Ever since the early twentieth century, and particularly since the publication of Chomsky & Halle’s The Sound Pattern of English (Reference Chomsky and Halle1968, SPE hereafter), English lexical stress has been one of the most studied topics of research in phonology. Many theoretical frameworks have used the highly complex and variable facts of English word stress as a testing ground for computational modeling, but while the generative and post-generative tradition has dominated the debates, the school of thought initiated by French linguist Lionel Guierre (Guierre Reference Guierre1979) has received less international attention. One of the enriching aspects of this collective volume is that it establishes a dialog between researchers following on from those two traditions. In addition, it seems that the quantitative analysis of English lexical stress has remained a relatively peripheral issue in the literature. Several contributions to this book adopt a data-driven approach (based on dictionary or/and oral corpora) to produce analyses of inter- and intravarietal stress variation.
The first two contributions are review chapters of theoretical approaches to word stress. Chapter 1, ‘On the treatment of English word stress within the generative tradition: History, concepts and debates’ by Jacques Durand and Eiji Yamada (pp. 6–52), is a historical account of the generative and post-generative theoretical approaches to English word stress. It begins with a review of the early twentieth-century observations on word stress by British and American authors who were relatively reluctant to formulate any generalizations on stress assignment, but provided valuable insights into the role of morphology and reflections on levels of stress, and paved the way for the concept of nuclear stress. Durand & Yamada go on to describe the breakthrough concepts introduced by Chomsky & Halle’s SPE, notably strong and weak clusters and their participation in the Latin stress rule and Main stress rule. The authors then discuss further theoretical developments within Metrical Phonology and Lexical Phonology, which provided new representational tools (Liberman & Prince’s Reference Liberman and Prince1977 tree and grid representations) and theoretical devices (Hayes’ Reference Hayes1982 concept of extrametricality which allowed the theory to do away with the underlying distinctions between [+/-long] vowels), as well as the exploitation of lexical strata in the application of phonological rules (Mohanan Reference Mohanan1982). The chapter finally describes the paradigm shift brought about by Optimality Theory (OT), with specialists (e.g. Hammond Reference Hammond1999) first rejecting phonological derivations altogether (strict parallelism), and later reintroducing them (serialism) in the form of Stratal OT. Durand & Yamada’s chapter is indeed an extremely rich and a valuable chronicle of what the (post)generative literature has proposed since the late 1960s to account for English word stress. However, it may be a challenging read for beginner students who are not already acquainted with one or several of the theoretical treatments discussed, due to the abundance of phonological and morphological concepts it introduces.
Chapter 2, ‘English word stress and the Guierrian school’ by Quentin Dabouis, Jean-Michel Fournier, Pierre Fournier and Marjolaine Martin (pp. 53–82), follows chapter 1 rather conveniently, since it provides an overview of an approach to English word stress initiated by Lionel Guierre (Reference Guierre1979), which has not received nearly as much international attention as the generative approach. While the Guierrian theory shares certain assumptions with the generative tradition, such as the central role of morphology and stress preservation, it does depart from it in some fundamental ways. The Guierrian treatment of stress is corpus-based (mostly exploits dictionary data), formulates stress rules based on spelling-to-sound relationships (which makes it convenient for L2 learners to whom the [+/- long] vowel distinction is initially unknown), takes into account semantically opaque constituents (inseparable prefixes), does not rely on syllable structure and distinguishes different subsystems (the ‘native’ English, loanword and neoclassical subsystems). The chapter presents the main tenets of the Guierrian approach and paves the way for further discussions within the volume, including phonologists’ traditional disregard for orthography, the treatment of opaque prefixes and stress variation among varieties of English.
