Aspectual suppletion and paradigm defectiveness in the Proto-Indo-European verbal system

Abstract

The existence of suppletion in the Proto-Indo-European language is still a question of debate (García Ramón 2002). While the evidence for such a phenomenon has been widely recognized within the verbal system of most Indo-European languages, some scholars describe it as a recent monoglot development which characterizes the history of each single language without involving a previous common stage (Van der Laar 2000). According to Strunk (1977), the hypothesis of a PIE suppletive paradigm based on the alternation of basic verbal root pairs such as *es-: *bhū- “be”, or *ei/i-: *gwā/ gwem- “go”, must be ruled out because it violates what he calls b-criterium, i.e. the complementary distribution of the forms involved in a suppletive relationship. More specifically, the existence of the present tense of Homeric Greek φύω and Sanskrit bhávati on the one hand, and Homeric Greek βαίνω and Sanskrit jgāti on the other, derived respectively from the PIE roots *bhū- “become” and *gwā-/gwem- “come” marked as [+telic], would testify the existence of full paradigms of verbs which already in Old Greek and Vedic Sanskrit were regularly inflected according to each tense. Therefore, this would indicate that those verbs were not “defective” as regards to one of the temporal or aspectual stems (here, the infectum stem marked as [–telic]). However, Vedic and Greek verbal paradigms show a striking lack of proportion in the distribution of stem forms. As I will try to show, such distribution is not random but depends on the [telic] lexical features of the original PIE verbal root. Moreover, the analysis of the contexts in which the verbs occur will confirm morphological data as well. I will test the validity of the criteria proposed by Strunk, considering defectiveness as a gradual phenomenon rather than a binary property related to a lexical stem within a paradigm. Furthermore, I will examine the origin of suppletion in a diachronic perspective, focusing on the role of “bridging contexts” in a case study, namely the hendiadic Homeric Greek expression βῆ δ᾽᾿ἰέναι “went” as an intermediate stage towards a suppletive relationship between the two verbs εἴμι “I go” and ἔβην “I went”. This periphrasis turns out to be a lexical strategy to express the [+telic] perfective past tense of the [–telic] verb εἴμι “I go”, which lacks an aorist stem in its paradigm. By means of comparison with data found in Vedic Sanskrit indicating the existence of a suppletive relationship éti “goes”: ágāt “went” (García Ramón 2002; Deshpande 1992) from the same PIE roots, the aim of this paper is to demonstrate that verb suppletion is an inherited Proto-Indo-European phenomenon. It is then hypothesized that suppletion was a “normal”, highly frequent, and not an irregular or even “unnatural” phenomenon (Dressler 1985; see Corbett 2005) within a system where the lexicon was gradually organized towards a more grammaticalized level. In a typological perspective, the phenomenon of suppletion is another piece of evidence in favor of the active-stative hypothesis for the most ancient PIE stage (Lehmann 2002), reflecting a system where the lexical aspect features of the verbal roots are still more important than tense features. Moreover, the aspectual nature of PIE suppletion is consistent with Bybee’s (1985) Semantic Relevance Hierarchy. Within such a typological system, the phenomenon of suppletion, crystallized in the most frequent verbal forms (Veselinova 2006), would reflect the older features (cf. Mel’čuk 2006), being interpreted as a gradual step towards the emerging of inflectional paradigms from lexicon to grammar. References Albright, Adam (2002). The Identification of Bases in Morphological Paradigms. UCLA dissertation. Website: http://web.mit.edu/albright/www/papers/Albrig