Distributed Morphology Today: Morphemes for Morris Halle

1. Variability and Allomorphy in the Morphosyntax of Catalan Past Perfective

1 1.1 Introduction The past perfective in Catalan illustrates a case of Labovian variability (Labov 1969 and related work) in that it shows up to three different forms (1): a synthetic form (S), spoken in some varieties of Valencian, Rossellonese, and Balearic Catalan (Majorcan and Ibizan) (1a); and two analytic forms, the standard (A1), which contains an inflected form historically derived from the present tense of the verb go and the infinitive of the corresponding verb (1b); and a nonstandard variant (A2) heard all over the Catalan-speaking area, whose first element resembles the synthetic past (1c). These forms do not express different lexical or truth-conditional semantics, nor do they show different morphosyntactic functions, and individual speakers use some subset of them without distinction. (1) a. b. c. Synthetic (S) Standard Analytic (A1) Nonstandard Analytic (A2) purificares vas purificar vares purificar purify.2sg.pst.perf stnd.aux.2sg purify nonstnd.aux.2sg purify ‘you purified’ ‘you purified’ ‘you purified’ As shown in (2), these data show special morphological features. On the one hand, the nonstandard analytic form A2 is especially interesting in that, at least descriptively, it seems to be built on the standard analytic form A1 with incorporation of the past perfective morpheme –re– found in the synthetic form S, and thus being apparently marked twice for past perfective. On the other hand, 1/2pl suppletive forms anem/aneu of the lexical verb anar ‘go’ contrast with the absence of suppletion in both auxiliary paradigms. Variability and Allomorphy in the Morphosyntax of Catalan Past Perfective Isabel Oltra-Massuet 2 Chapter 1 (2) Present indicative main verb go versus (non)standard past perfective auxiliary go 1sg 2sg 3sg 1pl 2pl 3pl PrInd lexical anar ‘go’ vaig vas va anem aneu van StndPstPerf Aux (A1) vaig vas va vam vau van +Inf NonStndPstPerf Aux (A2) vàreig vares va vàrem vàreu varen +Inf These data give rise to a number of empirical as well as theoretical questions regarding (i) the synchronic (partial) syncretism between the auxiliary of the A1/A2 forms and the present tense indicative of the irregular lexical verb anar ‘go’ in (2), which in turn brings up the question of the composition of meaning in the analytic past perfective, as well as the issue of the status, representation , and morphophonological realization of roots that seem to compete for insertion into f-morphemes; (ii) the contrast between the allomorphic variation in the 1/2pl present indicative and the paradigm leveling in the past auxiliary; and (iii) the kind—and locus—of structural microvariation that can derive both an analytic and a synthetic phonological realization from the same syntacticosemantic featural content within the grammar of an individual speaker. Adopting standard syntactic structures and following basic assumptions about the structure of the Catalan verb, I develop a Distributed Morphology account of the above-stated issues that is mainly based on a single functional morpheme motion with two phonological exponents, and the interaction between the internal syntax of these forms and the late insertion of underspecified functional Vocabulary items. I further show that intraspeaker variation is subject to morphosyntactic restrictions (synthetic S vs. analytic A1/2), whereas optionality is tied to the probabilistic application of an extralinguistically conditioned Impoverishment rule (standard A1 vs. nonstandard A2). 1.1.1 Catalan Main Verb Go vs. Perfective Auxiliary Go Different works have dealt with past perfective in Catalan, such as Badia i Margarit 1951, Colón 1976, Vallduví 1988, Pérez Saldanya 1996, 1998, Comajoan and Pérez Saldanya 2005, Juge 2006, or Jacobs 2011, among others. They are all diachronic approaches mainly hypothesizing about the origin, factors involved, and grammaticalization path from a lexical verb go to an allegedly desemanticized past auxiliary,1 which contrasts with other Romance languages —and most languages in fact, where the verb go has been grammatical- Variability and Allomorphy in the Morphosyntax of Catalan Past Perfective 3 ized as a future auxiliary. However, I am not aware of any synchronic analysis of the morphology of these forms beyond the classification of va- /ba/ as either a past (defective) auxiliary, as in Mascaró 1986, 2002b,Vallduví 1988, Wheeler 1998, Pérez Saldanya 1998, or Gavarró and Laca 2002, or as forms of the present indicative of the verb anar with analogic regulation for 1/2nd plural forms, as suggested in Badia i Margarit 1994.2 Treating the auxiliary of the analytic past perfective A1...