The next three chapters provide illustrations of the Guierrian approach and discuss its theoretical implications in the perspective of a comparison with mainstream treatments. In chapter 3, ‘Stress placement in etymologically prefixed disyllabic noun–verb pairs revisited: A semantic and diachronic approach’ (pp. 83–116), Jérémy Castanier proposes a corpus-based analysis of the relatively overlooked role of (diachronic) semantics in explaining synchronic and diachronic stress variation in prefixed disyllabic multicategorial word pairs. This study challenges previous work in which lexical frequency was argued to be the most influential factor in the change from isotonic to diatonic stress in such pairs. Castanier’s study is based on 99-word pairs for which he explored 55 dictionaries (from the eighteenth to the twenty-first century) and systematically noted their stress patterns (including variants) and meanings. The results suggest that besides the general tendency for verbal roots to attract stress, many historical noun-to-verb conversions (often involving semantic specialization) resulted in /10/ stressed verbs. Castanier emphasizes the role of semantic asymmetry in the emergence of /10/ stressed nouns which do not have a verbal meaning (e.g. suspect (N) is not the act of suspecting or the fact of being suspected), and that of semantic symmetry in nouns such as concern (/01/) which preserve verbal meaning and stress pattern. He also acknowledges the role of lexical frequency in the appearance of stress-alternating pairs during the twentieth century, whereby a rise in verbal frequency generally corresponds with the change from /01/ to /10/ in the noun. By using a diachronic dictionary-based approach, this chapter makes a convincing point that English stress is a multidimensional issue that needs to take semantics into account alongside phonology and morphology.
Chapter 4, ‘English phonology and the literate speaker: Some implications for lexical stress’ (pp. 117–53), also adopts the Guierrian graphophonological approach to tackle the question of the role of orthography in phonology. Quentin Dabouis starts by questioning the traditional exclusion of spelling from phonological theory and briefly reviews recent work in psycholinguistics, reading acquisition and SLA which has demonstrated the impact of orthography on the (re)structuring of phonological knowledge. The author therefore embarks on a review of various phonological conundrums that can straightforwardly be solved if one takes orthographical knowledge into consideration. These notably involve vowel reductions (atom-atomic) and alternations (satire-satiric), stress-shifts (partial-partiality), as well as consonantal phenomena such as velar softening, consonant alternations (damn-damnation), the treatment of syllabic consonants (balsam-balsamic vs. plasm-plasmic), pre-/r/ dentalization in non-rhotic varieties of English, or the observation that orthographic consonant geminates phonologically behave as consonant clusters. Dabouis goes on to question the traditional generative view that stress is derived from the segmental lexical makeup of words (notably the [+/-long] distinction), and argues that a model integrating orthographic information to account for stress can not only solve the problem of stress assignment in words with prefinal orthographic consonant geminates (e.g. vanilla), but also provide a general framework that accounts for vowel distribution. While it may at first look like the bulk of this chapter is concerned with the graphophonological treatment of segmental issues, it quickly becomes apparent that the author is using them as a basis to develop compelling arguments relating to stress assignment.
Chapter 5, ‘The stress patterns of English verbs: Syllable weight and morphology’ by Quentin Dabouis and Jean-Michel Fournier (pp. 154–91), is a large-scale, data-driven analysis of stress in British English verbs. The authors mainly seek to evaluate the validity of two generalizations proposed in SPE: stress placement in English verbs is sensitive to (i) final syllable weight (prevalent in the generative literature) and (ii) the presence of opaque prefixes (also referred to as the Germanic Law, mostly upheld in the Guierrian literature). After providing arguments from psycholinguistics, reading studies, phonotactics and distributional recurrence to justify their view that morphologically opaque verbs should be treated as morphologically complex, they provide a description of their methodology, which involves a rigorous morphological classification of verbs as prefixed, suffixed, compound, other (truncations, back-formations, etc.) and simplex, based on the Laboratoire Ligérien de Linguistique’s (LLL) Dictionary Database (aggregate of three dictionaries). Dabouis & Fournier then carry out a detailed description of verbal stress in the various morphological categories, and point out that the predominance of semantically transparent left-edged constructions (compositional or using a free base) supports the hypothesis that the same type of morphology is identified in constructions where an opaque prefix is attached to a bound base. The following sections highlight that while the two generalizations are highly efficient and make identical predictions for most verbs, they come at a different theoretical cost (i.e. assuming that speakers recognize opaque prefixes vs. refining Moraic Theory through extrametricality or catalexis and variable weight of segments), and have different empirical scopes (i.e. the Germanic Law applies to non-nominal prefixed words, while the weight-based analysis applies to other categories if one accepts extrametricality). Rather than taking an exclusive theoretical stance, this chapter adopts an empirical approach in order to confront two theoretical claims and acknowledges their individual strengths. While they do reaffirm the relevance of semantically opaque constituents, the authors also show how factoring in the weight-based generalization results in a more accurate description of stress assignment in English verbs.
