LINGUIST List 26.3057

Fri Jun 26 2015

Review: Historical Ling; Phonology: Cooper (2014)

Editor for this issue: Sara Couture <saralinguistlist.org>


Date: 27-Mar-2015
From: Nicholas Zair <naz21cam.ac.uk>
Subject: Reconciling Indo-European Syllabification
E-mail this message to a friend

Discuss this message

Book announced at http://linguistlist.org/issues/25/25-4187.html

AUTHOR: Adam Cooper
TITLE: Reconciling Indo-European Syllabification
SERIES TITLE: Brill's Studies in Indo-European Languages and Linguistics
PUBLISHER: Brill
YEAR: 2014

REVIEWER: Nicholas Zair, University of Cambridge

Review's Editor: Helen Aristar-Dry

SUMMARY

This monograph is an attempt to provide a model of the Proto-Indo-European
(PIE) system of syllable structure, using an optimality-theoretical framework.
It is aimed at both historical linguists and theoretical phonologists, After
an introduction providing an explanation of the aims and structure of the
book, and an explanation of the assumed theoretical basis, it is divided into
two parts. Part 1, Consonant Heterosyllabicity in Indo-European, attempts to
provide an analysis of Indo-European syllabification which will produce the
correct syllabification of sequences of consonants between vowels. In Chapter
2, Cooper assesses the evidence for the position of the syllable boundary
provided by Vedic Sanskrit, concluding that the sequence VCCV was syllabified
as VC.CV (where V stands for any vowel and C for any consonant), while
triconsonantal sequences were syllabified following a ‘strong’ variety of the
sonority sequencing principle, resulting in the syllabifications VRO.OV,
VO.ORV (where O stands for any obstruent, and R for any sonorant). Chapter 3
provides a formal analysis of these proposed syllabifications in optimality
theory. Chapter 4 examines the evidence for the position of the syllable
boundary in consonant sequences in Greek provided by the development of the
sequence *ty, in order to identify the syllabification of VRORV sequences,
about which the Vedic evidence did not prove conclusive. Cooper concludes that
the most plausible syllabification is VR.ORV. Chapter 5 discusses the apparent
syllabifications VOO.RV and VR.OOV, and the difficulties of generating these
within a coherent optimality theoretic model.

Part 2, Sonorant Vocalization in Proto-Indo-European, attempts to provide an
analysis of the apparently idiosyncratic behaviour of sonorant syllabicity in
Proto-Indo-European, whereby the right-most of two consecutive sonorants
between consonants is generally treated as vocalic, but with various
exceptions. Chapter 6 provides the background to this question and traditional
accounts, while Chapter 7 discusses previous attempts to model this process in
optimality theory. Chapter 8 provides a new model, which is consistent with
the constraints already proposed in Part 1. Chapter 9 proceeds to alter this
model to allow for the oddity whereby sonority is crucial in deciding which
segments are vocalised in the first place, but does not distinguish between
sonorants or differing sonorities (always preferring the rightmost). In
Chapter 10, Cooper discusses possible approaches to dealing with apparent
exceptions to the model proposed in the previous chapters and provides
typological parallels to the eventual model; while the failure to vocalise as
expected of the *n in the so-called ‘nasal-presents’ remains a problem,
Cooper’s suggestion that *m, which does not always vocalise when expected, was
less sonorous than the other sonorants, seems highly plausible.
Chapter 11 provides a brief summary conclusion, and possible directions for
future research.

