This article takes as its point of departure an unusual phonological change in a Czech dialect. It then proposes a model of phonological change which would make possible the understanding of structural innovations in the phonology of a homogeneous speech community. The model, which distinguishes two logically different modes of change (abductive and deductive), helps clarify the essential difference between ‘internally motivated’ change and change ‘induced from without’. The model uses our experience of observed phonetic changes, and may consequently have some bearing on our understanding of the structure of phonology.

Linguistic and psychological concerns are seen to converge in the area of semantics in such a way as to suggest the value, for both disciplines, of ‘psychosemantic’ studies. This paper illustrates some of the possibilities by providing linguistic evidence for three kinds of memory, called surface, shallow, and deep. The evidence deals with the occurrence of temporal adverbs in English and with their intonation and order, as well as with the non-generic perfective. A classification of memory by psychologists into sensory, primary (short-term), and secondary (long-term) is discussed, and is seen to coincide only in part with—though not to conflict with—the linguistically motivated classification. The relevance of memory to the use of tenses is briefly explored, and it is observed that the tripartite classification of memory is mirrored in our concept of the future.

Neither classical transformations nor global rules can account satisfactorily for the interaction of Quantifier Movement, Agreement, and Pronominalization in French. It is argued that an adequate treatment involves rules that enlarge networks of coreference in the course of syntactic derivations. Linguistic analogs of logical variables must be present at late stages of a derivation, giving rise to an interesting interplay between logic and syntax. The relevance of these facts with respect to some recently proposed extensions of the standard generative theory is discussed.

Truth is the most fundamental of all questions of appropriateness in language. Communication presupposes non-concealment between interlocutors, which logically excludes all forms of deception, not merely propositional lies. The lie, broadly conceived, is therefore a proper object of study for linguists, and a necessary one at a time when lying is cultivated as an art. As members of society, we have an obligation to contribute our skills in this as much as in other ways. Happily, a number of linguists have begun to respond by investigating the lies implicit in presuppositions, deletions, indirections, and loaded and jargonesque elements in the lexicon.

In Indo-European languages with mobile paradigmatic accent, fixed oxytones and movable athematic stems are accented by the same rule. When applied to syllables as accentual units, this rule generates the Sanskrit accent patterns. When applied to moras, it generates the Greek accent patterns—including, as a special case of mobility, the paradigmatic alternation of acute and circumflex. A modification of the the same rule accounts for the extension of mobility to originally oxytone -o- and -ā-stems in Bal to-Slavic, and for the rise of the Balto-Slavic acute and circumflex accents.

The claim that Rule insertion—rule addition anywhere in the middle of a grammar—is a possible type of phonological change is considered and rejected. Some putative cases of Rule insertion are shown to reduce to rule re-orderings governed by revised constraints on re-ordering; others involve non-controversial rule addition prior to the phonetic rules of a grammar; remaining cases of apparent Rule insertion are the fictitious consequence of incorrect analyses. New phonological rules can be added at only one point: at the end of the phonological rules but prior to the low-level phonetic rules.

Choice between different but equally informative proposals of ‘deep’ (or ‘underlying‘) structures appears to be arbitrary as well as undesirable. The appearance of rivalry among them is due to misrepresenting, in terms of alternative rules (constraints), what are in fact compatible tendencies. Realization of these tendencies, unlike the operation of rules, depends on specific lexical choice and context. Semantic tendencies operate within the framework of grammatical rules, but are no part of it. It is the freedom of ad-hoc choice among them, rather than any generative mechanism, that constitutes the truly creative aspect of language.

In many languages, focus constructions and constructions involving restrictive relative clauses show striking formal similarities, a fact which suggests that there is some deep, non-language-specific relationship between the constructions. The most tenable hypothesis about the nature of this relationship seems to be that both constructions involve the promotion of material from an embedded into a matrix sentence, a syntactic process that may be correlated with the semantic process of foregrounding one part of a sentence at the expense of the rest.

Germanic verbs of the 3rd weak class form presents characterized by an alternation between predesinential ∗ai (e.g. Go. 3sg. habaiþ) and ∗a (e.g. 1pl. habam). These verbs are usually compared with the ‘ē-verbs’ of Italic and Balto-Slavic, but no IE present built on the stative suffix ∗-ē- will account phonologically for the form of the suffix in Germanic. Instead, it can be shown that the characteristic Germanic paradigm results from the ‘activization’ of an older middle paradigm in which a 3sg. in ∗-ai (< IE ∗-oi; cf. Skt. duhé ‘milks‘) was further suffixed by the productive active ending ∗-þi (< IE ∗-ti).

Sentence qualifier markers, like those indicating interrogation and negation, are placed before verbs in consistent VO languages, after verbs in consistent OV languages. This placement principle permits us to identify other sentence qualifiers, like potential, desiderative, and reflexive. It also provides means for distinguishing qualifier markers from congruence as well as deixis markers. Moreover, it has morphological and phonological implications, leading to so-called agglutinative structure. Besides contributing to our understanding of linguistic structure, this principle enables us to clarify some problems in the historical development of languages.

The proposal made in Ross 1967, that transformations which move elements to the right are upward-bounded, is falsified by certain coordination reductions and certain pseudo-cleft constructions. Rightward movements are universally constrained by two weaker BEHAVIORAL RESTRICTIONS: (a) non-constituents are harder to keep in mind than constituents, and (b) two interrupted sequences which begin separately must not be separately continued. The second restriction independently accounts for the unacceptability of center-embeddings to a degree higher than one. The fact that SOME instances of pseudo-clefting are unacceptable, but not ALL, falsifies the claim made in Bach 1971 that Ross' proposed limitation on rightward movements constrains transformational components rather than derivations. The cases considered by Bach and Ross seem to be instances of grammatization and/or strengthening of the two proposed behavioral limitations.

