . 53. -4. boat, oar . . . 55. game

4. REPLACIVES 4 . 1 . Personal, possessive, and reflexive pronouns [ Nos. 1415-1434 ] Person 1. 2. 3. 4. 5. 6.

I you he, she we (inclusive and exclusive) you they

Person + reflexivity

Features of objects

1. 2. 3. 4.

I myself you yourself he himself we ourselves

1. 2. 3. 4.

my yours his, hers our

5. 6.

you yourself they themselves

5. 6.

your their

bfi

s