For those interested in understanding the historical origins of the Latin language, and tracing its development from its ancestor Proto-Indo-European, perhaps the most frequent difficulty is that of explaining precisely why a Latin word has the vowels it does. While Classical Greek preserved the inherited Indo-European vowels system more or less intact, the vowels of Latin words are often barely recognisable, showing the effects of numerous complex and interacting changes.
In this revised version of a University of Oxford D.Phil. thesis, S. investigates five problematic, related phenomena which primarily affected the vowels, but also the consonants, of pre-Classical Latin, all of which depend to one extent or another on the structure of the Latin syllable. After introducing the study of Latin historical phonology in Chapter 1, in Chapter 2 S. discusses vowel colouring before l (e.g. the alternation in the first vowel of velle, volo, vult, etc., which all derive from original *vel-). Focusing on the implications regarding the articulation of l, the investigation concludes that l was ‘dark’ at the end of a syllable, which S. interprets by reference to a phonological feature [+back], ‘clear’ when double ll, interpreted as a feature [–back], and underspecified elsewhere. ‘Dark’ l caused a preceding vowel to be coloured as u; ‘clear’ l had no colouring effect; underspecified l adopted degrees of ‘darkness’ depending on the following vowel, resulting in colouring of a preceding vowel to o or u.
Chapter 3 discusses inverse compensatory lengthening, commonly called the ‘littera’ rule, the apparently sporadic Latin development whereby a sequence of long vowel followed by single consonant became a sequence of short vowel followed by geminate consonant (e.g. lītera > littera). S. demonstrates that this change did occur sporadically in early Latin, in three phonetically motivated contexts, but he also shows that much of the supposed evidence for the rule can receive alternative explanations.
Chapter 4 investigates the syllabification of stop + liquid (‘TR’) sequences. S. provides a valuable summary of the rules of Latin vowel reduction, key to uncovering the syllabification of TR sequences since open and closed syllables show different reductions (cf. e.g. the alternation between vertebra and *verticla > verticula, both from original *verte-). S. shows that syllabification was primarily conditioned by morphology, but that in some contexts the evidence for this is obscured by phonologically conditioned vowel reduction patterns affecting some open syllables before TR.
In Chapter 5 S. provides an account of the constraints on epenthesis (vowel insertion to split a consonant cluster) in word-internal stop + l sequences (e.g. populus vs pūblicus, both from *popl-). He argues that the earliest change was in bl sequences (dated to the fourth century b.c.), followed by epenthesis in kl sequences (late-fourth to mid-second centuries), with epenthesis in pl sequences, constrained by several factors, the last (mid-third to second centuries). While epenthesis in bl and pl sequences is explained phonetically, that in kl is explained as a structurally motivated change, initially based on a morphological correspondence between suffixes in *-blom > -bolom and *-klom (→ -kolom).
Chapter 6 investigates the patterns in and constraints on consonantal assimilations, for example why we find affere < *ad-fer-, but Plautine ecfer not *effer. S. explains the patterns very effectively by reference to two phonological hierarchies referring to place and manner of articulation.
S.'s approach is careful, methodical and clear. The introduction to each chapter helpfully summarises the phenomenon, the problems and the proposed explanations. S. provides very detailed and balanced evaluations of the evidence for and against previous proposals, and detailed yet easily readable collections of data for the relevant sound rules (and morphological processes, where these are relevant). For the most part, the analyses account for the full range of data neatly and plausibly; only once or twice must he assume ad hoc explanations for exceptional forms.
The considerable volume of data is discussed clearly and usually in detail. For two of the phenomena, the ‘littera’ rule and vowel reduction before TR, the number of individual forms concerned is sufficiently large that detailed discussion in the main body of the work is avoided. For the ‘littera’ rule, a 26-page appendix provides brief discussion of each form, but for these forms and those relevant to vowel reduction before TR, S. refers the reader to his personal website, where he promises in-depth discussion of individual forms. At the date of writing, however, the relevant discussions are not available on the website. The book is not over long, and one feels that two long appendices containing the promised in-depth discussions would not have been out of place.
Increasingly, historical and comparative linguistic study of the Classical languages integrates or is informed by advances in modern theoretical and experimental linguistics. This is true of S.'s work, but he avoids the danger of thereby making his analysis incomprehensible to the average Classicist or traditional comparative philologist. His mastery of both the Latin language and modern phonetics/phonology is evident throughout, and interweaved in such a way that Classicists, traditional comparative philologists, and theoretical phoneticians and phonologists can appreciate and be informed by his work. S.'s approach to the explanations of sound changes is broadly ‘reductionist’, aiming to demonstrate that the phenomena under investigation can be explained mainly by reference to how words are articulated by speakers and perceived by hearers, rather than by reference to some more abstract principle of linguistic organisation in the mind. In this way, the abstractions of phonological theory, though appealed to where necessary, are largely avoided, explanation resting on the rather more concrete ground of phonetic reality. While S. aims at a reductionist account, seeking to explain as much of the data as possible purely by recourse to phonetic factors, he does not attempt to force an interpretation on data that does not fit this approach, but allows the data to speak for itself, making recourse to phonological, prosodic and morphological factors where necessary.
There are only one or two negatives, from a Classicist's perspective, all minor. Some technical phonological and phonetic terms are used with no introduction or explanation, for example ‘F1’/‘F2’/‘F3’/‘F4’ (the ‘formants’ or amplitude peaks in the sound spectrum), first used on page 18 and throughout the work, or ‘VOT’ (‘voice-onset time’, a feature of the articulation of stops), used on pages 176 and 185. Occasionally the use of such terms without explanation renders a paragraph rather confusing to the non-phonetician reader, but such paragraphs are by far the exception, and in general S. does a remarkably good job of making his phonetically based explanations comprehensible. Perhaps the most disconcerting thing, at least initially, is the way S. chooses to render long vowels, using the length mark of the International Phonetic Alphabet, instead of the more usual macron, so that for example implōrātōque is rendered as imploːraːtoːque on page 93 (while in all other respects a thoroughly ‘normal’, non-phonetic, Latin spelling is used, including the widespread but phonetically illogical use of i to cover both semi-vowel and vowel realisations of the Latin letter I coupled with a distinction between v and u for the semi-vowel and vowel realisations of Latin V).
Besides the more general value of the book, there are a number of valuable etymological notes on particular words which add to the interest of the work (e.g. the footnote on facilis on p. 130). The scope of the work spreads well beyond the five phenomena that are strictly the object of investigation, such that this work, while by no means an overview or introduction to Latin phonology, will undoubtedly have importance for any scholar working on Latin phonology.
