Date: Fri, 27 May 2005 16:38:30 -0400 From: Marc Picard Subject: Phonetics and Philology: Sound Change in Italic
AUTHOR: Stuart-Smith, Jane TITLE: Phonetics and Philology SUBTITLE: Sound Change in Italic PUBLISHER: Oxford University Press YEAR: 2004
Marc Picard, Concordia University
The purpose of this book is to analyze "how the Proto-Indo-European voiced aspirates developed in the ancient Italic languages" in order to demonstrate "how phonetic theory can be used to evaluate and motivate accounts of reconstructed sound change provided by philology" (p. 1). More specifically, given the following segmental correspondences between Proto-Indo-European (PIE) and Latin: PIE Latin *bh f/b *dh f/d *gh h *gwh f/k
Jane Stuart-Smith (henceforth JSS) sets out to determine what sound changes could most plausibly account for these radical transformations. Her basic tenet in this endeavor is that "phonetic theory provides the key to this particular problem of philology, because it seems to be the only way of establishing an appropriate route from reconstruction to reflex" (p. 1).
In Chapter 1, the author lays the groundwork for her study of the evolution of the PIE voiced aspirates in Latin and Italic in general. As she notes, the main problem is that although philology, by which is meant the study of texts in conjunction with comparative and historical linguistic analysis, is a time-tested method for phonological reconstruction, it is of little use in trying to establish the actual paths that lead from a proto-segment to its reflex(es) in a daughter language. In order to resolve what she terms "the route problem", JSS proposes to bring to bear the criterion of phonetic plausibility which is "evaluated according to two sources of information: phonetic predictions for sound change derived from applying a phonetic model of sound change to the sounds in question, together with an assessment of historically-attested parallel developments of voiced aspirates in Indo-Aryan" (p. 6).
Two preliminary issues are addressed in Chapter 2. The first involves the controversial question of whether PIE really had voiced aspirates given the widespread claims that no such segments can exist in a language without its having voiceless counterparts. Although a few Indo-Europeanists have posited the existence of a voiceless aspirated series, while others have tried to remedy the situation by adopting the so-called Glottalic Theory wherein the three traditional stop series *t, *d, *dh are replaced by *t(h), *t', *d(h) (where (h) represents redundant aspiration), no consensus has been reached on these proposals. Thus, given that "[m]any Indo-Europeanists would accept the voiced aspirates as part of an unusual three-series stop system", JSS simply assumes that "the starting point of our development was a series of voiced aspirate stops, realized phonetically as breathy voiced stops, similar to those which are found in contemporary Indo-Aryan languages.
The second issue has to do with the exact route that the PIE voiced aspirates took on the way to their final Italic destinations. The author notes that two principal explanations have been put forth in this regard (with *bh serving as a model for the whole series): the one by Ascoli in the 1860s whereby they were first devoiced (*bh > *ph) and then fricativized (*ph > f), with Latin undergoing a further word-medial development (f > b), and the one by Rix in the 1950s according to which the voiced aspirates were first fricativized (*bh > *B) and then devoiced (*B > f) except medially in Latin (*B > b). She thus sees her task as trying "to identify the most plausible explanation for the Italic development of the PIE voiced aspirates, be it one of the two existing, or something different." (p. 29).
The third chapter, which is entitled "Philology: The Evidence for the Italic Development" and which runs for well over 100 pages, aims to establish the Italic reflexes for the PIE voiced aspirates and to answer the three following questions: 1. Were the reflexes voiceless in word-initial position and voiced in word-internal position in all the Italic languages? 2. What conditioned changes can be established for Italic? 3. How feasible is it to assume a common Italic voicing of word-internal fricatives?
After an exhaustive examination of all the data relating to the complex developments of the segments in question in the two main branches of Italic, viz., Latito-Faliscan and Sabellic (Oscan, Umbrian, South Picene), JSS arrives at the following answers to each of her questions: 1. It seems more than likely that the reflexes for the PIE voiced aspirates in all the Italic languages showed a split distribution, voiceless in word-initial position, voiced in word-internal position. 2. A number of conditioned treatments of the PIE voiced aspirates are attested across the Italic languages, both by place of articulation of the original stop, and by its position in certain clusters: following a nasal; following a sonant nasal; and in clusters with *s or *t. 3. It is not easy to assume a common Italic voicing of fricatives word- internally in a voiced context.
In Chapter 4, the author endeavors to evaluate the explanations that have been put forth for the development of the PIE voiced aspirates in Italic. Incorporating the findings from Chapter 3 into the Ascoli and Rix models outlined above, she arrives at the following putative developments: Ascoli PIE *bh > Proto-Italic *ph > *f-/*-B- > Sabellic f-/-B-, Latin f-/-b- Rix PIE *bh > Proto-Italic *B > *f-/*-B- > Sabellic f-/-B-, Latin f-/-b-
Moreover, she states that the distribution of the Italic reflexes makes a third account possible: Word-initial position PIE *bh- > Proto-Italic *ph- > *f- > Sabellic f-, Latin f- Word-internal position PIE *bh- > Proto-Italic *-B- > Sabellic -B-, Latin -b-
"Phonetics, Predictions, Parallels" is the title of Chapter 5, and in it JSS assesses the phonetic plausibility of these three proposed routes for the Italic development of PIE voiced aspirates by resorting to evidence drawn from phonetic theory, on the one hand, and observed parallel developments in various languages, on the other. Underlying this analysis is John Ohala's model of sound change which, in her estimation, offers a motivated connection between synchronic variation and potential sound change. More specifically, this model rests on the premise that sound changes "result from the listener's inappropriate processing of variation which in turn is a direct consequence of universal constraints on the articulation, acoustics and perception of speech" (p. 10).
