The class of phonetic changes in which the loss of a consonant is accompanied by the development of distinctive length in a nearby vowel has traditionally been designated by the term ‘compensatory lengthening’. We first discard cases in which the development of length and the loss of a consonant are in fact independent phenomena, and then analyse remaining examples. It is argued that such changes always involve loss of a consonant immediately adjacent to the vowel in which length subsequently appears, and that the change can be analysed quite generally into the two independently-motivated processes of weakening of a consonant to a glide (either semivocalic or laryngeal) and subsequent monophthongization of a complex syllable nucleus. The category of compensatory lengthening is thus shown not to be an independent mechanism of phonetic change. The fact that synchronic grammars sometimes contain rules of this type is argued to be an instance of the difference, in the general case, between phonetically-motivated processes and phonological rules. Conclusions are drawn for the study of the role of phonetic explanation in phonology.

It is further argued that the monophthongization involved in these examples leads to the development of length if and only if a length contrast is independently motivated in the system of the language concerned. This provides a particularly clear example of the (structuralist) principle that the phonological interpretation of a phonetic change is dependent on the structure of the system in which it takes place.

Recent times have seen a marked growth of interest in historical linguistics, as well as an increased role for diachronic factors within synchronic theory. This suggests that a fundamental re-assessment of the relationship between synchrony and diachrony in linguistic theory is in order, particularly with regard to its role in linguistic explanation. A number of ways are indicated in which diachronic factors enter at various levels as integral and at least equal partners in the over-all explanatory structure of linguistic science.

People readily create and understand denominal verbs they have never heard before, as in to porch a newspaper and to Houdini one's way out of a closet. The meanings are best accounted for by a theory of interpretation that specifies what the verbs mean on particular occasions of their use. Our proposal is that their use is regulated by a convention: in using such a verb, the speaker means to denote the kind of state, event, or process that, he has good reason to believe, the listener can readily and uniquely compute on this occasion, on the basis of their mutual knowledge, in such a way that the parent noun (e.g. porch or Houdini) denotes one role in the state, event, or process, and the remaining surface arguments of the denominal verb denote others of its roles. This convention accounts for the meaning and acceptability of innovative verbs in various contexts; similar conventions may be needed to account for other innovative uses of language.

This paper falls into a shorter general part and a longer special part. Part 1 is concerned with the various solutions which a speech community may seek in the event of a threatening collision of two homonyms or near-homonyms ; a careful distinction is drawn, in almost every category, between strictly lexical and lexico-grammatical homophony. Four characteristic patterns are established: (a) The speakers, after some wavering, in the end accept both partners, because of contextual disambiguation or because the words at issue pertain to different social dialects (or registers)—or else because one is common, and the other relatively rare, e.g. technical, (b) The stronger word dislodges the weaker word, eliminating it altogether or driving it into a corner, sometimes only temporarily; strength, in this context, is measurable by sheer frequency, or by smoothness of structural integration, or by ready availability vs. total absence of suitable substitute items, (c) The two words merge, producing a third word, usually of extended semantic ambit. (d) The partners are gradually differentiated in meaning and, above all, in form, through complicated and less than obvious analogical processes—with the further dichotomy of either (α) a single partner or (β) both partners undergoing such polarizing changes. Throughout Part 1, examples are adduced from a mixed bag of European languages.

Word-initial consonants may be lengthened in Italian under certain phonological, syntactic, and sociolinguistic conditions. This phenomenon is known in the literature as Raddoppiamento Sintattico (RS). The syntactic conditions on RS for two principal varieties of Italian are examined here, and it is proposed that RS is sensitive to the syntactic notion of ‘left branch’. Thus RS is a phonological rule which has access to surface tree-structure. The proposal here accounts in a simple way for great masses of data and also delimits the possible structural analyses of any given sentence in Italian.

Mixture of Spanish and English, whether in isolated loan words or in code-switching of clauses and sentences, while socially motivated, is subject to clear linguistic constraints. Quantitative analysis of mixing in conversations of Mexican-Americans suggests specific functional constraints to express tense/aspect/mood and subject/object relationships, as well as structural constraints which permit only surface structures which are grammatical in both languages. Resolution of structural conflict plays a key role, so that lexical cores trigger longer phrasal switches if they govern rules which create non-shared surface structures. The relative frequency of mixes without structural conflict is constrained by discourse function.

