Abstract
In the wake of the long-standing acknowledgement that cognitive factors play a crucial role in historical linguistic processes (see Paul 1909; Weinreich, Labov & Herzog 1968; Aitchison 2003; Bybee 2003), the past decade has seen the emergence of what appears to be a new vibrating subfield in the study of language: the intermarriage of historical linguistics and psycholinguistics. A landmark paper, or rather series of papers, in this area was Jäger and Rosenbach’s (2008) study in Theoretical Linguistics on priming as an explanatory factor in grammaticalization, and the response articles in the same issue by Chang, by Eckhardt, and by Traugott, and a recent follow-up study by Hilpert and Correia Saavedra (2018). Other exemplar papers in this domain are Diessel (2012), drawing the connection between diachronic change and language acquisition, Grossman and Noveck’s (2015) target article on historical experimental pragmatics in the first issue of Linguistic Vanguard, Wolk et al. (2013), who set out to test the diachronic stability of processing factors, Hawkins (2012), invoking processing factors in the explanation of word order drift in the history of English, and a whole special issue of English Language and Linguistics (Bergs & Hoffmann 2017). It is this new current that has also resulted in the book under review.
The changing English language (TCEL) is the product of an inspiring workshop at the 2014 ISLE conference in Zürich, where the organizers were able to bring together several key scholars in the field of psycholinguistics and historical linguistics, and to pair them up to exchange thoughts on specific topics in their respective fields of expertise. This format was continued in the book, which consists of seven parts, discussing the role of (I) frequency, (II) salience, (III) chunking, (IV) priming, (V) analogy, (VI) ambiguity, and (VII) acquisition-and-transmission. With an introductory paper giving an overview of the field, this yields a total of fifteen chapters.
The editors have pulled off a marvelous job at laying out the field. As a collection of articles, this book is somewhat different from what one usually gets as the outcome of a conference or workshop: the collection of chapters attempts to give a near-exhaustive overview of major themes, rather than an assorted assembly of case studies, loosely held together by an introductory paper. The latter format may be suited for a specialized readership looking for state-of-the-art concrete case studies in a well-delineated field, but is far less suited for what the editors had in mind here: a programmatic introduction to an area that crosses sub-disciplines, each with a venerated research tradition of its own. The main attractiveness of the present volume is its tight top-down organization where contributors were asked to stick to these grand themes. The contributions in TCEL differ in their scope, and some are more case study-driven than others, but if you are interested in the role of frequency in historical corpus linguistics, in what exactly is understood by “salience” and how it is operationalized in psycholinguistics, or what the state of the art is in the field of priming, the chapters by Hilpert, Ellis, and Pickering and Garrod, respectively, are gems. This makes TCEL a valuable resource with a promised long shelf life that warrants the purchase of the hardcover.
A second quality of TCEL is that the chapters are rich in data, as well as methodological considerations illustrating the main points. All researchers brought together in this volume bring real data to bear, which can discriminate between competing theories. This can successfully be seen at work in the chapter on acquisition by Elena Lieven, for instance, where usage-based accounts are set off against Universal Grammar accounts, or in Felsers’s chapter on ambiguity, which lays out the discussion between serial syntax-first approaches and interactionist approaches, or Lopéz-Couso’s contribution on acquisition, where she weighs the evidence for and against the “recapitulation theory” (or ontogeny mirrors diachrony), or in De Smet and Fischer’s chapter, where they make the case that analogy is more constrained than the unpredictable, anything-goes force that some scholars have made it out to be.
