During late Old and Middle English, the distribution of short and long vowels in stressed syllables was profoundly altered. The changes involved have traditionally been understood as conspiring to optimize syllable quantity according to the position of the syllable in the word. However, Minkova's reformulation of so-called Middle English Open Syllable Lengthening (MEOSL) as a purely compensatory process appears difficult to reconcile with the traditional approach, which has recently been further compromised by suggestions that Trisyllabic Shortening was not a genuine historical sound change. In this article, Minkova's analysis is supported with new evidence of phonological conditioning behind the irregular lengthening of unapocopated disyllabic stems (e.g. raven vs heaven, body, gannet). I propose solutions to Riad's ‘data problem’ and ‘analytical problem’. Optimality Theory allows Minkova's revised statement of MEOSL to be integrated into a broader, non-teleological account of late Old and Middle English quantitative developments, including coverage of processes of lexical change such as borrowing and diffusion.

A characterization of the morphology of ‘regular’ Old English verbs is offered on the basis of a ‘word-and-paradigm’ approach to morphological structure. Thus rules of expression are formulated providing for the phonological realization(s) of (combinations of) morphosyntactic categories of tense, mood, person, number etc. However, it is argued that the organization of categories (in terms of compatibility and presupposition) differs with respect to the syntax–morphology interface and the morphology–phonology interface (the rules of expression). There is therefore proposed a set of rules of category realignment which relate the function-oriented system of categories to the expression-oriented system, which (among other things) minimize syncretism in the elements introduced by the rules of expression. Analysis of the preterite-present and weak-present verbs provides further support for this distinction in systems of category organization. The expression rules themselves do not introduce or realize ‘zeros’ and ‘mutations’ are interpreted as correspondences between base and derived/inflected form. An appendix offers a description of Present-day English verb inflexions for comparison.

This paper provides an account of be going to and will within the framework of Relevance Theory (Sperber & Wilson, 1986, 1995). Because of the range of interpretations derived from the use of these expressions in different contexts, many previous accounts have characterized them as polysemous. This polysemy has been attributed to semantic retention, whereby both old (lexical) and new (grammaticalized) meanings are recovered in certain contexts. The aim of this paper is to demonstrate: (i) that although be going to does exhibit semantic retention, in a relevance-theoretic framework this does not entail polysemy, and (ii) that interpretations of will previously attributed to semantic retention are, in fact, pragmatically derived, and hence will is also monosemous.

English testifies to a remarkable pattern of lexical gaps. In quite a few cases, particlestem nouns (e.g. the income) and adjectives (e.g. incoming) exist in the current language whereas their corresponding verbs (e.g. *to income) do not. Verbs with the opposite morpheme order (e.g. to come in) are, of course, attested. Since most relevant nouns and adjectives are derivatives of verbs, the word-class shift must either have entailed an inversion of the morphemes, or the verbs must have dropped out of the language. A diachronic analysis does indeed reveal that particle-stem verbs have a lower ‘life-expectancy’ than the corresponding nouns and adjectives. This difference is claimed to follow from differing degrees of compatibility between word classes and morpheme orders. The criterion of syntacticity shows particle-stem nouns and stem-particle verbs to be internally consistent but stem-particle nouns and particle-stem verbs to be inconsistent. This inconsistency gives rise to the observed historical instability of particle-stem verbs. While also disfavoured in this framework, the currency of stem-particle nouns (e.g. the handout) can be attributed to a conspiracy of several facilitatory effects which tend to override the arguments against this word type. The fact that linguistic units may fall into disuse necessitates certain revisions of the notion of optimality as the backbone of Optimality Theory.

This paper challenges the view that subject-orientation in English de-adjectival -ly adverbs is a syntactic attribute, and favours the hypothesis that orientation is, rather, a lexico-semantic feature. An important part of the evidence supporting this hypothesis derives from the examination of constructions in which an -ly adverb modifies an adjectival head (as in his genially informal manner) and displays a (potential) orientation to a co-occurring noun. The discussion is based on examples from the LOB Corpus.

Prevailing theories link the English periphrastic auxiliary verb do historically with Old and Middle English causative do. I argue that these and other accounts are inconsistent with modern dialect evidence and an analysis of the historical record suggested by that evidence. The primary source of periphrastic do was a habitual aspect marker which itself arose from the reinterpretation of bare object nominalizations as infinitive verbs.1