Surface alternations of English strong verbs suggest that they are highly irregular, and form at best a large number of small unrelated classes. But theoretical innovations, utilizing a markedness approach to characterize morphological irregularity, lead to the conclusion that these verbs are much more regular than they seem, and can be described in terms of seven conjugation classes. Six of the seven conjugation classes share, with minor variations, three ablaut rules. Class VII and one item of Class I require an umlaut rule. The ablaut and umlaut rules are not unique to verbs but are also needed to describe many noun and adjective alternations.

Hitherto, post-Saussurean linguistic theories have been unable to provide satisfactory descriptions of a creole continuum. This paper reviews previous attempts in the field, and indicates the need for a different theoretical orientation—one which would replace static, synchronic models of polar dialects with a single dynamic model incorporating both these and all the intermediate variations. In support of this argument, two sub-systems of the Guyanese creole continuum—the copulative and the pronominal—are described in terms of the dynamic evolution of the continuum as a whole; and it is demonstrated that, far from being an area of random variation, such a continuum represents a series of developmental stages ordered in accordance with basic principles of linguistic change. Finally, it is claimed that both the theory and the methodology advanced here cannot be limited to creole situations, but must have universal validity.

Resultative verbs in Mandarin Chinese are morphologically complex verbs which can undergo a certain affixation rule. Evidence is presented to support the claim that such verbs are best accounted for as being derived by a small set of lexical rules, rather than by syntactic transformations. The form and function of these lexical rules are discussed in detail.

Some characteristics of pronoun sequences are structured so as to necessitate the postulation of components of discourse structures—larger than the sentence—which 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.

Ancient Chinese metaphysics, as recorded some 4000 years ago in I ching, the Book of changes, expresses three fundamental characteristics of human thinking: (1) bipolar organization of the dimensions of cognitions; (2) attribution of positive polarity to the Yang pole and negative to the Yin; (3) parallelism in the orientation of the dimensions in terms of underlying Positiveness/Negativeness. In the present paper, eight postulates governing the dynamics of interaction among cognitions are developed, utilizing both linguistic and psychological evidence. In the theory, complex cognitions are reduced to simple cognitions consisting of signed and directed relations between pairs of meaning elements (e.g. M1 → M2) which in turn are analysed into simultaneous sets of bipolar semantic features capable of having only one sign and value at any one time. These postulates are used to predict English speakers' performance in a simple and specific type of cognitive interaction—the usage of and vs. but in the conjunction of pairs of adjectives. The measured affective meanings of the pairs of adjectives predict usage with high precision—in a frame which, linguistically speaking, will accept either conjunction—and the data are consistent with the three characteristics of Yang and Yin stated above.

The dying East Sutherland dialect of Scottish Gaelic shows grammatical change currently in progress in two syntactic environments, the passive and the case system. The rate of change has been rapid enough to show clear differences between the usage of the oldest and the youngest fluent speakers available, a span of just over 40 years. This permits the consideration of questions about the direction of change, the processes of change, and the rate of change, and of the utility of terminal speakers for systematic language description and historical reconstruction.

This paper challenges the arguments for abstract phonology based on the treatment of the Nupe vowels in Hyman 1970. Both formal and substantive deficiencies of Hyman's analysis are noted, and a non-abstract solution is proposed.

In this reply to Harms 1973, a critical look is taken at Harms' arguments against ‘abstract phonology’. Particular attention is focused on his unwarranted dismissal of the data from borrowing. In addition, it is argued that considerations of simplicity or economy do not provide any significant insights into the problem at hand. Finally, a modified approach to Nupe palatalization and labialization is outlined.