My overall positive evaluation of TCEL does not mean that I think the book is totally devoid of shortcomings. The main point of critique is that the book focuses on English alone. In a sense, this is understandable: many of the factors under scrutiny, like frequency, priming, and analogy can be perfectly illustrated with English, with the advantage that quick introductions to the grammatical intricacies of unfamiliar languages or glossed examples are not necessary. Also, some of the phenomena investigated in “historical psycholinguistics” call for data-hungry methods. It’s not easy, for example, to measure a frequency-related factor like “dispersion” in corpora of languages with a patchy historical record, and English just happens to have exceptionally good coverage in terms of available text corpora. Moreover, the impressive line-up of key scholars in the field often have firsthand experience with English, so why not exploit that? On the other hand, it can be legitimately called into question whether the findings can all be easily extended to other languages. Cognitive diversity among different cultures should not be underestimated (see, e.g., Everett 2013:38-45) and salience, analogy, or ambiguity do not necessarily play out in the exact same way in all languages. This does not only pertain to lesser known “exotic” languages. Drastic differences can already be observed in two closely related languages like English and German: psycholinguistic research by Carroll, von Stutterheim, and Nuese (2004) has shown that these languages organize their discourse in very different ways. These languages differ in meaningful ways in what they consider salient, and this has obvious ramifications for diachrony (see Los 2012; Los & Starren 2012). TCEL’s unabashed focus on English alone, while understandable from a practical point of view, holds a danger that sweeping claims are made on the basis of what we know from WEIRD (i.e., Western, Educated, Industrialized, Rich, Democratic) people in psychology and from English for linguistics. An intriguing case in point is an article by Lewis and Frank (2016), who show a strong correlation between the length of a word and its psychometrically approximated “conceptual complexity,” but the correlation is much stronger in English than in any of the other seventy-nine languages they investigate. There are several plausible explanations for this, and Lewis and Frank (2016) speculate it may be an artefact of the test design, but it also possible that English is an outlier, linguistically. While the authors in TCEL, especially of the historical chapters, are cautious enough not to jump to Universal Grammar-like conclusions themselves, and while the authors of the psycholinguistic chapters occasionally elaborate on other languages, it is good to keep in mind that the findings may have limited universality. In his well-informed chapter on ambiguity, for instance, Denison has a plethora of historical data showing how vagueness in parts of speech leads to diachronic change, but the process is especially geared to English, a language notorious for underspecification in its part-of-speech system.
A second point of concern is the rather limited time depth, on average, of linguistic history when authors discuss firsthand data. Hilpert, in what is otherwise a marvelous chapter on frequency, goes back to the nineteenth century only, Mair considers only contemporary English data, and Denison has a few observations from the OED which stretch farther back in time, but the bulk of his paper on ambiguity is devoted again to nineteenth and twentieth century data. On the other hand, Bybee and Moder do consider datapoints from the entirety of the Early Modern English period, and De Smet and Fischer go back all the way to Old English. The concern about the time depth is the same as the language-particular focus discussed above: do all mechanisms work in the same way across the different stages of English? This issue is addressed in the book (e.g., in Traugott’s chapter; 96), but it is not a central concern overall. Baayen et al.’s “Ecclesiastes Principle” seems to be relevant to Late Modern English and to Present-Day English only, and the type accumulation curves that are discussed in the chapter by Hilpert (58-59) could have mentioned research by Bentz et al. (2014) showing that such curves have consequentially different parameters in inflecting language stages, like Old English. It might have been worthwhile to broaden the temporal scope a little more systematically, especially in view of the fact that English is exceptionally well documented at larger time depths.
A last point of critique is that after having worked through the book, the reader is left with the feeling that basically anything that has ever been proposed by psycholinguists is a strong force in language change. None of the proposed factors is dismissed as such. It is hard to see which of the seven factors (frequency, salience, chunking, priming, analogy, ambiguity, and acquisition-and-transmission) is a stronger effect in language change. Of course, this is partly due to the interlocking of all these effects. Hundt, Mollin, and Pfenninger remark that: For example, priming contributes to chunking, ambiguity facilitates analogical reasoning, and various types of frequency [. . .] as well as salience, for instance, play roles in acquisition and change. None of these factors thus drives language change by itself, so future studies of specific changes will need to consider their interplay. (17)
These points of critique are basically just reminders of what needs to be done to have the cross-pollination of psycholinguistics and historical linguistics come to full maturity. Overall, TCEL is a very helpful and timely collection of papers, though probably more so for historical linguists than for psycholinguists, and probably more for linguists of the usage-based than Universal Grammar-minded persuasion.
