Categories of Eating in Tzeltal and Navaho Author(s): Brent Berlin Source: International Journal of American Linguistics, Vol. 33, No. 1 (Jan., 1967), pp. 1-6 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/1264023 Accessed: 18-08-2014 18:27 UTC Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to International Journal of American Linguistics. http://www.jstor.org This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions International Journal of American Linguistics VOLUMEX XX X TT January 1967 Number 1 CATEGORIES OF EATING IN TZELTAL AND NAVAHO BRENT BERLIN UNIVERSITY OF CALIFORNIA, BERKELEY 0. Problems and prospects My Tzeltal materials indicate that at 1. Navaho and Tzeltal least seven verbs of eating are in common 2. Tzeltal usage. The glosses for each class are based on 3. Modified treatment of Navaho an examination of the distribution of nom- 4. Generic versus historical explanations inal objects of each verb which I will discuss in detail below. 0. Herbert Landar's important note on Navaho verbs of eating' points up some in- TZELTAL teresting parallels in Tzeltal.2 I hope to out- Class Verb4 line below some of these similarities, utiliz- 1. Eating in general -tun ing methods analogous to those of Landar. 2. Chewy object with pulp expec- In addition, I will attempt to show that for torated the Tzeltal materials, conjunctive definitions 3. Meat -ti? for verbs of eating are at least plausible, a 4. Mushy or gelatin-like objects -lo? solution that Landar finds inappropriate 5. Individuated, hardish objects -k'us 6. Bread-stuffs -we? for his Navaho data. Finally, it is suggested 7. Foods which dissolve in mouth that more evidence concerning verbs of eat- with little mastication ing in language groups that lie between the Mayan and Southern Athabascan languages Landar points out, by utilizing what he may be worthy of collection in reference to calls a 'matricized method' for inspecting comparative problems in American Indian the distribution of topic comment construc- culture history. tions, that the Navaho verbs do not allow for conjunctive definitions, that is, one finds 1. For ease of reference, the Navaho verbs it impossible to provide unitary definitions of eating and their associated semantic cate- of the verbs in the set for the objects of such gories are presented below as given in verbs may belong to 'two or more alterna- Landar's note:3 tive classes'.5 NAVAHO The judgment that a term is susceptible Class Verb to a conjunctive vs disjunctive definition is 1. Eating in general yiy44? dependent (in Landar's and the present 2. Hard or chewy object yf?aal paper) on the analysis given the inventory 3. Long stringy object yflttog 4. Meat yisyal of the referent(s) of that term. Lounsbury's6 5. One round object yigkhit 4 Tzeltal verbs of eating are all radical transi- 6. Mushy matter yflts?ee? tive verb stems. There is some indication that a 7. Separable objects yisteel similar analysis given for the Tzeltal materials 1 Herbert Landar, Seven Navaho Verbs of Eat- will also be possible for Tzotzil verbs of eating. ing, IJAL 30. 94-96 (1964). 5 Anthony Wallace and John Atkins, The Mean- 2 Tzeltal is a Mayan language spoken in central ing of Kinship Terms, AA 62.65 (1960). Chiapas, Mexico. It has approximately 100,000 6F. G. Lounsbury, A Semantic Analysis of speakers. The dialect of Tzeltal from which the fol- Pawnee Kinship Usage, Lg. 32.158-194 (1956); The lowing data are drawn is that of Tenejapa, a Structural Analysis of Kinship Semantics, in dispersed community of approximately 10,000 Horace G. Lunt, ed., Proceedings of the Ninth speakers. International Congress of Linguists, (The Hague, 3 See fn. 1, p. 94. Mouton and Company, 1964) pp. 1073-1093. 1 This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions 2 INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS VOL. XXXIII usage of the term DENOTATA for the refer- The denotata of Tzeltal verbs of eating ents of a term is now common parlance, which will be of most importance in arriving especially in formal kinship analysis where, at the significata of such verbs are clearly for example, the biological kintypes a+m, the referents of the nouns with which they a+f, aOm7 are treated as the denotata of occur. It is also important to note that, the kinterms father, mother, and brother, in practically speaking, a 'complete inventory' English. From an examination of a more or of