This book describes the historical emergence and spread of the to-infinitive in English. It shows that to + infinitive emerged from a reanalysis of the preposition to plus a deverbal nominalization, which spread first to purpose clauses, then to other non-finite environments. The book challenges the traditional reasoning that infinitives must have been nouns in Old English because they inflected for dative case and can follow prepositions. In fact, as early as Old English, the to-infinitive was established in most of the environments in which it is found today, and its syntactic behaviour clearly shows that it is already a clause rather than a phrase at this early date. Its spread was largely due to competition with finite subjunctive that-clauses, which it gradually replaced. Later chapters consider Middle English developments. The book provides a measured evaluation of the evidence that the infinitive marker to undergoes a period of degrammaticalization. It concludes that the extent to which to gains syntactic freedom in Middle English is due to the fact that speakers began to equate it with the modal verbs, and therefore to treat it syntactically as a modal verb. The rise of to-infinitival Exceptional Case-Marking constructions is a Middle English innovation, triggered by changes in information structure that were in turn caused by the loss of verb-second.

English syntax used to have a version of the verb-second rule, by which the finite verb moves to second position in main clauses. This rule was lost in Middle English, and this article argues that its loss had serious consequences for the information structure of the clause. In the new, rigid subject-verb-object syntax, the function of preposed constituents changed, and the function of encoding ‘old’ or ‘given’ information in a pragmatically neutral way was increasingly reserved for subjects. Pressure from information structure to repair this situation subsequently led to the rise of new passive constructions in order to satisfy the need for more subjects; the change in the informational status of preposed constituents triggered the rise of clefts. If information structure can be compromised by syntactic change in this way, this suggests that it represents a separate linguistic level outside the syntax.

This chapter discusses the consequences of the loss of the verb-second rule in Old English and Middle English for the organization of information in the clause. The hypotheses about the consequences of the loss of verb-second are inspired by crosslinguistic/psycholinguistic differences between Present-Day English (PDE) on the one hand and Modern Dutch and German on the other hand, but also involve diachronic work, some signposting is in order. The chapter is organized as follows. Section 2 focuses on the differences between the first position in PDE, where only the subject is available to link to the preceding discourse in an unmarked way, and the first position in a modern verb-second language. It argues that the development of certain types of it-clefts and passives in early Modern English is a response to the loss of verb-second syntax in terms of functions of the first position. Section 3 zooms in on the local anchoring function of verb-second syntax, against the global anchoring system of PDE, which obviates the need for links to the immediately preceding context in that language. Section 4 looks at the evidence of Old English as preferring local rather than global anchors. As the infelicity of local anchors in PDE is more a matter of pragmatics and information structure than syntax, this means that the data cannot be expected to show an all-or-nothing situation from one period to the next, although the changes in preferences by users should be visible nonetheless.

In Old English we find considerable variation in the infinitival complementation patterns of verbs: some verbs take a bare infinitival complement, some verbs take ato-infinitival complement and some verbs can be complemented by either infinitive. This article evaluates two previous attempts to account for the distribution of the two infinitives and offers an alternative account; it also argues that the increase in the use of theto-infinitive occurred mostly at the expense of the finite complement clause rather than at the expense of the bare infinitive, which is the traditional view.