This well edited volume collects six new essays focused on the representation and distribution of laryngeal features in Dutch. That this particular Germanic language should be singled out for such analysis and discussion is appropriate, as Dutch is distinctly “ungermanic” in its laryngeal organization, closer in fundamental ways to its Romance neighbors than to German, a language with which, however, it does share the property of final laryngeal neutralization. For reason of its laryngeal hybridity alone, then, the volume is a welcome contribution to the field; but the laryngeal phonology of Dutch has also raised a range of incisive theoretical questions, many of which are addressed head-on in the essays gathered here.

Following a brief introduction by the editors (pp. vii–ix), Wim Zonneveld opens the volume with his survey article, “Issues in Dutch devoicing: Positional faithfulness, positional markedness, and local conjunction” (pp. 1–40). This chapter presents a valuable and com-prehensive overview of the facts relating to voice assimilation as well as final devoicing and how these have been described in the modern scholarship on Dutch phonology, from his own early rule-based work at the end of the 1970's to the interaction of markedness and faithfulness constraints today. The data are organized and laid out particularly well and will assure a lasting contribution of this overview article, even if its more speculative characterizations toward the end under modified versions of optimality theory do not hold up in the long run.

A key contribution to the volume is that of René Kager, Suzanne van der Feest, Paula Fikkert, Annemarie Kerkhoff, 9 Tania S. Zamuner, “Representations of [voice]: Evidence from acquisition” (pp. 41–80). This is an important psycholinguistic study evaluating the question of whether the laryngeal phonology of Dutch is organized along the lines some researchers (especially Joe Salmons and I) have proposed for most of the Germanic languages, with the voiceless stops distinctively marked (by the feature [spread glottis]) and the voiced ones unmarked (laryngeally empty), or along the lines of the Romance languages, with the voiced stops marked (by the feature [voice] or its analog) and the voiceless ones unmarked. The chapter shows persuasively that the latter characterization is correct for Dutch and the former for German, based on the distribution of learning errors in first language acquisition of these languages as well as English. This is a well-crafted piece of science, to my mind the premier piece in the volume, and anchors the book in experimental psycholinguistics as well as in theoretical phonology while highlighting the increasingly familiar distinction between “aspiration” and “voice” languages.

In “Exceptions to final devoicing” (pp. 81–98), Marc van Oostendorp investigates what appears to be the exceptional failure of final fricatives to devoice in certain dialectal varieties of Dutch. The illustrations all indicate that this phenomenon arises only in the (unin-flected) first person singular of verbs whose stems end in either a labiodental or velar fricative, as in ik bliv ‘I stay’ or ik ma ‘I may’. The author relates that no dialectal instances of voiced final coronal fricatives were found, and that there were more instances of labial than of velar voiced final fricatives; moreover, “in some cases the segment alternating with voiceless [f] was transcribed phonetically as [w] rather than [v]” (p. 82). Perhaps nothing more can be said about these observations, but the suggestion does present itself here that possibly these dialects continue to represent surface [v]~[w] as underlying /w/, which, as a phonological sonorant, naturally would not be susceptible to final obstruent devoicing. Whether the minority cases with final [] could be interpreted in these dialects in a parallel way is less apparent, and the author dismisses the /w/ idea in any case on p. 87, still noting, however, the celebrated phonological ambivalence of phonetic [v] between sonorant and ob-struent in other languages, most notably, Russian. Nonetheless, the paper raises interesting issues in the description of the final devoicing pheno-menon when the dialects of Dutch are taken into account, too, and is a core contribution to this volume surveying the full range of (de)voicing phenomena in Dutch.

Petra van Alphen, in “Prevoicing in Dutch initial plosives: Pro-duction, perception, and word recognition” (pp. 99–124), shows that “prevoicing” (negative VOT) is a chief characteristic of the Dutch voiced stops, as it is in the Romance languages, but that prevoicing is often absent in productions of word-initial voiced stops in Dutch, a fact that brings them acoustically quite close to their voiceless unaspirated counterparts in the language. Nonetheless, listeners generally keep the two stop types distinct due to the effects of “lexical priming:” if there exists no word in Dutch with an initial voiceless stop that would be homophonous with the word whose initial voiced stop has lost its prevoicing, perception of the “devoiced” stop is still as voiced, as in bloom ‘flower’, there being no word *ploom with which it could be confused; but if there is such another word, misidentification is more likely to arise, as in the devoiced variant of beer ‘bear’ being identified as peer ‘pear’. This is important work pointing to role of lexical access in phonological identification and is a worthy contribution to the volume.

Wouter Jansen offers another insightful phonetic study bearing on the phonological analysis of Dutch laryngeal distributions. In “Dutch re-gressive voicing assimilation as a ‘low-level phonetic process’: Acoustic evidence” (pp. 125–151), the point is established that a degree of voicing coarticulation occurs in stop+fricative+stop clusters, as in /fits+bel/ ‘bicycle bell’, so that, depending on rate of speech and perhaps other factors, the /s/ may be partially voiced or the /b/ partially devoiced. This is good experimental work and fits well into the volume at hand.

Finally, Mirjam Ernestus 9 Harald Baayen contribute an interesting experimental chapter titled “Intraparadigmatic effects on the perception of voice” (pp. 153–173). This paper reopens the discussion of the extent to which final devoicing is actually neutralizing in languages like Dutch, arguing that the phonetic qualities of nonalternating (“underlyingly voiceless”) voiceless final stops are somewhat different from those of alternating (“underlyingly voiced”) stops. Thus, the final burst at release of [t] < /t/ in a word such as krant ‘newspaper’ (plural kranten) has been observed experimentally to be more prominent than that associated with [t] < /d/ in a word such as mand ‘basket’ (plural manden). In one experiment, the authors presented a native speaker a written list of (real or possible) Dutch words—baard, baart, baaz, baas, band, bant—and instructed him to “realize final obstruents as voiced, when they were represented [in spelling] as voiced, and to realize them as voiceless, when they were represented [in spelling] as voiceless.” The speaker largely succeeded in this task, and listeners, moreover, were able to perceive the distinctions among finals that the reader had made. This is no surprise to those familiar with the work of Marios Fourakis and me (“On the ‘incomplete neutralization’ of German final obstruents,” Phone-tica 41.140–149, 1984), and follow-up studies by others on Polish (notably, Wiktor Jassem, Alexis Manaster-Ramer) and other languages. Fourakis and I showed that Labov's celebrated notion of the “observer's paradox” plays a confounding role in the laboratory analysis of German final stops unless these are elicited in a way that conceals the intent of the experiment rather than focusing on it, as the chapter under discussion does. This work nonetheless is a substantial contribution to the volume for its many careful measurements and the questions it raises with respect to listener awareness of paradigm effects in careful speech, underscoring that final voiceless obstruents which alternate with voiced in the same paradigm (for example, [t] in mand) can be slightly “less voiceless” than those which are voiceless in all forms of the paradigm (for example, [t] in krant). Still, the inherent limitation of the experi-mental methodology needs to be recognized, namely, that the elicitation of “natural speech” under laboratory conditions requires clever design in order to deflect the probable prospect of speaker hypercorrection and the production of artificial speech.

Voicing in Dutch: (De)voicing—phonology, phonetics, and psycho-linguistics is a timely, solid collection of important essays relating not just to the laryngeal phonetics and phonology of Dutch but to Germanic more generally. I think it will be influential and lasting.