the referents of all verbs in the set is ad- less complete set of denotata of a term it is mittedly unfeasible. What is significant is often possible to abstract certain common that the inventories be 'complete' enough as features which allow one to state 'the neces- to allow one to offer tentative definitions. ary and sufficient conditions for membership The adequacy of these definitions is tested in it'. Lounsbury and others refer to such and revised on the basis of newly encoun- definitions as the term's SIGNIFICATUM or tered referents. The process is continued its COMPONENTIAL DEFINITION. The goal of until the number of errors one makes are seeking to provide UNITARY componential totally eliminated or are reduced to insig- definitions in terminological studies has been nificance. stated by Lounsbury as follows: I present below six Tzeltal verbs of eat- ing accompanied by a typical inventory of A componential definition represents a Boolean common nouns with which they occur. The class product, and is thus a 'unitary' or 'conjunc- tive' definition. It is assumed that the meaning of general verb for eating, -tun, due to its any term belonging to a properly defined para- peculiar semantic restriction to interogative digm-one whose semantic field is itself unitary- expressions, (see below) is not included in will be susceptible to such a definition. This is per- the matrix. haps a stronger item of faith than we have a right to hold at this moment; but it furnishes the moti- VERBS vation for the analysis of kinship systems at least. NOUNS -ti? -lo? -k'us -?'u? -we? -buO' We proceed from extensional definitions (defini- ?ic chili pepper X tions by listing of denotata) to intensional defini- ti?bal meat X tions (definitions by specification of distinctive 5ehbew mushroom X features). We feel that we have failed if we cannot k'in kidneys X achieve conjunctive definitions for every termino- sehk'ub liver X logical class in the system. Were we to compromise pu4 lungs X on this point and admit disjunctive definitions Oalub chicken's X (class sums, alternative criteria for membership) comb as on a par with conjunctive definitions (class ca? gizzard x products, uniform criteria for membership), there 4ukum stomach x would be no motivation for analysis in the first 6'id' blood X (x) place ... 8 binam brains xX 7 The notation is that developed by A. Kimball cin bak marrow xX Romney as seen in his paper, Kalmuk Mongol and sepu grease xX the Classification of Lineal Kinship Terminologies, lo?bal banana xX AA 67. Pt. 2, 127-141 (1965). An earlier statement ?alcag orange xX has been presented in A. Kimball Romney and 6'um squash xX Roy G. D'Andrade, Cognitive Aspects of English ?on avacado xX Kinterms, AA 66. Pt. 2, 146-170 (1964). 8 F. G. Lounsbury, The Structural Analysis of e.g. kinship, is quite distinct from the task involved Kinship Semantics, in Horace G. Lunt, ed., Pro- in providing such definitions in less 'terminolo- ceedings of the Ninth International Congress of gized' areas of vocabulary. I do not suggest that Linguists, (The Hague, Mouton and Company, I have isolated all of the relevant dimensions in 1964) pp. 1073-1093. [quote from p. 1074]. the domain of eating in Tzeltal. The definitions I realize that the task of providing componen- which I offer will allow, however, correct usage tial definitions within lexical sets where the se- of the respective verbs with newly encountered mantic components of the total domain are known, objects. This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions NO. 1 CATEGORIES OF EATING IN TZELTAL AND NAVAHO 3 melones cantalope x form as meat + mushroom + chili pepper + 4'usub grape x blood.10H owever, additional data indicate Wijinp otato x clearly that the three anomalous nominals pahc' pineapple x cab honey x are actually instances of semantic extention. maniana apple x (x) I am fairly confident, after lengthy eliciting, manko mango x (x) that these three nouns are the only instances bok greens x (x) of such extension. kulis cabbage x (x) It is of particular interest psychologically wanes raddish x hi young corn x to note that informants, rather than being ?ahan roasting ear x ignorant of, or confused by, the 'apparent' cenek' bean x anomalous nouns, readily offer folk theories ?askal brown sugar x of explanation when queried to do so. The (in chunks) noun ?iMc hili pepper, occurs only with the p'ol