European and Brazilian Portuguese (henceforth EP and BP) are known to diverge considerably insofar as pronominal clitic placement is concerned. The first part of this paper attempts to show that most of these syntactic differences can be accounted for by a general phonological contrast between EP regular enclisis vs. BP native proclisis. How could such a difference be explained?Actually, clisis may provide remarkable evidence for a new claim on Portuguese stress and rhythmical patterns. In a second part, various phonetic considerations are adduced to support the view that EP has a syllabic-weight contrast and that lexical stress placement then follows a leftward mora-counting system. In BP, however, there is no such syllabic and accentual quantity. Finally, it is stated that such a structural difference parallels the fact that vowel reduction affects all EP unstressed (light) syllables, but only BP posttonic and not pretonic syllables.This may be accounted for by assuming that right-to-left quantitative stress systems (such as EP) involve pretonic integration within the preceding foot (and hence pretonic floating mora at the lexical stage), while nonquantitative systems (such as BP) show true (linked) pretonic strings. Clearly, EP enclisis and BP proclisis appear thus as crucial evidence for what may be a particular example of the general distinction among rhythmical patterns, recently made by Wenk and Wioland, between 'leader timing and 'trailer timing , here applied to lexical stress languages, which would then display either (as EP)/foot -^pretonic/or (as BP)jpretonic + foot I structures, respectively.

This paper aims to show that sonority-based generalizations on consonant phonotactics should directly follow from representations, not from stipulations on representations such as the commonly accepted licensing or government statements. The basic reason for this is that the second approach is both arbitrary and circular, as it entails a variable ranking of alleged well-formedness principles, if we want to explain, for example, why TR clusters may be either tautosyllabic or heterosyllabic depending on the language. I argue instead for a representational alternative assuming that (i) consonants and vowels are universally segregated, and (ii) involve two parallel CVCV sequences -one on the C-plane, the other on the V-plane -(iii) which may differ in length. It is shown how the major sonority categories, and thereby the phonotactic constraints based on these categories, naturally result from how the two CVCV sequences are synchronized if the one on the C-plane is longer than the one on the V-plane. It will also be seen how the proposed structures naturally account for several processes such as liquid metathesis and deletion, vowel epenthesis, plosive fricativization, etc., while providing a means for measuring the relative likelihood of some of them on the basis of representational markedness. IntroductionConsonant clusters pose an interesting challenge for phonological analysis and theory. This concerns the way languages diverge as to the amount of phonotactic constraints they are subject to, the empirical generalizations that can be drawn from cross-linguistic diversity, and how these generalizations are handled by theoretical models. In what follows, I will first show how two basic types of clusters may be distinguished (Section 2.1), and how both types have been accounted for by representational approaches (Section 2.2).1 It will be seen that past work on the subject has moved along different lines which are all based on the notion of "licensing". It will be argued that these accounts suffer from at least one major drawback for a representational theory, as they need to turn supposedly well-formedness principles into hierarchically ordered constraints. I will then defend (Section 3) and develop (Section 4) a radically new approach to consonant phonotactics. This proposal is based on the assumption that planar segregation between consonants and vowels is a universal feature of phonological representations, and that the consonant and vowel planes may have different lengths. The major phonotactic constraints, commonly seen as involving features like [consonantal], [sonorant] or [continuant], depend on how these planes are synchronized, "sonority" being something which results from the structure, not the other way round. This theory will be shown to provide a straightforward account of a dozen well documented facts, from several generalizations on the combinatorial properties of onsets and coda+onset sequences, to such processes as liquid metathesis and deletion, vowel epenthesis or compensatory lengtheni...

According to current unarist theories of vowel systems, minimal components (like A, I, U, etc.) are structured in relation to one another either in terms of dependency (or government) relationships (dependency and government phonologies), or by their relative weight (particle phonology). No attempt has been made in order to evaluate both formalisms, which are sometimes used simultaneously to represent vowel height. The present paper claims that both government and weight are required in the representation of vowel structure, insofar as each notion is specifically motivated. It is argued that government and weight express respectively the 'functional' and the 'structural' aspects of vowel systems. The main point is that such a distinction allows for a strict 'principles and parameters' approach to vowel structure and processes. This theory, which is formally an autosegmental development of particle phonology, provides, as is fully shown by Portuguese data, a unitary and natural account of complex processes by substituting one universal principle for several descriptive rules.* I am grateful to Marc Klein (Paris VIII) and Marc Plinat (Paris X) for helpful comments and suggestions on earlier versions of the present paper. I am also indebted to Mrs. Helen Dochney-Lepine for the revision of the manuscript.

No abstract

It is argued here that most phonological theory is not theoretical, but based on primitives and axioms (the so-called`constraints') which derive directly from the data they are supposed to explain. This article attempts to show what a non-circular conception of phonological theory may look like. The number of segmental primes, their markedness value, phonological content, and combinatory properties, as well as currently assumed constraints on syllable structure are shown to follow from a Boolean algebra, and, thus, to be independently motivated theory-grounded theorems. Hence, for example, neither the onset nor the no-coda constraints posited by OT are required. Another issue of the present theory is that segmental content and syllable structure and interdependent aspects, which emerge from the determination of skeletal units.* I wish to thank Marc Klein, Jean-Elie Boltanski, Tobias Scheer, Philippe Se Âge Âral and an anonymous reviewer for Studia linguistica for helpful comments on earlier drafts of this paper. I am also indebted to Helen Dochney-Le Âpine for crucial help in the preparation of the present version. 228 Joaquim Branda Äo de Carvalho # The Editorial Board of Studia Linguistica 2002.