One of the promises of Optimality Theory (OT) is that a diverse set of processes can be understood in terms of the interaction of a small number of output-oriented constraints. The analysis of vowel patterns in Vowel patterns in language delivers on this promise. Offering a survey of vowel patterns organized insightfully around a small number of constraints, the book constitutes a significant contribution to the OT literature and to the description of vowel patterns more generally. Analysis of metaphony in minor Romance languages, including a thorough case study of the Servigliano dialect of Italian, three stages in the diachronic development of German umlaut, and a broad sampling of vowel patterns in other languages are used to address the central question of how vowel quality interacts with word position.
A major accomplishment of the work is a unification of vowel patterns under the common theme of prominence and the umbrella of a common constraint schema, termed generalized licensing. In each of the more than 25 vowel patterns for which a formal analysis is provided, a constraint that licenses the occurrence of a marked feature in a prominent position is active. Attested patterns of vowel deletion, vowel reduction, feature spreading, feature copying, infixation, and coalescence, as well as static distributional patterns (for example, the distribution of non-high round vowels in Classic Mongolian) are all analysed in terms of generalized licensing constraints and their interaction with other constraints (including positional faithfulness).
The generalized licensing schema requires two elements: (i) a prominent position, which can be an initial syllable, stressed syllable, morphological root, or morphological stem, and (ii) a marked feature. Constraints in this schema function as markedness constraints penalizing instances of marked features that do not have a surface correspondent in the specified prominent position. Both shared features (‘indirect licensing’) and feature chains (‘identity licensing’) can satisfy licensing constraints, and Walker discusses languages making use of both of these strategies. In her analysis of Old High German, Jawaru, Central Veneto, and Lango, licensing constraints are satisfied by sharing marked features across prominent and non-prominent positions. In these languages, alternations in prominent positions are triggered by marked vowels in weak positions. The factorial typology also predicts indirect licensing patterns in which alternations target marked vowels in weak positions (unless the marked vowel also occurs in a prominent position), and the book provides a discussion of representative patterns in Buchan Scots, Macuxi, Ticinese, and C'Lela. Evidence for feature sharing (as opposed to copying) in these languages comes from the presence of intervening phonological material that blocks harmony at a distance. In Walker's analysis of Eastern Meadow Mari, several Romance dialects (Ascrea, Lena, Eastern Andalusian, Francaville Fontana), and a later stage of Old High German, licensing constraints are satisfied by copying marked features from a weak position into a prominent position. In contrast to cases of indirect licensing, each of these patterns shows evidence of long-distance harmony.
Two other types of licensing patterns are also given extensive coverage, ‘direct licensing’ and ‘maximal licensing’. Direct licensing refers to patterns in which a marked feature surfaces only in a specific prominent position and is prohibited elsewhere. Patterns of direct licensing in Esimbi, Modern Standard German and the Romance dialects of Liguria and Old Piedmontese are accounted for using the same licensing schema deployed for patterns of indirect licensing and identity licensing. Patterns of maximal licensing involve those in which marked features are required in every segment in a phonological domain (as, for example, a prosodic word). A maximal licensing constraint schema is proposed to account for these patterns in Baiyinna Orochen, Vinalopo Mitja, and Servigliano. The treatment of Servigliano goes into considerably more depth than the discussion of any of the other languages in the book. Making use of both maximal licensing and prominence-based licensing, the analysis covers three patterns of vowel harmony as well as a pattern of vowel reduction in unstressed syllables.
The functional grounding of licensing constraints is another recurrent theme in the book. Licensing constraints are hypothesized to have a functional motivation in that they reduce perceptual difficulty, enhance ease of articulation and/or otherwise facilitate processing. In many cases it is the reduction of perceptual difficulty that is singled out as the primary functional motivation for licensing constraints in a given language. The basic idea is that requiring marked features to occur in a prominent position enhances the perceptual salience of those marked features and thereby reduces perceptual difficulty. The book adopts the position that licensing constraints, at least in some cases, are induced by making use of a speaker's phonetic knowledge. Licensing constraints induced in this fashion asymmetrically favour the licensing of marked features. This leads to the prediction that if a given feature requires licensing by a prominent position, then features that are more marked will also require licensing by that position. With the caveat that the relative markedness of certain features appears to vary from language to language, the patterns presented in the book are consistent with this prediction.
Vowel patterns in language is organized into nine chapters. Chapter 1, ‘Introduction’, gives a brief overview of the book. Chapter 2, ‘Preliminaries: Functional grounding’, reviews literature on the relation between phonetics and phonological grammar, discusses evidence for vowel markedness and positional prominence, and presents the functional grounding hypothesis. Chapter 3, ‘Generalized licensing’, introduces the formalism used throughout the book; and Chapter 4, ‘Typological predictions’, reports the patterns predicted by factorial typology of the proposed constraints (computed using OTSoft 2.1). The core analysis is presented in Chapters 5–8. Each of these chapters discusses a different type of vowel pattern: Chapter 5 analyses patterns of ‘Indirect licensing’, Chapter 6 patterns of ‘Identity licensing’, Chapter 7, ‘Direct licensing’, and Chapter 8, ‘Maximal licensing’. Chapter 9 presents ‘Conclusions and final issues’, including a brief discussion of trajectories of historical change in vowel patterns and possible avenues for addressing the ‘too-many-solutions’ problem for licensing constraints.