popcorn x k'o?os toasted tortilla x Tzeltal numeral classifier, koht, which in nues nut x turn occurs with (excluding ?i6) animal refer- sakil pumpkin seed x ents. (Note, for example, 6a?koht O'i? two tuP?ak onion x (animal class) dogs; cankoht mut four (ani- wale? sugar cane x mal class) chickens; ho?koht wakas five (ani- te?el ?isim corn x stalk mal class) cows.) The ethnolinguistic theory wah tortilla x is simple: Since koht occurs with animals kaslan wah bread x (flesh) AS WELL AS WITH ?ic, why would the pa*' tamale x occurrence of ?ic with the verb of eating, -ti?, lulse candy not also be explained on the ground that turn cab sugar cane residue ?ic too, in some sense, was conceptually flesh? 2. One distributional feature is immedi- As for ceh6ew mushroom, it should be ately apparent from the above chart that is noted that the designation of one variety of quite distinct from Landar's Navaho data. this class is ,ahal ti?bal, literally, red meat. With few exceptions-which will be dis- In discussing the major class affiliation of cussed below-the Tzeltal verbs of eating are all mushrooms (i.e., are they plants, ani- complementarily distributed expressions as mals, or what?) several informants offered defined by the nouns with which they occur. the descriptive phrase lumilal ti?bal, liter- It remains to be seen to what extent the ally, earth's meat. The use of -ti? with mush- denotata of each verb have in common a rooms thus appears to be a case of extention semantic feature or features which sets verified by the occurrence of the lexeme them apart from each other in the set. I will ti?bal meat in both general and specific refer- discuss each form in sequence. ence to the class. In observing the nouns which occur with In discussing the potential occurrence of -ti, one may note that all but three refer the noun c'ic' blood with the verb -ti?, and specifically to referents which may be glossed the verb -lo? (the less frequent occurrence is as meat or flesh. The three exceptions are ?i6 indicated in parentheses), one must again chili pepper, 5eheew mushroom,9 and c'id' utilize additional ethnographic information. blood. Without additional ethnographic data It becomes clear in elicitation sessions that one might be forced to offer a disjunctive the choice of the two verbs, -ti? or -lo?, is definition of the class implied by this verbal 10S ee fn. 1. Landar's suggestion is relevant 9 The term ceheew is a general form encompass- here: 'It is important to recognize the need for ing a set of thirteen specific types of edible mush- further ethnographic study of unexpected pres- rooms. ences and absences in the matrix', p. 95. This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions 4 INTERNATIONAL JOTJRNAL OF AMERICAN LINGUISTICS VOL. XXXII1 determined entirely by the physical form of vidual pieces, it takes the form -k'us, a verb the blood when eaten. If it is eaten in the which denotes individuated, masticatable ob- form of blood sausage, as it normally is, jects. As for bok greens, and kulis cabbage, then it occurs with the verb -ti?, fulfilling the verb -lo? is required when these foods are the criteria of that verb in that the blood is boiled, as is normally the case, resulting in a eaten in elongated sausage-like sections of soft, mushy mass. When eaten raw, imply- cleaned intestines. If it is consumed as ing mastication of pieces of individuated ob- mushy globules formed when fresh blood is jects, the verb -k'us is obligatory. (Inform- poured into boiling water, the verb -lo? is ants, in discussing the importance of the dis- appropriate, a verb which will be seen below tinction of amount of mastication required to denote a class of soft, gelatin-like foods. compare the eating of a ripe mango or an While the exact form of the referent of the avacado, -lo?, with that of a roasting ear or noun is perfectly predictable to a Tzeltal mouthful of peanuts, -k'us.) speaker given either verb form, greater The verb -O'u?h as a highly restricted dis- nominal specificity may be obtained by use tribution and is quite specialized in meaning. of the expressions sut'bil c'ic' blood sausage It occurs only with two nouns, wale? sugar (literally, pinched blood) and payem c'ic' cane and te?el ?i6im corn stalk. It may be boiled blood. I think it is evident that to eat awkardly but correctly glossed as ingestion flesh/meat is the basic meaning of the Tzel- of