The Metre of Beowulf: A Constraint-Based Approach (review)

Why are we never done with the structural analysis of Beowulf? The short answer is that no one has as yet fully understood the meter of Beowulf, and the reason could be that the meter—if there is one—can’t be understood before the phonology of Old English (or perhaps phonology in general) is sufficiently well understood. Therefore, the hope remains alive that it will be fruitful to apply new phonological models to the old text. And it is. The knowledge of the structure of Beowulf has been accumulated over several schools of analysis, and the research interest in the topic remains alive. The sense of a real understanding is not quite there, however.

Michael Getty’s monograph provides a constraint-based analysis of meter that brings some of the insights of optimality theory (OT; Prince & Smolensky 1993) to bear on the analysis of Beowulf. In a meter with the incongruous properties of highly variable line length in terms of syllables or moras, highly variable distribution of lexical stress (both pattern and number), and with frequent stress clashes and lapses, it seems reasonable to regard the 3,169 noncorrupt lines of poetry as the outputs of a grammar in which several conflicting constraints are ranked at a level where each gets a say at least once in a while. A loose grammar, if you like.

Many Beowulf scholars will like G’s approach. It is well written, internally consistent, and explicit in motivating the theoretical positions taken. Its greatest virtue is perhaps not the answers it provides, but rather the fact that it restates some of the basic issues, and prompts discussion of several more. What is the relationship between language and meter? Why should some meters be immediately perceptible from the text and others so well concealed? Where do we find a contemporary meter that is like Beowulf, so we could better investigate it? To what extent should the reception of the meter go into the analysis? Greflects on such issues in interesting and thought-provoking ways. My chief concerns with this work have to do with the concrete implementation of G’s general program.

Ch. 1 provides a useful outline of the book and an excellent review of Sieversian approaches to Old English (OE) meter. Issues like the status of Kuhn’s laws and stress in finite verbs are discussed extensively and put in a science–historical perspective.

Ch. 2 deals with lexical and phrasal stress in Old English. G reviews previous approaches and proposes a set of constraints that place primary stress on roots and allow recursive stress feet (a complicating fact that is difficult to evaluate with respect to the subsequent matching to metrical structure). Relevant phrasal constraints are introduced and interesting comparisons are made with Finnish (for word prosody) and Modern English (for phrasal prosody).

Like most previous researchers, G acknowledges a basically rhythmic character of the meter, but ‘chiefly at the levels of abstract phonological representations’ (172). The high-ranked constraints proposed in Ch. 3 ensure that (i) verse feet are uniformly left-headed (trochaic), (ii) verse feet and half-lines branch, (iii) weak positions can be left out or expanded, (iv) the maximum size of a metrical position is a phonological foot, (v) phonologically weak syllables are barred from strong metrical positions, (vi) strong syllables of heavy-stemmed, polysyllabic words are barred from weak positions, (vii) no text is extrametrical, and (viii) alliterating syllables occur in the head position of verse feet.

Ch. 4 introduces the lower-ranked constraints, which allow for the vast rhythmic variation encountered in Beowulf (*Clash, *Lapse) and which stop the metrical model from overgenerating (so-called Balance constraints). Here we also find a group of meta-constraints including Boundary which makes metrical boundaries maximally distinct, and Hanson and Kiparsky’s (1996) Fit, which sees to it that languages select meters that can use the native vocabulary...