Temporal Properties in the Ancient Greek infinitive and their manifestation

Christina Sevdali

This paper wishes to contribute to the understanding of finiteness as non-uniform category cross-linguistically. I will discuss the Accusativus Cum Infinitivo (AcI) construction of Ancient Greek (AG) infinitives, which is exemplified as follows:

(1) Orgilous kai traheis umas esesthai nomizo:.

Angry-Acc and rough-Acc you-Acc to-be-Fut think-I.

“I think you will be angry and rough.”

Demosthenes, Philippika II: 34:1

In the above example, the subject of the infinitival clause (umas) is in the accusative Case and it is different from the subject of the main clause. This construction, that is restricted to the so-called ECM cases in English, is very frequent in AG and in fact is used every time there is disjoined reference between the subjects of the main and the infinitival clause. My account for this phenomenon will be based on the temporal properties and the morphology of the AG infinitives. I will argue that contrary to previous assumptions (Binnick, 1991 among others) that claimed that AG infinitives show aspectual distinctions alone, AG infinitives also show temporal distinctions and possibly mood distinctions. The crucial thing however for my analysis and for the account of AcI is Tense and this will be argued on the basis of morphological, empirical, syntactic and semantic evidence. I will argue against previous analyses in the literature that discussed AcI in Latin and AG. These analyses revolved around two major trends to account for AcI: either they claimed that AcI should be treated like ECM, postulating a raising-to-object rule for Latin (Pepicello, 1977) or they proposed the existence of a null C that is responsible for the Case of the infinitival subject from within the infinitival clause (Cecchetto & Oniga, 2001 for Latin and Tantalou, 2003 for AG). Although I share a lot of the latter view’s assumptions I will not subscribe to them fully. Therefore, I will argue against both these types of analyses following Bolkenstein, 1979 and Pillinger, 1980 against the ECM approach and also using data from AG. In my analysis I will argue that AG infinitives have Tense (following Philippaki-Warburton & Catsimali, 1989) and it is responsible for the accusative of their subject. The infinitival tense will be characterized as relative following Comrie, 1984. Moreover, an articulated theory of anchoring of the infinitival tense through a C position will be put forward following Stowell, 1982 and Enc, 1987. In my system all AG infinitival clauses are CPs and independent evidence for that will be the adjunct clauses of AG that are introduced with prin “before” and -ste “so as” that are arguably CPs. The anchoring system I propose will be based on features of finiteness and tense, following Boscovic, 1997 and Martin, 2001 and the operation Agree as proposed in Chomsky, 2001. The temporality of the AG infinitives will be further supported by data that illustrate that AG infinitives obey sequence of tenses rule and also by data that show infinitives modified with temporal adverbs. The claim that AG infinitives also have Mood will be supported by the double negation system of AG, that is used to mark the distinction between indicative and subjunctive in finite clauses (main and embedded) and it is also preserved in the infinitives. Finally, the ability of the Tense node to be responsible for the Case of the accusative subject of infinitives will be supported by claims on Case assignment by Pesetsky & Torrego (2002), Alexiadou & Anagnostopoulou (2002) and Iatridou (1993). A correlation will be established between Agreement and Nominative Case but it will be shown that Tense cannot be linked to one and only one Case cross-linguistically. I will conclude with certain remarks on the inadequacy of the present system of non-finiteness as I will have demonstrated hopefully that even the use of +/- Agr/T features is not enough to capture the attested cross-linguistic variation.