As a result of analogical change, a bundle of phonetic features may arise which occurs simultaneously only in a given morphological category. Such a segment can be accounted for by a low-level morphologically conditioned phonetic rule. A subsequent phonetic change may normalize the segment and delete the first rule from the synchronic grammar. While the diachronic grammar will include both phonetic rules, the resultant change will appear in the synchronic grammar as a change in the morphophonological structure of the underlying form. This principle is illustrated by the anomalous nasal vowel letter attested in certain Old Church Slavonic participles, and by the fate of the velar stop in adjectival and pronominal inflection in northern Russian dialects.

The traditional explanation of Proto-Romance geminate initials (as in Italian rafforzamento), in terms of a progressive assimilation subsequent to Latin, is here rejected in favor of a theory of cognation between the strikingly parallel initial morphophonemics of Romance and Celtic. To sustain this, it is necessary to redefine geminates as units and medial so-called single consonants as phonemically composite. Romance syntactic ‘gemination’ is thus replaced by syntactic ‘lenition’, as in Celtic. Lenition actually handles certain Italian features better than does gemination.

This study presents two main theses: (1) that distinguishing ‘primitive’ from ‘derivative’ forms and functions can be accomplished in a linguistic description by a base containing two types of rewrite rules, expansion and substitution; and (2) that adjuncts to a noun can, under certain conditions, dissociate from that noun and pass from ‘attributive’ to ‘adverbial’ function—in particular, that clausal modifiers to the nominative can so detach and become predicate appositives and absolutes. These two proposals suggest a number of corollaries—among them, that the major adverbial clause types (condition, concession, purpose etc.) are basically adjuncts to the subject noun.

The importance of the morphophoneme as a structural unit has not been fully recognized. Post-Bloomfieldians explicitly rejected this unit and methods of description making use of it. This paper contends that morphophonemes are indeed required for the most efficient description of morpheme alternants. It is further argued that morphophonemes are real phonological units and not simply abstractions fictions. These points are illustrated by a problem in Tera, a Chadic language of northern Nigeria. In order to account for word-final vowel alternations, some /i/'s must be interpreted as the morphophoneme |ə|, others as |i|. Surprisingly, allophones of preceding consonants are variably determined by these underlying morphophonemes.

Indo-Iranian ∗-rt- is represented in Avestan sometimes as /Ř/ (written š), sometimes as rt. It is maintained that /Ř/ is the normal phonological reflex of ∗-rt-, and that rt has been restored according to Kuryłowicz's ‘Fourth Law of Analogy’ in motivated categories where there was still awareness of a morpheme boundary between the /r/ and the /t/.

Russian stress alternations are usually considered capricious and unrelated to grammatical meanings, but this paper argues that that is true only on the phonemic level. An analysis of such alternations in terms of underlying morphophonemic entities and ordered rules makes it clear (1) that Russian stress shifts are clearly correlated with the grammatical categories of number and case, and (2) that all stress shifts in the Russian substantive can be described by a single alpha-switching rule with varying environmental conditions. By listing the unmarked singular stem in the lexicon, generating the plural stem therefrom, and then generating the several case forms within each number stem, one obtains a simplification of the stress-shift rules and a generalization of their applicability. It is thus shown that a number of Russian words heretofore considered exceptional can all be covered as members of a single class.

The following alternative analyses of the Korean vowel system are compared: (1) By the view of traditional phonemics, we recognize nine vowels /i e ø ε ə a u o/, plus autonomous /w y/. (2) By distinctive feature analysis, regarding w and y as on-glide realization of features [+round] and [+palatal], we recognize nineteen vowels: i e ε i ə a u o, wi we (= φ) wε wa wə, ve vε və va vu vo. (3) By the view of generative phonology, we recognize four vowels—high , low a, mid rounded o, and mid unrounded ə—with w and y as on-glides, and with y as a fronting off-glide. (4) If we extend the analogy of off-glide y, and recognize a backing off-glide w, then we have three vowels, ə a, and two distributionally free glides, w and y. It is argued that alternative (3) gives us the truest and simplest picture, when considered in conjunction with morphophonemic phonomena, and especially in the light of the generality that several phonological rules gain by this analysis. It is hoped that this paper illustrates the possibility of choosing the ‘descriptively adequate’ solution from among several competing, non-unique, ‘observationally adequate’ solutions.

Arabic verbs are traditionally considered to be inflected for number when they follow the subject, but not when they precede the subject. The resultant ‘nominal’ sentences of the former case can be related to the ‘verbal’ sentences of the latter case by a focus transformation which duplicates a noun before the verb and leaves the initial instance of the noun after the verb to be pronominalized. This pronominalization, in the case of subject nouns, is realized as the so-called number inflection of the verb. Such a treatment also simplifies the analysis of relative clauses, ostensibly subjectless sentences, and sentences containing copula pronouns.

A proposal for modifying the tagmemic model is here illustrated in the description of the Huixtec Tzotzil verb phrase. Pike's notion of slot-class (Longacre's function-set) is retained, but the close correlation of slot and class as properties of a unit of grammar (the tagmeme) is rejected. An attempt is made to partially specify the form of rules in the lexical hierarchy of Tzotzil, here termed the semantic component. Rules for relating its semantic, syntactic, and phonological components are also presented.

This paper presents positive evidence for a genetic relation between Dravidian and Uralian. The major evidence consists of systematic phonemic correspondences between 153 Proto-Dravidian and Proto-Uralian lexical items. Impressionistic structural resemblances support the lexical evidence.

Similarities between Aymara and contiguous Quechua dialects have been considered by linguists simply as evidence of borrowing. This study, based on the application of standard comparative procedure to 520 Quechua sets and 255 Quechua-Aymara sets, gives evidence rather for genetic relationship between the two languages. A section on isoglosses clarifies the relationship between the northern and southern areas of the Quechua-speaking territory and proposes an older and important dialectal stage in central Peru.

Ten terms for Buddhist religious functionaries in standard Thai are subjected to semantic analysis. The data are grouped into a semantic set which is exploited to reveal redundant, hence predictable, semantic features. In addition, semantic patterning within the set is discussed and is related to the relevant subset of classifiers.