juices from foods exhibiting pulpy portions tal verb of eating, -ti?. which must be obligatorily expectorated,i .e., Due to the substantial overlap in the dis- sugar cane pulp or pulp from masticated tribution of nouns occurring with -lo? and corn stalks. -k'us, it will be useful to discuss these two As noted above, the verb of eating, -we?, verbs jointly. occurs only with nouns which denote bread- In examining the denotata that occur stuffs.11 with -lo?, one may observe that they refer The verb -bu^' occurs only with foods to foods which, in their natural or mature which are eaten by being allowed to dissolve (in the case of plant fruits) states have a in the mouth with little or no mastication, mushy-gelatin-like texture. The quality or i.e., lulse candy, or the candy like substance, texture of the referents which occur with -lo? tum 6ab, hard candy-like residue of sugar cane is referred to repeatedly by informants by juice. use of the attributive k'unil soft, mushy. The general eating verb, -tun, has a re- when attempting to offer ethnolinguistic stricted syntactic distribution. As far as I data on the 'meaning' of the verb. In addi- have been able to determine, -tun occurs tion to the noun c'ic' (see above) note the only in interrogative expressions and is partial over-lap in the distribution of -lo? utilized when the speaker does not know the and the verb -k'us in reference to the nouns referents being eaten. Informants note that maniana apple, manko mango, bok greens, if the referents were known, then the ap- and kulis cabbage (again the rarer occur- propriate specific verb of eating would be rences of a form are indicated in paren- used. Such interrogative expressions into theses). Here, as in the case of blood, the which -tun enters are: binti ya a?tun What is choice of -lo? or -k'us depends on the state that you are eating?, or ya bal stak' a?tun of the object when eaten. If the object is Is the thing you are eating or about to eat well ripened, i.e., soft and mushy in its gen- eral consistency, it occurs with the verb -lo?. 11T here is some indication that -we? occurs in certain contexts as a general verb of eating. It However, if the referent is not fully ripe, if would never be utilized, however, if the meal did it has a hard-like consistency which re- not include some bready portion, especially wah quires a good deal of mastication of indi- tortillas. This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions NO. 1 CATEGORIES OF EATING IN TZELTAL AND NAVAHO 5 edible? If a verbal form is utilized in re- found that they are all in complementary sponse, the specific verb of eating is em- distribution with every other noun in the ployed, e.g., ya hlo? lo?bal I'm eating (a) set, certainly suggesting a conjunctive defini- banana; ya hti? ti?bal I'm eating meat; ya tion. And, while nouns which occur with hk'us cenek' I'm eating beans. No native the verb to eat meat, yisyal, also occur with Tzeltal speaker would respond to the above verbs indicating one round object, separable interrogative expressions with an utterance objects, and long stringy objects,t here appears such as ya htun lo?bal* I'm eating (with the to be no question as to the basic meaning of generale ating verb)a banana. this class. Landar clearly recognizes this in In summary, each Tzeltal verb of eating his discussion of gender when he notes: is seen to be defined by a set of statable semantic criteria which are unique for each The Navaho verb is a predication which has a sub- verbal form. I have suggested that these ject and often an object; it qualifies, therefore, as a sentence-word. The seven verbs listed above can definitions, derived from an inspection of stand, that is, without any accompanying nouns, the distribution of each verb with its nom- as complete sentences. I can say yis al, for exam- inal referents and supplemented with ethno- ple, and that would mean to a Navaho auditor graphic data, are unitary and provide 'the I ate it, and he would know moreover that what I ate was meat and not let us say an apple or a necessary and sufficient conditions' that a cracker.12 referent must exhibit to be used appropri- ately in Tzeltal. A modified treatment of the Navaho data The possibility of testing the definitions would suggest that some, but not all, of the depends largely on the number of possible verbs of eating are subject to conjunctive referents that a particular