This paper examines the main phonological properties of vowel harmony, with facts drawn exclusively from Uralic and Altaic languages. It is demonstrated that vowel harmony rules apply iteratively; and that exceptional stems and suffixes differ in that the former do not serve as context to the vowel harmony rule, while the latter do. Evidence is presented for the incorporation of abstract segments and absolute neutralization rules into phonological theory. Specifically, abstract and non-abstract analyses are compared side by side in a Turkish dialect, Buriat, Manchu, Vogul, Cheremis, and Hungarian: in each case the abstract analysis is shown to be superior or the only one tenable. The abstractness controversy is discussed in detail, and a principled resolution is offered.

Attempts to justify the principle of asymmetrical ordering have failed to consider some of the alternative relations among phonological rules and have ignored the problems raised by free alternation. Many free alternation patterns, like the i/y/iy, u/v/uv of Sanskrit and the complex vowel sequence alternations of Chorti (Mayan), cannot be generated economically by rules applicable only in fixed linear or cyclical order, but can be by rules applicable in alternative or symmetric order, a–b or b–a. Variable coefficients like Labov's are adaptable to systems of symmetrically or QUASI-ORDERED rules, permitting description of idiolectal and stylistic features in terms of the frequency of rule-choice—or, for closer approximations, of rule-sequence choice. A preliminary application of this approach to Chorti morphophonemic data leads to the isolation of three narrative styles and demonstrates the independence of the alternation rules within the community.

Even with the recent interest in correlating linguistic performance with sociological variables, the tendency in discussions of lexical borrowing is to focus on the words themselves and to ignore the process. This paper argues that the process of borrowing determines what is borrowed. Most borrowings may well be nouns for items new to the borrowing language. However, given sufficiently pervasive cultural contact, speakers may borrow certain types of core vocabulary items in great numbers. Data to support these contentions come from a study of the dialects of Ateso, an East African language.

This paper reviews the accentual patterns found in the various inflectional paradigms of Russian and develops a number of rules to characterize these patterns. It then surveys the most common accentual patterns found in different types of word formation, and discusses the manner in which the rules developed in the first part of the paper account for these patterns. A number of theoretical issues in phonology (cyclic rule application, the role of disjunctive ordering, etc.) are briefly discussed, as are certain accentual phenomena in other, mainly Slavic, languages.

Halle & Keyser cast their theory of iambic pentameter in the form of a set of rules against which lines of poetry are matched to ascertain the degree of correspondence; the result of the matching process is a description of the line in terms of its metrical complexity. A sub-part of their rules is examined here and found to be needlessly complex. A new set of rules for this sub-part is then proposed which is simpler and more revealing.

Although morpheme-structure conditions state generalizations about the distribution of morphophonemes of a language, there is no mechanism in the current theory of generative phonology that directly captures generalizations about phonetic patterns. This paper discusses the need for explicitly formulated surface phonetic constraints (SPC's) that state distributional constraints of segments and features at the phonetic level. On the one hand, SPC's are explicit statements about surface phonotactics and redundancies; on the other hand, they function as phonetic structure well-formedness conditions in sentence generation, SPC's will have direct relevance to discussions of the psychological reality of internalized phonological structure, rule conspiracy, lexical borrowing, and phonological change.

It is shown that presuppositions and implications associated with verbs distribute themselves relative to the time-axis (action-time) of the verb. Presuppositions pertain to the time preceding this time-axis, while implications pertain to the time following the time-axis. The semantics of two universal verb classes, modality verbs (M-verbs) and cognition verbs (C-verbs), is analysed in light of these findings.

Two widespread diachronic processes of sense extension are studied. One involves the development of sentence modalities from M-verbs. It is shown that a great amount of predictability exists in this process, and that the type of tense aspect marker likely to arise from an M-verb may be inferred, with great accuracy, from the specific presuppositional and implicational structure of the precursor verb. A considerable degree of predictability is also shown in the development of M-verb senses from C-verbs. It is suggested that our temporal and aspectual notions, including our notions of time, negation, certainty, and possibility, are already present in the semantic structure of verbs. Diachronic extensions of verb-senses merely serve to illuminate this.

An analysis of the deep structure and transformations involved in the syntactic derivation of long and short forms of adjectives and participles in Russian serves here as the basis for a discussion of the notion ‘parts of speech’ in a transformational grammar. It is suggested that a grammar must distinguish between deep-structure parts of speech (introduced by the base rules) and surface-structure or derived parts of speech, i.e. deep-structure categories in transformationally derived phrase-marker configurations which are impossible at the level of deep structure. For example, it is demonstrated that the active participle in modern Russian is a deeper verb transformationally introduced into the constituency of an NP; thus active participles will occur in Russian whenever a [… V …]NP configuration occurs in the surface structure.

The syntactic evidence provided by Robin Lakoff for the existence of a negative transportation rule is examined, and is found to be insufficient to establish the rule. It is shown that performatives are not the only verbs requiring tag questions to be formed on the next S down, and that this is of crucial importance in arriving at an adequate account of tag questions. Attention is drawn to certain semantic regularities which seem to account for the pattern of occurrence of tag questions without recourse to a rule of negative transportation. In the final section, an attempt is made to widen the field of discussion in the negative transportation debate by raising a new problem.