The aim of this article is to present a unified analysis of Romance past participle agreement. Data from 19 standard and non-standard Romance varieties are considered. It is shown that in order to account for the variation found across the various Romance varieties, past participle agreement should not be considered as following simply from another module of the grammar, such as Case Theory, coindexation or the theory of functional categories, but should rather be considered from the point of view of a general theory of grammatical agreement. It is also shown that past participle agreement must be treated as an example of nominal-modifier agreement rather than as an example of verb-argument agreement.

In this article, it is argued that perfective aspect in Serbo-Croatian is licensed in the syntax. Licensing of Aspect requires the presence of a hierarchically higher category. In compound tenses, perfective aspect is licensed by a governing auxiliary in Infl while, in the present tense forms, it may only be licensed by an overt complementizer in Comp. A Comp-Aspect Agreement mechanism is proposed based on the requirement for licensing in the present tense forms. This agreement also obtains between a covert complementizer and perfective aspect in imperative constructions (perfective imperatives move to an empty Comp position to satisfy the Aspect licensing requirement) and between an overt complementizer and the auxiliaries in compound tenses (auxiliaries move to an empty Comp to satisfy the Tense licensing requirement). Although Tense may be licensed by any governor in root sentences, the presence of an overt complementizer, which is a hierarchically higher category, takes precedence for licensing in embedded sentences.

The introduction of empty nuclear positions in syllable structure goes back to the early 80s. Since then, empty nuclei have been proposed in phonological analyses to account for vocalic epenthesis (vowel/Ø alternation). The analysis of Gen put forth in this article argues for empty nuclei although no vowel/Ø alternation can be observed in the language. Nevertheless, on the basis of the tonal patterns of Gen, it is shown that empty positions must be part of the underlying representation of syllables in the language. Another aspect of the study deals with consonant clusters. Thus far, all analyses supporting empty positions within a cluster have also referred to the absence of cooccurrence restrictions among the consonants surrounding the empty vowel. This paper provides new insights since, in Gen, the consonant clusters in which empty nuclei are proposed are typical of those observed in branching onset languages (like French, Italian and English) or in languages having light diphthongs. We argue that all attested consonant clusters in Gen are heterosyllabic.