Chapter 6, ‘Lexical stress in varieties of Australian English: A corpus-based exploration’ by Anne Przewozny and Marjolaine Martin (pp. 192–239), shifts from a theoretical to a variationist scope. The authors propose a corpus-based analysis of Australian varieties of English with a threefold goal: (i) to describe the stress patterns of Aboriginal loanwords, (ii) to provide a systematic account of lexical stress in Standard Australian English, and (iii) to examine the claim that speakers of Standard Aboriginal English tend to shift stress to initial syllables. Following a review of the literature on variation in Australian English and Aboriginal varieties of English, Prezwozny & Martin first demonstrate that, across a corpus of 280 Aboriginal loanwords (two to five syllables), the Guierrian graphophonological rules governing lexical stress in English make slightly better predictions with regards to stress placement than the indigenous-influenced initial stress hypothesis. They also find, through a thorough analysis of dictionary data for standard American, British and Australian English, that among 3,700 stress-variable items, stress placement is highly stable across the three standard varieties (94 percent identical for main pronunciations), and that Standard Australian English exhibits less intravarietal stress variation than the other two. Through their exploration of the AusTalk and PAC-Australia corpora, the authors disprove the hypothesis according to which Standard Aboriginal English typically shifts stress to initial position; they find that among 40 items, only those with a dictionary /201/ pattern shift stress significantly, and reckon that that is in fact a general tendency in English. This chapter illustrates the way theoretical and descriptive accounts of English word stress can be refined by systematic analyses of dictionary and oral corpora.
Another outer-circle variety is the focus of chapter 7, ‘Melodic complexity and lexical stress in Singapore English: An experimental study’ by Gabor Turcsan and Oriana Reid-Collins (pp. 240–63). The goal of their chapter is to contribute to the description of word stress in Singapore English, based on a PAC corpus, and on the results of a nonce-word production experiment which was previously submitted to inner-circle speakers, thus allowing for comparability. The background section of the chapter looks at previous literature on the phonological structure of the variety and compares it with the authors’ own observations in their PAC-Singapore corpus. With regard to the vowel system, they do observe the type of neutralizations (compared to RP) reported in other studies in their mesolectal (conversational) speech data, but argue that acrolectal (wordlist) productions maintain the same phonological contrasts as RP, demonstrating the importance of exploiting data giving access to several speech registers. Similarly, their observations confirm that greater prominence and lower amount of vowel reduction, due to a tendency to treat grammatical words as lexical words and lesser incidence of reduction of weak syllables in lexical words, contribute to the perceived syllable-timed rhythm of Singapore English. Following a description of their experimental protocol (reading of carrier sentences with disyllabic nonce-words controlled for syllable-weight in either nominal or verbal position), Turcsan & Reid-Collins show that the level of agreement in stress placement among their Singaporean participants is similar to the one observed among inner-circle subjects. However, they point out that Singaporeans have an overall greater tendency to stress these nonce-words on the right. It is especially true in verbs, where even a light final syllable is most often stressed, but can also be observed in nouns, where the preference for left-syllable stress is lesser than for inner-circle speakers in all weight configurations, except H(eavy)-L(ight). The authors conclude that experimental data can be a useful complement to corpus data in uncovering mesolectal phonological regularities.
