Kenneth L. Pike

Immediate Constituents of Mazateco Syllables

1. It is well known that sentences have an internal structure which can be analyzed in terms of successive layers of immediate constituents.l Thus, the sentence Poor John ran away divides first into Poor John and ran away, then Poor John divides into Poor and John, while ran away divides into ran and away, and so on. It is convenient to describe syllables of Mazateco2 in a similar fashion. The structure of these syllables does not consist of a series of sounds equally related, like beads on a string, but is rather like an overlapping series of layers of bricks. The different

Dimensions of Grammatical Constructions

with one set of contrastive features in rows, another in columns, and with phones in the cells of the matrix? Would this approach allow for presentation-simply, comparably-of the construction types of a large number of languages as a basis for a universal etics of grammar? Our answer to these questions is affirmative. Several difficulties in early tagmemic descriptions would be lessened by such a presentation. Redundancy-most annoying, perhaps, of the characteristics of current tagmemic presentation-would be sharply cut. Where sentence (or clause, or phrase, or word) features have been repeated in order to allow for multiplied sentence types, single presentation of the intersection of two features could be made, with later cross reference to this total complex. Where a basic kind of sentence is modifiable in various ways for emphasis, or to make interrogative sentences, various devices have been used to avoid a listing of each basic type repeatedly, plus the modifications. Helen L. Hart2 sets up for Amuzgo a single emphatic sentence type to cover relevant emphatic variants of all basic sentences. Similarly, one query sentence type, one response type, and one quotative type are set up to handle varieties of all the basic sentence types. Velma Pickett for Zapotec3 points in the direction of dimensions by charts of constructions in reference to contrastive classes of clause4 and phrase construc-

On Tagmemes, Née Gramemes

0. Some recent concepts concerning universal grammar which we have begun to develop in the Summer Institute of Linguistics and at the University of Michigan have already proved fruitful. They have helped to provide a theoretical framework into which many hitherto awkward odds and ends of linguistic data fit as essentialeven crucial-parts of a total view of language structure; they have implied and pointed the way to analytical procedures which have already proved helpful in the solving of some difficult problems on the field, both of phonology and of grammar; and they have provided the background against which our pedagogy for the teaching of syntactic analysis has been effectively revised.

Sociolinguistic Evaluation of Alternative Mathematical Models: English Pronouns.

Some characteristics of pronoun sequences are structured so as to necessitate the postulation of components of discourse structures-larger than the sentencewhich can now be treated by the mathematical theory of groups (and in a manner accessible to the non-mathematician). Since dozens of variant formalisms can represent the six pronominal axes studied, criteria outside formalism are needed to evaluate them. In seeking for naturalness, sociolinguistic criteria (beyond grammar or lexicon as such) must be used. Once the choice of best representation has been made on sociolinguistic grounds, it turns out-to our delight-to combine, as mathematical characteristics, those of the minimum non-commutative group with those of the minimum (non-trivial) commutative group. This article studies the interplay of linguistics, social situations, and mathematical formalisms as they affect certain speaker-addressee relationships involving three people. In ?1 a few questions are raised about linguistic use of such relations, to suggest to the reader the general area of interest. ?2 develops, through the use of graphs, the small amount of mathematics needed by the mathematical layman in order to follow our argument. ??3-9 discuss and evaluate the many alternatives, partially ranking them in order of naturalness. ??10-13 return to more general considerations, and end with the conclusion that the choice arrived at empirically, for greatest naturalness, has interesting mathematical properties as well.*

Compound Affixes in Ocaina

The identification of the morphemes and morpheme classes of a language frequently presents difficulties. Some of these involve problems in morphemic theory itself. This is true of two problems in Ocaina,l one concerning morphemic segmentation and fusion, and the other concerning morphemic compounding2 as it distorts morpheme distribution classes. The data are presented and interpreted in ??1-2. The theoretical implications are summarized in ?3, where it is pointed out that the interlocking of the lexical and grammatical hierarchies must be distinguished from effects of the interlocking of the lexical and phonological hierarchies.

A Note on Allomorph Classes and Tonal Technique

0. In the late fall of 1950, I had the opportunity in the field to check Cornelia Maks data on the tone system of the dialects of Mixtec spoken in San Miguel el Grande and in San Esteban, with several of her informants. In addition to the striking differences in the complexity of the two systems, certain problems arose in the handling of the SE material which provided experience and/or warnings potentially valuable to other workers in register-tone languages.

A Text Involving Inadequate Spanish of Mixteco Indians

In Mock Spanish of a Mixteco Indian2 I showed how Mixteco tone and Spanish intonation were again mixed. This time, however, a Mixteco Indian was deliberately using incorrect tones in an attempt to mimic Spanish intonation and to persuade his monolingual neighbors that he had learned Spanish in addition to his own speech. The present text3 portrays a different aspect of problems arising in a community where speakers are passing from monolingualism to bilingualism. Here, two Mixteco speakers know only a very few phrases of Spanish4 and even these phrases they tend to use incorrectly. Some soldiers, searching for criminals, interrogate them. In replying, the Mixtecos use these phrases which are the only ones they know even though they understand neither the questions of the soldiers nor their own replies. By chance their phrases constitute a confession of murder, and as a result they are thrown into jail.