Ans van Kemenade (Radboud Universiteit)

Discourse particle then in English : clause structure and discourse organisation
Ans van Kemenade, Radboud University, Centre for Language Studies

The paper focuses on then as a discourse particle in the history of English. It will be shown, on the basis of the parsed corpora for Old English, Middle English and early Modern English, that the information structural status of then as a discourse particle was more or less stable throughout the history of English, even though its syntax changed profoundly. We will focus on the use of then in questions and imperatives, as these two clause types are characteristic of interactive language and occur robustly in the corpora. The precursor of then in Old English (þonne) is analysed by Links & van Kemenade (submitted) as an adverb grammaticalised to a functional head which occurs high in the left periphery, preceded only by (wh) – finite verb – (discourse-given subject) -. The analysis is much inspired by the massive literature on discourse/modal particles in German and Dutch. There is a second temporal particle þa, which we will here ignore as it is used as an exclamation and an assertion particle, comparable to German ja. Þonne occurs on a large scale in questions, imperatives and conditional correlatives, and is presuppositional in the sense that it reflects a speaker response to the context, such as surprise or disapproval in questions, reinforcement or downtoning of the directive in imperatives. The combination of this clause external link and the fact that it marks the boundary between discourse-given and discourse-new participants, gives it a clear discourse-structuring character. Þonne occurs robustly in Old English. Its frequency of use dropped over the Middle English period, but both its syntax and pragmatic function at first remained stable. From the late Middle English period onward, this pragmatic use of then became associated with the right periphery. The clause-final position presumably arose as a competing option in the spoken language, comparable to Dutch (Foolen & vd Wouden 2015). It was the clause-internal position that was lost, and it will be argued that this is the result of an interesting complex of syntactic and pragmatic factors : 1) The loss of V2 from late Middle English onward, hand in hand with the loss of V to I movement and the auxiliation of the modals and the rise of do-support, rigidly reduced the scope for discourse structuring elements in the higher clausal domain. 2) the earliest attestations of final then occur in late Middle English, according to Haselow (2012) as resumptive adverbs in the apodosis of conditional clauses. 3) This use of then was grammaticalised as a clause-final particle, implying paratactic reference to the context, and was subsequently broadened to a variety of clause types.