The development of contrastive tone because of the articulatory reinterpretation of segmentally-caused perturbations in intrinsic fundamental frequency is well attested in a number of unrelated languages. Considering the wide-spread character of this process, it is likely that its ‘seeds’ can be found in the functioning of the human articulatory and/or auditory mechanisms. This paper reviews what the authors consider promising explanations for well-attested tonal sound patterns, e.g. tone originating from the effect of prevocalic stop consonants or postvocalic glottal consonants, and tone rarely or never originating from the influence of postvocalic non-glottal consonants or from vowel height.

While tone terracing appears superficially to be a purely phonetic phenomenon, it is frequently subject to syntactic, grammatical, and lexical conditions which are not phonetically definable in surface structure; and this raises the question of how such conditions are to be encoded at the phonetic level to which pitch assignment rules apply. Some studies (e.g. Johnson 1972, Peters 1973, Anderson 1978) have proposed treating downstepped tones as separate ‘tonemes’ (distinguished from other tones in their feature characterization) to which pitch assignment rules apply in a left-to-right iterative mode of application. Such proposals fail to provide an explanation for the phonetic equivalence of basic and derived downstepped tones, and for the equivalence of downstepped and the corresponding non-downstepped tones with respect to following tones. In the alternative treatment proposed here, tone terracing systems are described in terms of a model which provides an independent level of representation to which intonational processes, including register lowering (downstep) apply; such a model is applied to the description of the tone terracing system of Akan. Further evidence for this model is drawn from Kikuyu.

The account given by Joos 1952 of the emergence of a contrast between laminai and apical sibilants in Old French and Old High German, and its reduction in Later Old French and Middle High German, is brought into sharper focus by examining the evidence of a 9th century MS containing a text in Old Picard and one in Old Rhine Franconian written by the same hand. This evidence suggests a re-interpretation of the development of the sibilant systems in these border dialects—and also in the ‘standard ‘literary dialects—in terms of separate features of manner and point of articulation for sibilants; and it raises questions concerning the relationship between contrast and sound change.

Morphological marking, whether realized by nominal case inflections or cross-referencing on the verb, can be either (i) ergative, marking transitive subject [A] function, vs. absolutive, marking intransitive subject [S] and transitive object [O]; or (ii) nominative, marking A and S, vs. accusative, marking O function. Absolutive is always the unmarked term in an absolutive/ergative opposition. Nominative is most frequently the unmarked term in a nominative/accusative system, but there are some languages in which accusative is unmarked. A language whose morphology mixes accusative and ergative marking has the split determined by (a) the semantic content of verbs, (b) the semantic content of NP's, (c) aspect/tense choice, or (d) a combination of these.

A, S, and O are universal semantic-syntactic primitives. A universal category of ‘subject’ can be defined as the set {A,S}, and is valid only for the level of deep structure. Language-particular syntactic operations, such as coördination and subordination, work in terms of a (shallow-structure) ‘pivot’; this is most often S/A, but can be wholly or partly S/O (languages of the latter type are said to be ergative at the syntactic level). A major function of antipassive or passive derivations is to place A or O NP's (respectively) in derived pivot function, S. Many languages which have some morphological ergativity are entirely accusative (S/A pivot) at the syntactic level. All languages that show syntactic ergativity have some morphological ergativity.

Dependency network representations, which eliminate the need for transformations, are proposed as a means for showing the essential semantic relationships among the components of sentences, especially those that contain oblique complements (including by-phrases). Evidence from Quileute, a Northwest American Indian language, provides strong support for the position, advanced by Langacker & Munro 1975, that passive sentences are basically agentless, and for their suggestion that all oblique complements have external sources.

English word stress is analysed in terms of rhythm—of alternating strong (S) and weak (W) syllables. A set of S-placement rules assigns S's to words. A set of rhythmic constraints defines well-formed configurations of S's and W's. The interaction of rules and constraints determines the accentual patterns of most words, whether simple or morphologically complex. The analysis accounts for all ‘stress shifts’ in morphologically related forms (e.g. solid, solidity), as well as certain phrasal stress shifts (e.g. thirteen students). The relation between S/W and degrees of stress is explored. Similarities are noted between the rhythms of discourse and those of verse. The analysis is compared to that of Liberman & Prince 1977, who also make use of alternating strong and weak syllables.

