Backward control in Ancient Greek and Latin participial adjuncts

Abstract

This paper discusses evidence for backward control in Greek and Latin participial adjuncts. The data differs from all previously reported cases of backward control in not involving multiple case assignment and from all previous literature except Haddad (2011) in discussing adjunct control rather than complement control. The paper first briefly surveys backward control and how it has been accounted for in the literature. Section 2 brings theoretical background and looks closer at the mechanisms that are typically used in theoretical analysis of control and raising, namely movement/identity and coindexation. Section 3 describes some basic facts about Greek and Latin participles and gives some background on word order, which is crucial to the analysis. Since there is no case assignment in the lower (adjunct) clause, my analysis of the controller’s position relies crucially on word order facts. Section 4 offers a corpus study corroborating the principle of clause-bound discontinuity: word order is not so free as to allow scrambling across clauses, including adjunct participle clauses. Section 5 provides an analysis of the Greek and Latin data within Lexical-Functional Grammar and discusses the distribution of backward and forward control. Section 6 gives conclusions.