The paper is intended to shed light on the so-called Kuhn's laws pertaining to early Germanic word order and intonation. This is done by investigating the metrical function of different morphosyntactic categories in Old Icelandic meters and interpreting their behavior on the basis of insights from modern intonational phonology. Verb forms belonging to independent clauses cannot fill the last foot in lines in dróttkvætt poetry. This is taken to show that phonology was involved in the word order restrictions in Old Icelandic, among them the verb-second (V2) constraint, which is often assumed to be syntactically governed.The work reported in this paper started during a sabbatical at Stanford University under a Fulbright fellowship in 1998–1999. I would like to thank Paul Kiparsky, my sponsor there, for inspiration and help with some crucial issues. I would also like to thank the participants in the Metrics conference in Toronto, October 8–9, 1999, and the GLAC–7 conference in Banff, April 21–23, 2001, for comments on presentations of parts of the work. Finally, thanks go to the editor, Mark Louden, and two anonymous reviewers for suggestions for improvement and editorial guidance.

In this paper, I argue that quantitative data on relative clause variation in Old English provide additional evidence against the theory of multiple authorship of Beowulf. I show that aspects of various relative clause types in Beowulf are unique in the Old English poetic and prose corpus and homogenous throughout all three proposed parts of the poem. Comparison with the poems attributed to Cynewulf indicates that the frequency, distribution, and selection of relative clause types are consistent throughout an individual poet's compositions. Statistical analysis of the general distribution of relative clause types and factors such as the type of antecedent and its distance to the relative clause supports the hypothesis that Beowulf is a unified poem.I would like to thank R. D. Fulk, Kari Ellen Gade, and Rex A. Sprouse for comments and critiques of drafts of this paper, and audiences at “Studies in the History of the English Language I” at UCLA and “Philology in Germanic Studies at Illinois and Indiana V” at the University of Illinois for helpful discussion. Special thanks to Julie Auger for advice and her generous help with the VARBRUL analysis. Any remaining errors are of course my own.

Introduction

Traditionally, research on language change has been a post-mortem activity, focused on isolated changes that are complete and often only documented in written texts. In the 1960s the field was advanced considerably by Labovian sociolinguistics and the investigation of “change in progress” adduced through patterns of community-internal linguistic variation correlated with external facts about speakers such as age and class (see Labov 1994 for an overview). However, despite the many benefits of such work on “dynamic synchrony,” we still know relatively little about how language change unfolds over the lifetimes of individual speakers, that is, in real time (cf. Bailey et al. 1991). The logistical challenges of such research are, of course, considerable. Whereas it is straightforward for psycholinguists to observe language development in children over the course of a few years, documenting changes in the verbal behavior of individuals over several decades is by contrast much less feasible. Nevertheless, present theoretical models of language change could be considerably improved by the results of real-time studies.

The notion of SYLLABLE CUT (Silbenschnitt), which was introduced and motivated in classical works by Sievers, Jespersen, Trubetzkoy, and others, has been discussed and developed recently in Germanic linguistics, including by Vennemann (1991), Maas and Tophinke (1991), Becker (1996, 1998), and Murray (2000). Along with the issue of tonal accents, syllable cut is also featured in a volume edited by Auer, Gilles, and Spiekermann (forthcoming).