AUTHOR: Julia Schlüter TITLE: Rhythmic Grammar SUBTITLE: The influence of rhythm on grammatical variation and changes in English SERIES: Topics in English Linguistics PUBLISHER: Mouton de Gruyter YEAR: 2005
Mark Campana, Department of English, Kobe City University of Foreign Studies
This book presents a corpus-based test of a simple idea, that English favors an alternating pattern of stressed/non-stressed syllables in word and phrase structure. The Principle of Rhythmic Alternation can be discerned in texts dating from the 16th century to the present, and has (the author argues) influenced the development of the language in subtle ways. It begins by examining the distribution of competing forms, e.g. 'worse' and 'worser'. Although 'worse' has always been the suppleted form of 'bad', most other comparatives had an –er ending (e.g. 'richer'), so there was considerable pressure to fill out the paradigm. These two forms competed with each other from late medieval times, but 'worser' persisted longer than it should have in prenominal position before eventually dying out. The reason is that 'worser' contains an extra (weak) syllable, which the rhythmic grammar favors as a buffer between the stressed syllable of the adjective itself and the typically stressed first syllable of the noun it modifies. 'Worser' gave way to 'worse' much sooner in other syntactic environments where the specter of a stress clash did not arise. In other words, the preference for rhythmic alternation tipped the scales in favor of one syntactic variant over another.
This book represents an expansion of ideas originating in previous and ongoing work, in particular the author's dissertation. Chapter one lays out the major issues and offers an outline of the book. Chapter two presents some background information on the Principle of Rhythmic Alternation (PRA), with discussion of some other phonological concepts and their instantiation in natural language. Never is theory far below the surface though, and right from the beginning the reader is reminded of the kind of consequences that lie ahead: How can we ignore the impact of sound-based phenomena on word and sentence-structure, as well as meaning? Sooner or later, any theory must address this question, and the argument leans towards functionalist accounts motivated by general forces of cognition. In this regard, the PRA is considered as a prime candidate.
Chapter three is an outline of the methodology taken, and a description of the texts used for the survey. These are all machine-readable and represent various stages of British English from 1150 to the present. The search for rhythmic patterns entails the exclusion of verse, so only prose forms are considered: fictional and non-fictional, dramatic and non-dramatic, written-to-be-spoken and even spoken (modern English only). In total, the database contains many millions of words, statistically sufficient for detecting the subtle effects of stress-timing on word and sentence structure. To balance the results, special care is taken to accommodate the idiosyncrasies of each text, including the social context in which it was written.
Chapter four is the first of two chapters dealing with the results of the survey, and evaluating them against the major predictions (i.e. given two or more variants--lexical or syntactic--the rhythmically grammatical one will win out). The story of 'worse'/'worser' typifies the situation in noun phrases, along with other pairs of variants: so-called 'a-adjectives' ('ashamed', 'asleep'); the participles 'drunk'/'drunken', 'broke'/'broken', 'struck'/'stricken' 'knit'/'knitted' and 'lit'/'lighted'. The choice between 'a quite' vs. 'quite a' is syntactic (the latter one persists), as are combinations of color terms (e.g. 'red & yellow' vs. 'yellow & red'). In each case, the evidence shows a marked preference for the 'rhythmically correct' alternant in attributive (pre-nominal) position, where stress clashes are likely to occur. The same variants in post-copular position show a more even (and predictable) distribution. In each case, the results of the corpus-based survey are presented graphically through charts and tables and explained in detail.
Chapter five extends the study of rhythmic effects on variation and historical change to items within the verb phrase: multi-word S-adverbs ('not so unexpectedly'), monosyllabic V-adverbs ('scarce/scarcely'), and -ing forms prefixed by an a-affix ('set a-going'). A similar scenario unfolds in the case of 'marked' infinitival (to) complements of the verb 'make' vs. unmarked (bare) ones: if the infinitive verb has initial stress, a potential stress clash ensues. In this environment, 'to' will function as a rhythmic buffer. As before, the evidence gleaned from the survey shows that 'to' was retained much longer than expected in the history of English, before completely disappearing from the active construction ('make him go'). Its retention in the passive ('made to go') has more to do with syntactic stabilizing factors than with rhythm, according to the author. Throughout this chapter, syntactic and prosodic environments are closely controlled, owing to the ease with which stress clashes can be avoided by other means. Semantic factors in particular obscure the PRA as the sole arbiter of distribution. The sheer size of the database reveals some plainly significant trends, which appear to have no reasonable explanation beyond a rhythmic basis: potential stress clashes are avoided when an extra-syllabic variant is available.
