The morphological trigger of V-to-T: The case of Old English

1.Introduction The phenomenon of V-to-T movement, which is considered to oppose Romance languages on the one hand and a large majority of Germanic languages on the other hand, has been the target of analysis of the so-called Rich Agreement Hypothesis or RAH, which contends that the movement of the finite verb to the T(ense) head is caused by rich agreement morphology, that is, rich person and/or number morphology, combined or not with rich tense morphology: see e.g. Roberts (1985,1993), Platzack & Holmberg (1989), Rohrbacher (1994, 1999), or Vikner (1997), where it is stated specifically that "V-to-T movement applies if and only if person morphology is found in all tenses" (p. 201). The simplified labelled-bracketing configuration in (1a) shows in an informal way a sequence where no V-to-T applies, just the movement of V to the v head, and (1b) corresponds to a sequence where the movement of V to v is followed by V-to-T movement. (1) (1)[TEXT NOT REPRODUCIBLE IN ASCII] The RAH has not thus far nevertheless appeared to make precise the proper formulation of richness relative to V-to-T, and it has come to be criticised both on the grounds that there are languages with rich verbal morphology that appear to be V-in situ languages and on the grounds that there are languages with scarce verbal morphology that exhibit V-to-T movement. It must also be noted that accounting for the trigger of V-to-T movement is further complicated by the lack of consensus on the use of diagnostic tests in order to acknowledge a given language as V-to-T or V-in situ, as can be easily inferred among others from the works of Thrainsson (2010) or Koeneman & Zeijlstra (2014): such diagnostic tests are typically the ones postulated mainly since pollock (1989), namely, the placement of negation or also of medial adverbs like always or often. Thus, the order V-Neg or V-Adv would arguably indicate that V-to-T movement has applied, whereas Neg-V or Adv-V would indicate just the opposite. However, the syntactic status of negation as e.g. a phrase or a clitic, or the variable position of medial adverbs, can arguably mask the position of V itself. Also, the V2 phenomenon, that is, the movement of V up into the C(omplementiser) position above TP can prove to be an obstacle for the actual verification of the movement of V-to-T. In this paper I argue that the trigger of V-to-T lies in rich tense morphology in the sense of variations or asymmetries across teases, which happen to be provoked by the so-called stem or thematic vowel. The proposal consists in that the stem or thematic vowel segment gives rise to variations or asymmetries across tenses in certain languages, which makes the processing or derivation of verbal forms more complex or longer: such languages are V-to-T languages. In minimalist terms, the stem or thematic vowel morpheme or segment is made to correspond in the present approach with a v feature that T must value in the course of the derivation at narrow syntax. I defend the hypothesis that Germanic languages are all in their old periods V-to-T languages, in a parallel fashion to Romance languages, though later on, in contrast to Romance, they become non-V-to-T. The exceptions to the latter are: Icelandic, Yiddish and also the Faroese language in its original version, since these three Germanic languages have arguably gone on being V-to-T. (2) Because of the relevant role that must be conceded, in my view, to the historical perspective, my main focus in this paper is on Old English (OE), and I first use Romance as a testing ground for the account of V-to-T that I would like to propose. The focus of the discussion is specifically on explaining V-to-T in OE in a parallel fashion to V-to-T in a Romance language like Spanish. In order to support the present account of V-to-T I also use Icelandic on the one hand and German on the other, since these languages are very similar to each other as regards rich agreement morphology, and also rich tense morphology (in the sense of number of markers as compared to number of persons), though they have each received a different treatment in the literature. …