In chapter 8, ‘Input optimization and lexical stress in English’ (pp. 264–87), Michael Hammond proposes a theoretical and statistical analysis of exceptional right-syllable stress in English nouns within the framework of Optimality Theory (OT). Hammond first lays down generalizations on the patterning of stress in English nouns relating to syllable weight and morphology. He then proposes an analysis that involves, among others, constraints on footedness of the rightmost syllable and on foot structure, as well as constraints on foot alignment with the right edge of the word. He also suggests some refinements of moraic theory distinguishing vocalic and consonantal morae. The following section introduces and illustrates two key concepts of the chapter: phonological complexity is defined, in optimality theoretic terms, as an assessment of the number of constraints winning candidates violate and their rank within a phonological system, and Input Optimization is construed as the selection of input forms which minimize the phonological complexity of that system. The predictions Input Optimization makes about the distribution of stress patterns in English nouns are confirmed by dictionary data, which show a statistical preference for winning candidates which violate the lower-ranked constraint (all feet are aligned with the right edge) over those which violate the higher-ranked one (the rightmost syllable is not footed). As is the case with other contributions within this volume, this chapter shows how phonological theory can advance with the support of the type of statistical data corpora and dictionaries provide.
Finally, in chapter 9, entitled ‘A solution to theoretical shortcomings in the stress assignment of words in English’ (pp. 288–324), Eiji Yamada presents an overview of Positional Function Theory (PFT), a new approach to word stress assignment. First, he describes some limitations of SPE and classic OT approaches in accounting for the absence or presence of subsidiary (pre-tonic tertiary or post-tonic secondary) stress. Specifically, SPE cannot account for the unstressed antepenultimate in e.g. information, lamentation /2010/, and classic OT struggles to account for the tertiary stress in condensation, importation /2310/. The following section proposes a unified system capable of correctly predicting subsidiary stress in the two exceptional word types mentioned above, with the help of a battery of positional functions. Main stress assignment can demonstrably be accounted for with a reduced set of positional functions, although the latter require (directionality and environment) adjustments. However, Yamada acknowledges the fact that his treatment yields a number of incorrect stress patterns (notably in verbs with a heavy final syllable (e.g. designate) and certain adjectives in ‑ory or -ative), which require lexical marking in order to shift the order of the incorrectly assigned primary and secondary stresses. The author further discusses the advantages of PFT over OT in handling the secondary–tertiary stress differences, and explains how the optional activation of a positional function can account for stress variation. He explains that PFT is a superimposition theory, in that it is not derivational and does not rely on constraint hierarchy, but rather computes a stress value for each syllable through the interaction of constantly activated positional functions which are mostly conditioned by adjacent or neighboring syllables.
Books dealing with word stress usually either exploit a range of languages to discuss theoretical aspects (Hulst Reference Hulst2014; Goedemans et al. Reference Goedemans, Heinz and van der Hulst2019), or review word stress placement within broader descriptions of individual varieties of English, sometimes adopting a more pedagogical stance (Staun Reference Staun2010; Cox & Fletcher Reference Cox and Fletcher2017). Unlike other collective volumes, New Perspectives on English Word Stress focuses on English lexical stress exclusively, and provides a solid overview of current theoretical and empirical research on lexical stress. In my view, it is therefore intended for advanced students and researchers who are already versed in phonological and morphological theory. For instance, introductory readings in OT and PFT would certainly help the reader grasp the subtleties of chapters 8 and 9, respectively. The various chapters are conveniently organized so as to offer a logical progression from reviews of theoretical models, to empirical approaches to stress variation, before circling back to remaining theoretical issues. The many cross-references to chapters within the volume testify to the editors’ effort to establish a dialog between the contributions. However, while the contributors take heed of the clarity of their explanations, the many end-of-chapter notes would have been more helpful in the form of footnotes, especially in the more theory-heavy chapters. This criticism does not affect the quality of this volume, which is a must-read for students and researchers interested in English lexical stress.
