2.2 The verbs with present in -ō

Transcription

1 Proto-Indo-European conjugation conventional label of Indo-European perfect, regardless of the aspectual or temporal functions assumed by this morphological category through the historical development of the various Indo- European languages. The real function it performed is therefore one of the issues still open for scientific debate and there is not yet any uniformity of views on this matter. 1 It seems clear, however, that the solution to the problem must take account of all the functional, semantic, pragmatic and syntactic aspects related to each verb category where the original morphological features of this archetypical flexional pattern are reflected. With these preliminary considerations in mind, we shall right away resume this issue, bringing a new piece into the investigative framework. 2.2 The verbs with present in -ō The conventional description of the verbal system of the attested Indo-European languages draws a fundamental morphological distinction between verbs forming the present indicative with the ending -mi and verbs, which, in the same function, add the ending -ō. The former ones use to be called 1 Cf. the following comment of LURAGHI (2012: 29), footnote 21: «Verbs in this inflectional class have been connected with the PIE perfect and/or middle (ROSE 2006). Synchronically, verbs in the -hi conjugation cannot be considered to be belonging to a semantically coherent and well defined class; they have both active and middle forms, and, apart from morphological aspects, nothing keeps them distinct form verbs in the other inflectional class, the -mi conjugation. However, if they really go back to the PIE middle, their separation from other media tantum in Hittite (and Anatolian in general) should be accounted for in any assessment of the PIE middle voice. Unfortunately, there is no consensus regarding the way in which these verbs came to constitute a separate group». 15

2 Chapter 2 athematic and the latter thematic verbs. Such descriptive model is most of all employed in the study of Greek grammar, and is applied to Latin as well. It shows indeed a certain practical value also for the other Indo-European languages as a whole. The Indo- Iranian languages too, which show generalization of the ending - mi, are usually included in this bipartite scheme, assuming that -mi was analogically extended to the forms that originally belonged to the ō-conjugation. In this case, anyway, the analogical ending would not have replaced the original one, but would just have been added to it; as shown below: *b h érō I bear Gr. phérō OI bhárā + -mi scheme 7 Such reconstruction is fully plausible. As for its relative chronology, it is to be considered likely that the analogical addition of the morpheme -mi first took place after the merger of the three originally long vowels *ā, *ē, *ō into Indo-Iranian ā. It would have been the loss of distinctivity caused by this phonological change, that would have promoted redetermination of the morpheme denoting the first person singular of present tense. Avestan actually fully confirms such development, as this language still made use of the first-person singular ending -ā (<*-ō), only for the thematic verbs, side by side with the analogical one -ā-mi. 2 As for Hittite, we stand once again before an apparently different situation from that envisaged for the other Indo-European languages. Here, even the vestiges of the ō-present seem to be lacking. As we have observed, Hittite actually shows, beside the present in -mi, that in -hi. But could it not be that the same original dichotomy may hide behind that? 2 See WILLIAMS JACKSON (1892:131) and SKJÆRVØ (2003:71). 16

3 2.3 Two basic personal conjugations Leaving Hittite aside again, let us at now look how, in Greek, Latin and Germanic (Gothic), an action at the present temporal level can be expressed by three different morphemes of first person singular: Greek Latin Gothic 1. *b h erō > phérō ferō baira I bear 2. *esmi > eimí sum im I am 3. *me-mona > mémona meminī (ga)-man I have in mind scheme 8 From a taxonomic point of view, the endings 1. and 2. are traditionally considered as allomorphs of the present indicative, while 3. is considered a resultative perfect with present value. It must be recalled here that the Latin ending -ī has to be diachronically analyzed as *-a-i, and thus referred to the same original form which gave rise to the Greek and Germanic ones, with the addition of the same morpheme *-i that distinguishes the so called primary endings of the present in -m-i. If we now transpose the three reconstructed forms into the phonological shape which they probably had before the vocalization of the Proto-Indo-European laryngeals, then we get: 1. *b h eroh 2 2. *h 1 esmi 3. *(me-)monh 2 e scheme 9 17

4 Chapter 2 At a merely formal level, we can sketch a situation where two different basic morphological behaviors can roughly be distinguished, in respect of the first-person singular (present) ending: a) personal ending -mb) personal ending -h 2 - scheme 10 In Hittite, where the original laryngeals are partially preserved, the original dichotomy is still fully visible in the twofold conjugation of present in -mi / -hi. In a way similar to Latin, Hittite too had determined the ending in laryngeal by the morpheme -i of the so called primary endings, whose original function is probably deictic, both spatially and temporally, giving action the semantic connotation of hic et nunc. The fact that this morpheme is absent in the corresponding forms of Greek and Germanic, reveals that it originally was a clitic element, which appeared only to the extent that the pragmatic, syntactic and semantic rules of the language required it. In other words it was not part of the ending to which it possibly was added. In the case we are examining, the truly original personal endings are consequently represented by -m and -h 2. It can also be deduced that, in a similar way to suffixes, personal endings appeared with different apophonic degrees, in accordance with certain grammatical rules, in interaction with the apophonic structure of the verb root itself: *b h ér-oh 2 I bear vs. *u ói d-h 2 e I know (Gr. oîda, OI veda, Goth. wait) scheme 11 At least formally, we can consider the two forms exemplified above as belonging to two different subgroups inside the major 18

5 class of the verbs in -h 2, which, in turn, is in opposition to that of the verbs with first-person ending in -m. 2.4 The original function of the twofold verb conjugation Now, since all the analyzed different kinds of ending can occur in verb forms that are compatible with the temporal category of present, we can deduce that such endings are not in temporal opposition to each other. Thus, their original main function is not marking verbal tense. They certainly express verbal person, though referring to it by two thoroughly different morphological sets: 1. pers. -m / -h 2 2. pers. -s / -th 2 3. pers. -t / -(e) scheme 12 It is most likely that such a striking morphological opposition originally correlated with a functional opposition, and that the latter concerned in particular with the relationship between verb and subject. Hence, it is the diathesis that lies at the basis of the dichotomy of the Indo-European verb person. 19