verb may poten- definitions. tially take. For example, the denotata of the 4. In closing, some remarks concerning the verb -#'u? are limited, as far as I have been rather striking semantic parallels between able to determine, to two and only two Navaho and Tzeltal in a domain such as physical referents, sugar cane and corn stalk. verbs of eating may be worth making, even These forms exhaust the universe, or at if somewhat speculative. When one looks at least, the cultural universe, for Tzeltal the semantic categories suggested by Atha- speakers. One possibility of testing the bascan classificatory verb stemsl3 and nu- definition would be some artificial situation meral classifiers in the Mayan languages,14 where the investigator presented the inform- (categories which Landar correctly suggests ant with some plant stem with a juicy pulp that we treat as GENDER in its most general which one would expect to be eaten in the manner specified by -O'u?. Some classes, 12 See fn. 1. 13 Harry Hoijer, Classificatory Verb Stems in however, are open classes, i.e., mushy-gela- the Apachean Languages, IJAL 11.13-23 (1945); tin-like foods. For classes such as these, one Some Navaho Linguistic Categories, Lg. 27.111- is able to test one's formulations continually 120 (1951); William Davidson, L. W. Elford and on newly encountered, culturally common Harry Hoijer, Athapaskan Classificatory Verbs, referents. in UCPL 29.30-41 (1963). 14B rent Berlin and A. Kimball Romney, De- scriptive Semantics of Tzeltal Numeral Classifiers, 3. In reference to the Navaho materials, it AA 66, Pt. 2, 70-98 (1964); Brent Berlin, Tzeltal is not totally clear that Landar's negative Numeral Classifiers: A Study in Ethnographic conclusions at deriving conjunctive defini- Semantics, Janua Linguarum, series practica, tions for Navaho verbs of eating are correct (The Hague: Mouton and Company, In Press); Brent Berlin, Christopher Day and Nicolas A. for each of his seven verb forms. If one Hopkins, Numeral Classifiers in Three Mayan examines the distribution of nouns that Languages: Chuj, Jacaltec and Tzeltal (In prepa- occur with yflts?ee? mushy matter, it will be ration). This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions 6 INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS VOL. XXXIII sense),'5 one is led to look for GENERIC rather lection of verbs of eating in languages which than HISTORICAL explanations. This is lie between Athabascan and Mayan. strengthened by the occurrence of numeral A tendency to neglect materials which classifiers in several languages of South may bear on historical development is as America, Melanesia, and particularly, South- detrimental to our cumulative knowledge as east Asia, which appear to exhibit rather is the premature statement of historical con- similar semantic ranges. nection.17 However, I know of no functional fea- tures, or, in Hymes' terms, features 'inher- 17A portion of the data presented here was brought to light while engaged in ethnobotanical ent in all units within the frame of analysis',16 investigation among the Tenejapa Tzeltal. This which would predispose the segmentation of research is supported by the National Science 'eating' into the semantic categories which Foundation, grants GS-383 (Comparative Ethno- are seen in Tzeltal and Navaho. What is botany of Two Tzeltal Communities, A. Kimball clearly called for is a concerted effort at col- Romney and Peter H. Raven, Co-principal in- vestigators) and GS-1183 (Studies in Tzeltal 16 Herbert Landar, Class Co-Occurrence in Botanical Ethnography, Brent Berlin and Peter Navaho Gender, IJAL 31.326-331 (1965). Landar H. Raven, Co-principal investigators). This notes (personal communication) that he uses gen- financial aid is gratefully acknowledged. A Uni- der in reference to concordial sub-classes following versity of California faculty research grant made Lounsbury's usage as given in the latter's course the presentation of the paper in its present form in language and culture. possible. I appreciate the helpful criticisms of 16D ell H. Hymes [ed], Language in Culture and Herbert Landar on an earlier draft, as well as Society, (New York: Harper and Row) p. 567 discussions I have had with Terrence Kaufman, (1964). John Gumperz, Jan Brukman, and Paul Kay. This content downloaded from 95.102.60.81 on Mon, 18 Aug 2014 18:27:04 UTC All use subject to JSTOR Terms and Conditions