Past Event

The Rise and Fall of the Relative with a Leftward Island

Dr. Robert Truswell
University of Ottawa

About the Presentation

Over the course of its history, the relative clause system of English has undergone several changes. The major details of these changes, and their relation to other aspects of English grammar, are now well understood, thanks to work by Cynthia Allen, Suzanne Romaine, Ans van Kemenade, and others. However, interesting in a different respect are a series of relative constructions which, to borrow Anne Breitbarth’s phrase, “live fast, die young”. These are constructions which appear and disappear over a century or two, are sometimes quite infrequent even in the period in which they are found, and often seem tangential to the broader narrative which seeks to explain how we move incrementally from the Old English relativization system to the very different system of Present-Day English.

Such apparently peripheral constructions pose interesting questions of their own. As with any diachronic study, we want to know where they came from and where they went to, but these questions take on extra urgency as a result of the evanescence of the constructions in question. How can it be that a whole construction can appear, flourish to a very limited extent, and disappear so quickly that there is little concurrent change in the rest of the system?

These questions are particularly uncomfortable for believers in Principles & Parameters: the obvious course of action, for each short-lived construction C, is to stipulate a microparameter of the form “C can/cannot be formed in language L”. However, this approach faces two problems: it goes against the general lexicalist project of localizing variation in properties of lexical items, and it has no obvious way of addressing the question of why learners would suddenly choose a new setting for the parameter, at the very beginning of a change. P&P theorists therefore hope to find some antecedent trigger, related to particular lexical items, to explain why we see the slight changes we do, when we do. In other words, the constructions in question are hopefully less peripheral than originally thought, and may even cast light on less peripheral changes.

This talk examines one such short-lived, “peripheral” construction, namely the Relative with a Leftward Island, or RLI. These are biclausal structures, with one clause left-adjoined to the other, and a dependency between a wh-pronoun and a gap within the left-adjoined clause, as in (1).

(1) Mr Hoby, my Mother, and my selfe, went to visitt some freindes [[who, _ beinge not at home], we retourned]
Lady Margaret Hoby, Diary, 1599-1601

This complex, low-frequency construction appeared around 1500, peaked in the first half of the 17th century, and declined in frequency from that point on, apparently dying out in the mid-19th century. By analysing its syntax and semantics from a synchronic perspective, we arrive at viable hypotheses concerning the diachrony of the construction. The key, as represented in (1), is that the relative pronoun never leaves the left-adjoined clause. Even when the antecedent of the relative pronoun immediately precedes it, then, as in (1), the pronoun is structurally very remote from its antecedent. Given this analysis, I show that RLIs are amenable to the kind of diachronic story which is desirable from a P&P perspective: RLIs emerged when independent factors first made it possible for a learner to posit a relative pronoun and a gap in this configuration, and they disappeared when relative pronouns began to require structurally more local antecedents.