An attempt to construct a generative grammar of French with a coverage comparable to that of traditional grammars has failed. A description has been arrived at in the course of this work, however; it is much more complex than expected, and turns out to be entirely taxonomic. This result calls into question the validity of the so-called theory of generative grammar.

The material presented here is intended to illustrate the arguments presented in our article on Noun Phrase Accessibility (Keenan & Comrie [1972] 1977, hereafter NPA); it was originally intended that these data would be published as an appendix to that article. Exemplification is given for each strategy used to form relative clauses (RC's) in each language cited in Table 1 of NPA (pp. 76–9), except where adequate data appear in the body of the text. All information is based on work with native speakers, except where explicitly stated otherwise. Acknowledgments made in NPA (p. 63, fn.) also apply to the present publication.

In the Boston dialect, the variable or as in short, forty, has two phonetic variants: dialectal [Də], and more standard [Oə]. Speakers of three ethnic groups—Irish, Italian, and Jewish—showed systematic differences in stylistic and diachronic use of the variants, and in subjective evaluation of them. The observed differences are consistent with a three-stage model of variation and change based on ethnicity and, secondarily, on education. However, an additional factor of ‘cultural force’ is required to explain why the model may operate at different rates. The model has implications for the acquisition of dialectal vs. standard variants, and lends support to the view that the ‘family’ provides a stronger influence than the ‘peer group’ in the development of phonological systems.

One important function of the imperfective aspect in Modern Greek is to indicate indefinite repetition; but when a modal element such as ‘can’ or ‘must’ is present, the perfective may be selected instead. This apparent vacillation may be accounted for by making the assumption that indefinite repetition reflects what in the logical form of sentences is a structure ∀t (p at t ⊃ q at t)—where t stands for a variable ranging over the ‘occasions’ of events—and that the aspect of verbs, including those representing the modals, is determined by the relative scope of the universal temporal quantifier ∀t and the modal.

It has been shown by Keenan & Comrie 1977 that restrictive relative clauses (RC's) in many or perhaps all languages are subject to several constraints which make reference to the Noun Phrase Accessibility Hierarchy and to a classification of RC's based on their surface-structure properties. It is argued here that our understanding of RC-formation in universal grammar is further enhanced if a modified version of the classification of RC's, proposed informally by Givón 1975, is used instead of the classification of Keenan & Comrie.

Implicational universals of the form ‘If a language has some word order P, then it also has word order(s) Q’ are widely used to explain word order change. A language which develops word order P without having Q is said to be universally inconsistent, and it is predicted that it will re-introduce consistency by subsequently acquiring Q. This paper, however, argues against the validity of such a paradigm. The current assumption that the early western Indo-European dialects are inconsistent is founded on inadequate formulations of the synchronic universals. First, more precise formulations are given for some basic word order universals in a large body of languages ; then their predictions are tested against all the early daughter languages of Indo-European (using data from Friedrich 1975). None of these dialects exhibits inconsistency; hence we have no evidence for the reintroduction of consistency as a valid mechanism of syntactic change. It is shown, using historical data from Germanic, that implicational universals can nonetheless make correct predictions for the relative timing and manner of word order changes, but only on the assumption that languages in evolution obey synchronic universals. Some purely theoretical arguments are provided to support the empirical findings, and an alternative explanatory model of word order change is outlined.

The conversational historical present tense (CHP) occurs only in a specific type of narrative, here labeled a ‘performed story.’ In contrast to previous analyses of this feature, it is the thesis of this paper that CHP in itself has no significance. Rather, it is the switching between CHP and the past tenses which is the relevant feature. Two important constraints—that of sequence of tenses in coördinated sentences, and that of nonoccurrence of CHP in subordinate clauses beginning with when—support the theory that CHP alternation operates to separate events from one another. The fact that alternation is, in itself, an expressive feature, leads to speculation about the nature of other types of linguistic variation.