Published online by Cambridge University Press: 22 April 2026
Isthmus Zapotec belongs to the Zapotec language family spoken in the state of Oaxaca, Mexico. There are at least seven different “dialects” of the Zapotec “language,” but they are often considered to be separate languages, inasmuch as they are mutually unintelligible. The Isthmus dialect is spoken in an area bordering the Pacific coast in the Isthmus of Tehuantepec.
1 This represents a syllable stress-length component which was previously analyzed in terms of phonemic vowel length and nonphonemic stress, (cf. Velma Pickett, Nonphonemic stress : a problem in stress placement in Isthmus Zapotec, Word 7.60-65 [1951].) The present analysis, treating vowel length as dependent on stress, makes a somewhat simpler description and handles more easily certain recent Spanish loans. Further problems in high-level phonology must be solved before this is entirely satisfactory, but these do not affect our grammatical analysis.
2 The terms nucleus and satellite are taken from Richard C. Pittman's Nuclear structures in linguistics, Lg. 24.287-92 (1948). There is, however, one striking difference in their use here, which is discussed in Section 3.2.
3 C. C. Fries, The Structure of English 23 (New York, 1952).
4 For the stimulus which led to the analysis in terms of dependent and independent structures at each level, I am indebted to Viola Waterhouse, who made available to me her unpublished paper, Independent and dependent sentences.
5 See Section 2.2.1.
6 Including proclitics and enclitics, which are here classified as dependent words, in the syntactic description.
7 Velma Pickett, Isthmus Zapotec verb analysis I and II, IJAL 19.292-96 (1953) and 21.217-32 (1955).
8 Kenneth L. Pike, Gramemic theory, General Linguistics 2.39-40 (1957); idem, Language in relation to a unified theory of the structure of human behavior (preliminary edition, Part I, Glendale, 1954); and idem, On tagmemes, née gramemes, HAL 24.273-78 (1958).
9 In English, one word such as “ouch” can fill slots in all levels up through discourse (if discourse level is defined as I have done in Isthmus Zapotec). But such a tagmeme at discourse level would be the unusual case rather than the norm.
10 Note, however, that at word level, where types based on internal structure are not as readily definable, distribution classes are primary.
11 For further discussion of the problem of word classification (or identification of “parts of speech”) see Section 3.4.
12 Compare the following quotation from Zellig Harris (Co-occurrence and transformation, Lg. 33.285 [1957]). “On the other hand, to describe a language in terms of the cooccurrence of the individual morphemes is virtually impossible: almost each morpheme has a unique set of co-occurrents; the set varies with individual speakers and with time ...; it is in general impossible to obtain a complete list of co-occurrents for any morpheme; and in many cases a speaker is uncertain whether or not he would include some given morpheme as a co-occurrent of some other one.”
13 Frequent in occurrence either in terms of text count (the independent and dependent pronouns) or in terms of both text count and membership list.
14 One-member classes determined by unique phrase combinations are not included in the list. They are indicated in formulas by the form itself. For example, see Section 2.2.3.12.3.
15 The term “total list” implies a closed list of the members given and no more. In some cases the implication is considered by the writer to be an accurate one (e.g., the pronoun classes). In other cases the term is used to mean “total list of forms observed,” with the reservation of possible addition of a few more members after further study. The term is used in both senses for simplicity of statement.
16 Third person plural forms are phrases composed of quantifier ka “plural” plus third person singular words, e.g., ka bě “plural he/she” (= “they”).
17 Other persons are indicated by independent pronoun phrases composed of person base word ‘la?a and dependent pronouns or dependent-pronoun phrases, e.g., ‘la?a a bě “person he/she” (= “they/them”).
18 This subclass includes one word which does not fit the class meaning—'liǰi “home.”
19 Note, for example, the two following definitions of phrase: “any group of two or more words that form a sense unit, either expressing a thought fragmentarily or as a sentence element not containing a predication but having the force of a single part of speech.” Webster's New collegiate dictionary. “A free form which consists entirely of two or more lesser free forms, as, for instance, poor John or John ran away or yes, sir, is a phrase.” Leonard Bloomfield, Language 178 (New York, 1933).
20 Except for unique phrases and a few others where class names are not identifiable.
21 Some nouns seem to occur more frequently as heads of possessive1 phrase, others as heads of possessive2 phrase. The amount of freedom in this distribution varies from one speaker to another, however, and hence it is not considered to be a structural limitation.
22 Dashes (—) mark the breaks between clause-level tagmemes, both in the citation and the literal translation. Free translations are added where the literal translation is not clear. Where English requires two words for one word in Zapotec, hyphens are used between the translation words.
23 The preposed Independent Subject may also have a double function of sentence introducer when it is filled by some sequence form such as 'ómbre rí ? “this man,” referring to a person introduced previously. This is especially true when the subject is preposed to certain satellite tagmemes, as mentioned below.
24 The Purpose tagmeme has a semantic variant of benefactive. The meaning differentiation makes the homogeneity suspect, but in the present data the two variants cannot be adequately separated by formal characteristics.
25 Under certain conditions transitive verbs sometimes occur without the Object tagmeme, e.g., in response to a question which has already stated the object. These are considered to be dependent clauses; see Section 2.3.2.9.
26 Two clause nuclei in juxtaposition which are not united by the presence of a satellite tagmeme which functions with both nuclei are considered to be compound sentences made up of two clauses.
27 This phrase could alternately be treated as a special verb phrase, different from the other general verb phrases, which would add another class of three verb phrases and a further distribution class composed of the one word sǐ.
28 Base tagmemes in sentence, utterance, and discourse structures are labelled SBase, DBase, and DBase respectively. At each level, contrastive base tagmemes are distinguished by the filler lists given.
29 In Sections 2.5 and 2.6 only free translations are given for illustrations.
30 See the bibliography, Section 4.9, for other presentations under this model.
31 Robert E. Longacre, String constituent analysis, a paper presented to a linguistic colloquium at the Summer Institute of Linguistics, Norman, Oklahoma, 1958.
32 The clause nucleus is the principal unit under question, since it is the most consistently complex nucleus. It does contain one optional tagmeme, the Independent Subject, which might be considered to be satellite in an intermediate (between phrase and clause) level. The Independent Subject, however, is not of the same status as optional satellites in other structures, for it is a redundancy, a duplication of the function of the Dependent Subject, thus more closely united to the minimum form (obligatory tagmemes) than most satellite structures are to their nuclei.
33 Isthmus Zapotec verbs II.
34 E. A. Nida, Morphology: the descriptive analysis of words 110 (Ann Arbor, 1949): “An external distribution class consists of a set of forms which occur in the same environment.”
35 Charles F. Hockett, A course in modem linguistics 225-8 (New York, 1958).