The book offers a sophisticated balance between the detailed analysis of individual languages and commonalities observed across languages. This balance is captured eloquently by the OT framework. The factorial typology presented in Chapter 4 lays out typological expectations which are instantiated through the analysis of individual languages in Chapters 5, 6 and 7. In these chapters, licensing constraints drawn from the generalized licensing constraint schema are deployed alongside an established set of formal tools from the OT literature. To handle vowel patterns restricted to specific morphemes or to classes of morphemes, constraints are indexed to lexical items, following Pater (Reference Pater and Parker2009) and antecedent work on constraint indexation. Patterns sensitive to minimal contrast, for example, those found in Western Asturian (200–206), are dealt with by modifying the Gen component of the OT grammar to enumerate systems of contrast, following Campos-Astorkiza (Reference Campos-Astorkiza2009). Devices such as the lexical indexation of constraints and the representation of minimal contrast allow the analysis to shape the core typological predictions to fit the idiosyncrasies of different languages without losing sight of what is common across them.
Most of the vowel patterns under study have alternative analyses in the literature. In many cases the spirit of the past analysis is easily incorporated into the generalized licensing approach. In some cases, however, alternatives offer different types of intuitions about the vowel patterns under study. Each of Chapters 5–7 end with a succinct summary of alternative analyses and arguments in favour of the generalized licensing approach. A central line of argumentation is that the generalized licensing approach captures a broader range of patterns than the alternatives. The discussion of alternative analyses of metaphony in the central Asturian variety of Lena at the end of Chapter 6 offers a good example of this. Lena metaphony is licensed by a stressed syllable, it is non-local, and it is only triggered by certain affixes. In this case and (as mentioned above) throughout the book, morpheme-specific patterns are dealt with by indexing constraints to specific lexical items. In the analysis of Lena, lexically-indexed licensing constraints are used to restrict the triggers of metaphony to inflectional suffixes, and lexically indexed faithfulness constraints are used to block vowel raising in stems that exceptionally resist metaphony. Alternative analyses discussed at the end of Chapter 6 include Hualde (Reference Hualde1989), which focuses on the metrical aspect of the pattern, and Finley (Reference Finley2009), which pursues the intuition that metaphony is the realization of morphemes that trigger it. The alternatives offer plausible formal accounts of Lena metaphony that highlight different aspects of the pattern and capitalize on different intuitions. Walker argues that neither alternative can be extended to capture the range of vowel patterns within the scope of generalized licensing. While metaphony in Lena is consistent with the metrical foot as the domain for feature assimilation, other vowel patterns covered by generalized licensing are not metrically driven. In some, the licensing position is prominent by virtue of being word-initial (but not necessarily stressed) or by being the morphological root. Similarly, not all vowel patterns are restricted to particular morphemes, as in Lena. Some (for example, in Central Veneto) are purely phonological and show up both as distributional patterns and alternations. The generalized licensing schema draws this range of patterns under the same formal analysis by abstracting the notion of prominence over different types of word positions.
The idea that prominence is an abstract phonological entity enables generalized licensing to achieve broader typological coverage than its alternatives. From the generalized licensing perspective, a wide range of vowel patterns are driven by constraints relating marked vowel features to various types of positions that can be categorized as prominent. Since the positions that are argued to be prominent for the purposes of generalized licensing are prominent for different reasons, they are likely to be prominent to different degrees. Nevertheless, all prominent positions are treated identically within the generalized licensing schema. Whatever processing differences may exist between the various prominent positions, they are flattened in the phonology. This disconnect between processing and the phonological grammar is consistent with the position that grammar is influenced by (but distinct from) ‘raw phonetics’.
Although the contributions of the book in the areas of description, typology and formal theory are substantial, the support offered for the functional grounding hypothesis is rather weak. Firstly, the book outlines (in Chapter 2) a particularly careful position on the debate about the source of phonetically natural phonological patterns. It maintains that functionally grounded phonological constraints can be induced synchronically from phonetic knowledge, but also that some patterns, for example, those for which historical changes have obscured the functional grounding, might no longer be induced with reference to phonetic knowledge. This position seems to soften the main prediction that licensing constraints asymmetrically favour marked features. Second, direct evidence for whether (and to what extent) the vowel patterns in the languages under study do or do not facilitate perception, processing or articulation, appears to be entirely lacking. It is noted that ‘much remains to be explored in issues surrounding markedness of specific vowel qualities in weak contexts, and its relation to … perception … ease of articulation … and so on’ (32), but little coverage is provided even of the large body of existing work on vowel perception or on listener-motivated perceptual enhancement (see, for example, Uchanski Reference Uchanski, Pisoni and Remez2005) – areas which could be helpful in evaluating the functional grounding hypothesis. In short, the functional bases of the patterns under study are somewhat speculative. Nevertheless, the typology of patterns described in the book provides the landscape for future work in this area.
Overall, Vowel patterns in language is a substantial contribution to the OT phonology literature and destined to become a standard reference for vowel patterns. The theoretical proposal unifies analysis of a diverse range of phonological phenomena, including both static distributional patterns and the gamut of vowel-related phonological processes, with a small set of OT constraints. Well-written and well-organized, the book will satisfy a range of different types of readers, serving as both a typological survey and as a theoretical treatise on the abstract nature of prominence in phonological grammar.