Sentential aspect presents a situation (event, state etc.) from a particular point of view: it represents the speaker's choice of perspective on the situation. In this article, a speaker-based approach to aspect is proposed which crucially invokes abstraction, namely the idealization of different types of situations. A speaker relates an actual situation to an idealized type by using the linguistic forms associated with that type.

Speakers may choose an unusual way of presenting a situation, as with a progressive stative such as I'm really loving this walk. Here a state is presented as an event. Analysis of stative sentences in English shows that they have a unique aspectual interpretation, and allows for the interpretation of progressive statives.

Two components of sentential aspect are recognized. SITUATION ASPECT involves type of situation, e.g. event or state; VIEWPOINT ASPECT involves type of perspective, e.g. simple or progressive. Separating these two components leads to a unified account of simple aspect in English.

Two types of bound morphemes—clitics and inflectional affixes—are found attached to (free) words in many languages. At least six lines of evidence separate the clear cases on each side: the degree of selection between the dependent morpheme and the word to which it is attached; arbitrary lexical gaps; phonological idiosyncrasies; semantic idiosyncrasies; syntactic operations affecting the combinations; and restrictions on the combinability of clitics with inflectional affixes. These criteria all indicate that English contracted auxiliaries (She's gone) are clitics, but that the English contracted negative (She hasn't gone) is an inflectional affix—a rather surprising conclusion that turns out to have satisfying consequences.

Two competing theories exist about the Proto-Germanic syllabication of and similar sequences: one proposes that the division was , the other, . Furthermore, three major phonological developments in Germanic groups have never received a satisfactory explanation: consonant gemination in West Germanic, vowel lengthening (with differential constraints) in North Germanic, and glide strengthening in Gothic. This paper proposes a uniform solution for all these problems, based on independently established principles of syllabic phonology: a preference (at least in the Germanic languages) for stressed syllables that have two morae (Prokosch 1939), and a universal preference for syllable onsets that are at least as strong as the preceding syllable offset (Hooper 1976). We provide various kinds of evidence for a Proto-Germanic syllabication which is in accordance with Prokosch's principle, but not with Hooper's; and we then explain subsequent developments as different ways of rebalancing the syllable structure to accord with both Prokosch's and Hooper's principles: in West Germanic; in North Germanic; and in Gothic.

A drastic restructuring of the Common Mayan system of pronouns resulted in the Modern Yucatecan pronominal system. This restructuring involved, first of all, a paradigmatic regularization of new, formerly non-existent grammatical categories.

The restructuring was internally constrained by the hierarchy of the pronominal system itself. The organizing laws which shaped the Mayan system are (a) the well-known theorem of Huntington 1904 that ‘the number of elements in every logical field must be 2m‘ and (b) a constraint on the theorem provided by the Peircean notion of SYMBOL (as against ICON). Grammatical systems of a more SYMBOLIC character formally distinguish relatively fewer 2m spaces. Common Mayan, which was a symbolic system, made fewer distinctions in the logical field of its pronominal system than do the modern, Yucatecan reflexes of that system.

If a sound change has lexically diffused without completing its course, one finds that among the lexical items qualified for the change, some have already changed (c), others have remained unchanged (u), and still others show variant forms (u/c). When such a change has affected a group of genetically related languages, the consequent comparative pattern u—u/c—c can be used to set up subrelations among languages. In this paper, we draw on data from six languages belonging to the South-Central subfamily of Dravidian, with reference to an atypical sound change called ‘apical displacement’. There are 63 etymologies which qualify for the study. A total of 945 possible binary-labeled trees fall into six types for the six languages under study. In terms of our postulates, that tree is the best which scores the lowest m, i.e. the minimum number of independent instances of change needed to account for the u-c-o (o = no cognate) pattern of a given entry. Each of the 63 entries has been applied to the possible 945 trees, and the trees have been scored for the value m by computer. The one tree which scored the lowest (71 points) is identical with the traditionally established tree for these languages. This paper shows that: (a) one shared innovation is sufficient to give genetic subrelations among languages, within the framework of the theory of lexical diffusion; (b) unchanged cognates are as important as changed cognates in giving differential scores for possible trees; and (c) the notion of shared innovation can be further refined within the theory of lexical diffusion.

Unmarked terms in the lexicon, compared to marked ones, are typically more frequent in use, less complex in form, and acquired earlier by children learning a language. Terms which are unmarked in single languages are often unmarked in all languages; however, marking is not always invariable across languages, or through time within individual languages. The present work focuses on variation in cultural importance as a factor which influences marking. As the importance of a referent changes within a speech community, the marking value of its label alters, often resulting in lexical change. Introductions of previously unknown referents in culture contact situations—e.g. domestic plants and animals—frequently have led to shifts in cultural importance. Such examples illustrate how cultural factors, by influencing the assignment of marking, often play an important role in lexical change.

Data collected from the Mayan language Quiché suggests that considerable differences may exist across languages in the form of children's first words. Several classes of Quiché words, including the verbs and the positional element k'o:(lik), take ‘terminations’. Although the termination does not encode a simple semantic or syntactic meaning, the children used it correctly in over 86% of their first verbs. The syllable structure and stress rules conspire in Quiché to promote the early acquisition of terminations, suggesting the primacy of intonation in morphological development.