Chapter six considers the theoretical models that are necessary to accommodate the observed effects of rhythm on nominal and verbal structures. Generative (''syntactocentric'') theories are rejected, as their modularity does not allow phonological rules, etc. to affect higher levels of representation. Optimality Theory (OT) is given a better chance, where rules with different functions can be ranked more-or-less as needed for any given dialect. Still, it is ill-equipped to handle variation, which both reflects synchronic reality and feeds the process of historical change. Instead, the author opts for a spreading activation model based on language processing. In it, rhythm can be traced to the behavior of neural clusters in their cycle of activation and inhibition (stressed and unstressed, respectively). These are represented by abstract nodes organized into levels of increasing complexity: phonetic, phonemic, syllabic, morphological, and semantic. The deep-seated rhythmic layer consists of just two nodes, i.e. the stressed and unstressed. Information-sharing between nodes at different levels is bidirectional, so that choices made at one can be communicated up or down. Since processing involves stages of both planning and execution, an imminent stress clash can be avoided via feedback and 'feedforward.' Various arguments are given in support of such a model, the most persuasive being that it reflects (what are thought to be) verifiable cognitive properties. Chapter seven is a summary and conclusion.
Following the conclusion, there is an appendix showing the typical stress patterns of nouns, verbs and adjectives--useful information in understanding the potential for stress clashes. The searchable sources on which the survey was based are listed in a primary reference section; the bibliography contains secondary references from other such studies, as well as from the literature of historical linguistics (history of English), neurolinguistics, psycholinguistics, language processing and more. The book is indexed, which helps immensely in recalling the various principles that periodically interact with (sometimes obscure) the PRA.
Generally speaking, I found this book a fascinating read, if only because it draws together so many different strands of research. The PRA is not new, but the emerging field of corpus studies offers a legitimate testing ground for trends that might otherwise go undetected. Great care was taken to eliminate any evidence that might prove inconclusive or obscure the effect of the PRA. Caveats were plentiful, leaving little doubt as to the correctness of interpretation. The graphs and tables were well-presented and explained. This is good science.
At the same time, the argumentation was usually convincing, even when the survey results were inconclusive. The author makes a good case for using written texts as evidence for spoken forms—a hard sell, to say the least. Nor was the author reluctant to offer alternative views of various phenomena. In many cases, potential support of the PRA is compromised by semantic, prescriptivist, or paradigmatic forces, which the author did a good job of factoring out.
The style of writing is clear and well-organized, following the format of a standard doctoral thesis. The topic of each chapter is presented in the context of the big picture, and subsections are outlined in advance, later summarized. There is some cross-referencing between subsections of different chapters, as some issues cannot be resolved all at once. The language is near-perfekt (!) and sometimes colorful, with a good balance of technical terms and idiomatic expressions. Undergraduate students will find this book to difficult to navigate unless they have a solid background in at least one of the areas of research; graduate students have a better chance. For general linguists, the book offers rare insight into other disciplines, i.e. what other researchers are up to, and how to assimilate their results and methodologies with their own work.
There isn't really much to say that's negative about the book: it has a clear agenda and carries it out. The focus on rhythm may not seem terribly exciting, but there are so many different subplots—from literature and prescriptivism to neurology and theory—that anyone with the patience to learn will find something in it. The fact that results are sometimes difficult to discern can be frustrating, as when the impact of the PRA is countenanced by other factors. The effort to tease them apart is slow and meticulous. The discussion of spreading activation models could have been expanded to anticipate counterarguments of, say, generativists—perhaps too handily dismissed. At a deeper level, the synchronic (and possibly diachronic) influence of rhythm on English must find its expression in syllable-timed (and other) languages. This would be a direction for future researchers, though.
ABOUT THE REVIEWER:
ABOUT THE AUTHOR
Mark Campana is a professor of English and Linguistics at Kobe City University of Foreign Studies. He has written on the morphology and syntax of polysynthetic and ergative languages (Amerind, Austronesian) in the generativist tradition. More recently, his interests lie in the field of pragmatics, in particular intonation, discourse markers, and other audible cues to intended meaning.