‘Tense’ Marked Obliques in Malagasy

In Malagasy, spatial deictic adverbials and certain PPs may carry the prefix t-. This prefix is usually analyzed as a past tense marker, since oblique predicates take twhen denoting past states, while oblique modifiers typically require twhen the verb is marked for past tense. It is generally assumed that ‘tense matching’ is mandatory. Here, however, I present data showing that past tense verbs may take non-t-marked obliques (mismatching). When the oblique denotes a goal, mismatching generally indicates that the theme of motion is currently occupying the endpoint of motion, or has not yet reached it. When the oblique denotes an instrument, location, etc., mismatching triggers a past habitual reading. I propose a provisional account of these facts, which treats tas a marker of temporal boundedness and captures the contrast between goal and non-goal obliques in terms of where they merge in the syntactic structure. In this paper I discuss the distribution of the prefix tin Malagasy, a language of the Philippine type spoken on Madagascar (here I focus on the standard dialect). Tattaches to spatial deictics and certain types of PPs and adverbials, and is generally analyzed as a past tense marker. This is illustrated in (1), where the deictic phrase constitutes the main predicate of the sentence (note the absence of a copula). Sentence (1c) shows that when the deictic takes t-, the clause receives a past tense interpretation, while (1a,b) show that tabsent when the clause has a present tense or irrealis/future interpretation (future/irrealis being marked by the particle ho): (1) a. Any anati-n’ ny ala ny gidro there inside-Lnk Det forest Det lemur “The lemur is there in the forest” b. Ho any anati-n’ ny ala ny gidro Irr there inside-Lnk Det forest Det lemur “The lemur will be there in the forest” * This paper supersedes my earlier work on this topic, which appeared as Pearson (2000, 2001). These articles were based on work with a single native speaker in Los Angeles. Subsequent fieldwork with multiple speakers in Madagascar (Antananarivo and Nosy Be) showed that the first speaker’s judgements were rather idiosyncratic, necessitating a reassessment of the phenomenon. Many thanks to the following individuals for providing the data for this paper: Raherimandimby Rija, Ranaivoson Elia, Randria Aina, Randriamihamina Mihaingosoa Hasiniaina (“Hasina”), Rasanimanana Lova, Razafindrakoto Laza, Razanajatovo Rado, and Razanarisoa Clarisse. Thanks also to audiences at AFLA 12 (the Austronesian Formal Linguistics Association) and 10 ICAL (International Conference on Austronesian Linguistics) for comments on earlier versions of this work. All errors and oversights are of course my own. 1 Malagasy is a verb-initial language with a relatively fixed word order (traditionally characterised as VOS). Clauses generally consist of a predicate phrase followed by a clause-final trigger (usually called the subject, but analyzed in Pearson 2005 as an A’ element). As in other Philippine-type languages, the grammatical function of the topic is indicated by the voice form of the verb. In this paper, two voice forms appear in the examples, the actor-topic (or ‘active’) form, used when the external argument of the clause acts as trigger; and the theme-topic (or ‘passive’) form, used when an internal argument is the trigger. For more on the Malagasy voicing system, as well as general information on the morphosyntax and word order of the language, see Keenan (1976), Guilfoyle et al. (1992), Pearson and Paul (1996), Paul (1998), Pearson (2005), Rasoloson and Rubino (2005), and references cited therein. 2 The following abbreviations are used in the examples: 1s = 1 singular pronoun/enclitic, 12 = 1 inclusive pronoun/enclitic, 3 = 3 (singular/plural) pronoun/enclitic, AT = actor-topic, Det = determiner, Irr = irrealis, Lnk = linker, Loc = locative proclitic, Pst = past, TT = theme-topic. The tprefix is glossed simply as “T” throughout.