Friday April 13th 15:30 Greenberg Room
Augustin Speyer
University of Pennsylvania
Was Old English a Verb Second Language?
For two decades, generative historical syntacticians, relying on the work of van Kemenade (1987) and of Pintzuk (1991), have largely accepted the claim that Old English, like its Germanic cousins, obeys the verb-second constraint. This idea is used to account for the inversion of the tensed verb with the subject in sentences with topicalized objects like the following:
(1) Thaet hus haefdon Romane to thaem anum tacne
geworht.
that house have the Romans with the one
feature constructed
However, as was known to van Kemenade and to Pintzuk, sentences with pronominal subjects fail to exhibit such inversion:
(2) Aelc yfel he maeg don.
every evil he may do
This latter fact led Pintzuk to propose that, in Old English, unlike in German or Mainland Scandinavian, the tensed verb raises only to I (NFL) rather than all the way to C(OMP) and to formulate a mechanism that derives the difference in word order between (1) and (2) from an interaction of verb movement with clitic pronoun scrambling. Pintzuk's approach, however, has problems. For one thing, as van Kemenade noted, it wrongly predicts that Old English should allow V2 word order in subordinate clauses. Still, more recent accounts of Old English word order have not yielded an unproblematic explanation for the contrast between (1) and (2).
The analytical problem posed by the observed variation in Old English V2 word order depends, however, on the empirical assumption that the order in (1) is obligatory. We will give a statistical analysis of word order patterns in the now available York-Toronto parsed corpus of Old English which indicates that this empirical assumption is false. Although the order in (1) is favored, it appears not to be the only one possible. If we are correct, it follows that Old English is not a verb-second language in the ordinarily accepted sense and the inversion of subject and verb in (1) will require a new explanation. We argue that the patterns of verb-subject order in Old English arise from variation in the position of subjects, as discussed in Haeberli (2000). This variation is grammatically free with non-pronominal subjects and the tendency of the language to favor inversion of subject and verb in cases like (1) is due to an interaction, in usage, between prosody and syntax that is governed by considerations of information structure.