Having thus applied the principles of this phonetic model of sound change to the case at hand by first making a number of predictions for change and then confirming them through observations of attested parallel developments in the history of Indo-Aryan, the author arrives at the following conclusions on the basis of which an evaluation of the merits of the three proposed diachronic routes should now be possible: 1. Ohala's model of sound change is a reliable and constrained predictor of possible sound change. 2. The traditional reconstruction of the PIE voiced aspirates as breathy voiced stops is phonetically the most plausible option. 3. There now exists a body of reliable and independent information in the form of phonetically-predicted changes supported by parallel developments against which the routes for the Italic development may be assessed.
In the sixth and final chapter, JSS sets out "to fulfill the explicit task of the book, namely to provide the most plausible reconstruction for the stages between the PIE voiced aspirates and the Italic reflexes" (p. 226). Her first conclusion is that Rix's account is the least phonetically plausible, mainly because it presupposes a word-initial shift of *bh to *B which she estimates to be unlikely on the basis of phonetic predictions and parallel developments. Of the remaining two scenarios, Ascoli's is deemed to be less plausible than the third option which alone assumes that word-initial and word- final voiced aspirates went their separate ways from the outset. This is because "[g]eneral observations of the synchronic and diachronic behaviour of stops often demonstrate predictable differences between word-initial and word-internal position'" (p. 198).
As a longtime proponent and exponent of the importance of naturalness in sound change, I cannot but applaud the author's efforts to reconstitute the relative chronology of such a complex set of sound correspondences. Too often, such correspondences are taken to be the sound changes themselves by historical phonologists who seem to lack any basic methodology for distinguishing the two. As I have tried to show over the course of many years (see Picard 1994 and references therein), any attempted reconstitution of the phonological history of a language, be it in whole or in part, must be guided by a fundamental principle which is that of the minimality of phonological change. This major constraint on the nature of sound change, which was first put forth by Donegan & Stampe, states that "processes represent responses to phonetic difficulties . . . and each process makes substitutions by altering a single phonetic property to remedy the difficulty. Since the substituted sound should, in each case, be as perceptually similar to the original target as possible, it follows that the processes make will be minimal" (179:136-7). More succinctly, the claim is essentially that "it is usual for change to proceed in small steps which involve the alternation of only one feature at a time" (Bynon 1977:86).
The application of this universal principle to any set of correspondences involving obstruents, as in the case at hand, will ensure first and foremost that any change that affects this class of segments can never simultaneously involve more than one one of its three major phonetic properties, namely voicing, manner of articulation, and point of articulation. On this score, the third scenario - the one proposed by JSS - fares very well since the pathways she proposes all conform to this constraint, albeit with one minor adjustment. Thus, the shifts *bh- > *ph- (voicing), *-bh- > *-B- (manner) and *-B- > /b/ all involve a single phonetic property, and although *ph- > *f- changes both voicing and point, this can easily be remedied by positing the well-attested (but often transitory) intermediate stage /P/, i.e., a voiceless bilabial fricative.
Although the author's failure to give heed to these types of general conditions on sound change has no serious negative effects on her own analysis, it does lead her to posit a faulty assessment of the other two scenarios. In ranking Rix's proposal below Ascoli's in terms of phonetic plausibility, she overlooks the fact that the latter has *f going to *-B-, a shift that must be considered implausible not only because it involves two features (voicing, point) but also because it involves a strident fricative going to a non-strident fricative, something that violates another general constraint on phonological change. Admittedly, Rix does postulate the opposite shift of *-B- to *f- but since an intermediate /v/ would be a most natural intermediate stage, his analysis must be deemed at least as plausible as hers in this regard. Other factors, which I cannot hope to examine here, would have to be invoked in order to resolve this question.
In sum, although the sheer exhaustiveness and consummate professionalism of this study are highly commendable, the methodology JSS employs is not one that promises to travel well, especially when one has to deal with languages for which written records are unavailable. Decisions about whether to posit this or that particular intermediate sound change, which ultimately determine the plausibility of a reconstructed diachronic scenario, should not be made at a local level, as was done in the case at hand, but rather should rest primarily on pre-established conditions and constraints on phonological change.
Bynon, Theodora (1977) Historical Linguistics. Cambridge: Cambridge University Press.
Donegan, Patricia & David Stampe (1979) "The study of natural phonology". In Current Approaches to Phonological Theory, Daniel Dinnsen, ed. Bloomington: IndianaUniversity Press, pp. 126-73.
Picard, Marc (1994): Principles and Methods in Historical Phonology: From Proto-Algonkian to Arapaho. Montreal and Kingston: McGill- Queen's University Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Marc Picard teaches phonetics, phonology and written communication at Concordia University in Montreal. He has published extensively on synchronic and diachronic phonology and morphology, particularly in the areas of Romance and Algonkian languages.