All Germanic languages with the exception of English are subject to the Verb-second-constraint (=V2), as is well known (e.g. Vikner 1995).
English lost V2 during the Middle English period. In other words: examples like (2a) are after a certain time ungrammatical and are replaced by examples like (2b), whereas sentences like (1) are unaffected.
(1)John likes beans.
a. Beans likes John.
b. Beans John likes.
At the same time another development takes place: The rate of topicalization, i.e. the fronting of any constituent other than the subject, declines also:
In the period 1150-1250, 11.7% (575of 4913) of sentences with direct object have the object fronted, in the period 1640-1710, only3.6% (128 of 3541). This latter development has not generally been noticed in the literature.
The question arises whether the loss of V2 and the decline of topicalization are related to each other. In my paper I argue that they indeed are related and that, given the loss of V2 inEnglish, the decline of topicalization is a consequence of a prosodic constraint common to at least all Germanic languages.
This constraint, which we might call ‘Trochaic Requirement’ (=TR), is a reflex of the Obligatory Contour Principle (Goldsmith 1976), which requires that adjacent units differ infeature content, in the domain of prosody: Between two phrasal accents a prosodically weakelement must intervene.
Topicalization involves fronting of a constituent which meets certain pragmatic requirements (see Prince 1999). The topicalized phrase is accented, as is another phrase inside the open proposition (i.e. the remainder of the sentence), viz. the variable of the open proposition. In principle, these two accents could wind up adjacent to one another, which would violate the Trochaic Requirement.
The sentences in (3) illustrate how discourse context determines accentuation.
(3) a. John doesn’t like peas, but beans John likes.
b. John is eating peas today, but beans John ate yesterday.
Now consider example (4).
(4) Mary likes peas. Beans, John likes.
Here the variable of the open proposition is the subject, as is most often the case. Since sentences like the 2nd of (4) violate the TR, they are disfavoured. As long as V2 was still an option, the problem could be resolved by inverting the subject and the verb (= 2a). As V2 becomes impossible, topicalizing leads in many cases to a violation of the TR and isconsequently avoided too.
I came to these results by conducting a quantitative study using Corpus Search on all available parts of the Penn-Helsinki Corpora of Middle and Early Modern English (1150-1710). Only sentences with a full noun phrase subject and a non-auxiliary verb were taken into account (32749 sentences with direct object; 1780 thereof with topicalization, 585 thereof with full-NP-subject). The accent on the open proposition was assigned manually, using an algorithmexperimentally tested for modern English.
Goldsmith, John A. (1976) Autosegmental Phonology. Doctoral dissertation, MIT (published1979, New York/London: Garland).
Prince, Ellen (1999) “How Not to Mark Topics: ‘Topicalization’ in English and Yiddish,” in:Texas Linguistics Forum. Ch. 8. Austin: University of Texas.
Vikner, Sten (1995) Verb Movement and Expletive Subjects in the Germanic Languages,Oxford / New York: Oxford University Press.
"It is no good to try to stop knowledge from going forward. Ignorance is never better than knowledge."
~~Enrico Fermi (1901 - 1954)~~