No abstract

No abstract

This chapter characterises the empirical and theoretical bounds of the concept of allomorphy. We demonstrate how allomorphy may be distinguished from phonological, morphological and syntactic sources of variation, and we attempt to highlight the core concerns that a full theory of allomorphy should address. These include the types of elements (for instance, roots versus affixes) that may enter into allomorphic relationships, how far conditioners of allomorphy may be from loci of allomorphy, when and how competition between allomorphs is resolved, and whether grammars impose upper bounds on numbers of allomorphs.

Eulàlia Bonet the Person-Case constraint (PCC). This constraint, claimed there to be universal, is present in languages that have pronominal clitics, like the Romance languages, languages with weakened pronouns, like English, and languages that have a rich agreement system, like Southern Tiwa. The constraint, thus, affects complexes of ϕ-features related to the argumental structure of the verb. The most common context for the Person-Case Constraint is ditransitive clauses, even though other constructions that can trigger it are causative constructions, and constructions with datives of inalienable possession, for instance. In (1) the effects of the Person-Case Constraint are shown with a first person clitic corresponding to the Direct Object, but ungrammaticality would also arise with a second person clitic (singular or plural). Combinations of two third person clitics do not usually lead to ungrammaticality, even though they often trigger changes not relevant here. The judgements on combinations of first and second person clitics, illustrated in (4), seem to vary a lot. In some languages, these combinations are ungrammatical, while in others, like Catalan, they are grammatical for some speakers, and plainly ungrammatical for others. An additional set of speakers of Catalan accept them in only one of the possible readings, but the judgements as to which one is preferrable seem to vary from speaker to speaker. 3,4 (4) (*) Te m' ha recomanat la Mireia 2Sg 1Sg has recommended the Mireia a. 'Mireia has recommended me to you' b. 'Mireia has recommended you to me' This difference in behavior led Bonet (1991) to posit a strong version of the constraint, for speakers who do not accept sentences like (4), and a weak version of it, for speakers who do accept such combinations. These two versions were stated as follows ((5) corresponds to Bonet 1991: 182, (11)). (5) *Me lui / I-II Constraint a. STRONG VERSION: the direct object has to be third person b. WEAK VERSION: if there is a third person it has to be the direct object

(to appear in Probus)In Catalan vowel epenthesis and consonant deletion seem to have a different conditioning in simple words, in verb-clitic or clitic-verb sequences, and in clitic-clitic sequences (where an emergence of the unmarked effect with respect to syllable structure is found). In this paper, it is argued that, in spite of these domain effects, which would suggest the need for a serial analysis, all the facts concerning epenthesis and consonant deletion can be accounted for in a parallel optimality-theoretic approach. The differences in behavior are the consequence of the different ranking of morphological Alignment constraints with respect to other constraints and an Alignment constraint that makes reference to subsyllabic constituents.