EVALUATION

The book is well-written and produced: I noticed very few problems (one typo:
blockedz, on p.230; the formatting went mildly wrong in the tableaux on pages
175, 177, 179 and 181), other than my perplexity at the convention by which an
author refers to himself in the plural. In many ways, this book is an
exemplary case of how to combine phonological theory with comparative
linguistics. It shows a solid knowledge of Indo-European reconstruction with -
as far as I am able to tell - equal engagement with the literature on
optimality theory and other aspects of phonology. Particularly striking is its
clear emphasis on seeing Indo-European syllabification as a synchronic
phonological process, with a single model being proposed which represents a
Proto-Indo-European ‘grammar’. Cooper does not in general make any claims to
‘explain’ diachronic developments within different stages in the
proto-language or in the attested languages, thus avoiding the problems of
vacuity identified by McMahon (2003): differences in grammars over time can be
modelled in optimality theory by re-ranking, but this does not explain why the
ranking changed. Where Cooper does take actual historical developments into
account, this seems to me to lead to a misstep. In his discussion of the
perfect union vowel in Vedic in Chapter 2, where i appears between consonants
in some V:CCV --> V:CiCV and VCCCV --> VCCiCV sequences, Cooper discusses the
possibility that this should be seen as a process of deletion rather than
epenthesis, and observes that although “the fact that such an analysis would
not conform to our understanding of the historical developments involved is
arguably not enough to discount it outright, it nevertheless does raise some
doubt” (p.68). But in fact the creation of the perfect union vowel is at least
as much historically a process of deletion as epenthesis. It is clear that it
originally comes as the result of epenthesis of i which arose from most *CHC
sequences (where H stands for one of the so-called ‘laryngeals’, whose precise
phonetic status is somewhat unclear). In the perfect, roots ending in *VCH
followed by endings beginning with a consonant would have resulted by regular
sound change in VCiC sequences. Historically, therefore, we have both
‘deletion’ and ‘epenthesis’: epenthetic i was deleted in VCiC sequences, and
extended to VCCC and V:CC sequences which did not originally contain it.

The only real problem that arises when reading this book is the old question
of what sort of artefact PIE really is. Is it a ‘real’ language which is
susceptible to phonological theory of the type used here, or is it not?
Despite the caution of scholars such as Clackson (2007: 16-17),
Indo-Europeanists are, in the main, apt to treat PIE as if it were.
Furthermore, the comparative method can only get us so far, and as a result,
internal reconstruction is used to reconstruct sound changes which took place
within PIE itself (for example, the rules involving obstruents laid out in
Mayrhofer 1986: 109-18). But, as Lass (1975: 17) points out, internal
reconstruction “is of no independent value as a technique for the recovery of
linguistic history”; consequently, we may well be better off treating
explanations of these ‘sound changes’ as falling within the purview of
synchronic rather than diachronic phonology, i.e. representing inputs to and
outputs of a grammar rather than change over time. Nonetheless, difficulties
arise when it comes to deciding what exactly to posit as these inputs and
outputs, since PIE itself is not attested: in this way the proto-language is
different from a real language, where at least there is direct evidence of the
outputs. These problems seem to me to be particularly acute when dealing, as
here, with aspects of the phonology which do not come, in the main, with
phonetic correlates. With regard to the phonemes of PIE, for example, rules
and principles of varying sorts have been developed which allow for a
reasonably well-accepted idea of what the distinctive features of the phonemes
are: so, for example, most Indo-Europeanists would accept that e.g. *t could
be described as a voiceless coronal plosive - even though its reflex in some
languages might be voiced or a fricative (for attempts to elucidate some of
these principles see Lass 1993 and Kümmel 2007). But it is not clear to me
that we have such principles, at least stated explicitly, for determining how
the differing approaches to the position of the syllable boundary in the
different languages should be reconciled. In this book, Cooper wishes to
establish where the syllable boundary lay in sequences of two or more
consonants in PIE, but what he actually examines is where Vedic - and,
secondarily, Greek - put the syllable boundary in these sequences. How do
other languages treat these sequences, and, if they treat them differently,
how do we know which syllabification we should reconstruct for PIE? This is
not a problem which is unique to Cooper’s work, and one could argue that it is
not his duty to deal with a problem that Indo-Europeanists have not solved (to
the extent that some scholars do not think it is possible to reconstruct PIE
syllabification at all, e.g. Beekes 1988). But, in a book which aims “to
demonstrate the continued relevance of both Proto-Indo-European and the
ancient Indo-European languages to contemporary linguistic theory” (p.1), such
questions certainly require discussion, since theoreticians are unlikely to
consider PIE relevant to the questions at hand unless it can be shown to rest
on a reasonably firm footing that PIE was a real language, which actually
existed at some point in time, and that we can, in some reasonably principled
way, reconstruct how it syllabified particular sequences.

Particularly interesting from this point of view is the discussion in Chapter
5, in which Cooper attempts to formulate a system which will produce both the
syllabifications VOO.RV and VR.OOV in PIE. He comes up with two possible
approaches to produce the desired syllabifications, both of which involve the
positing of new and theoretically undersupported constraints, and concludes
that supporting their existence would require searching for parallels through
extensive typological work. But it would surely be much easier to convince the
scholarly community to carry out such research if it were an attested language
which were the basis for positing such novelties. And this is particularly the
case, because, as acknowledged by Cooper (p.130), much of the evidence for
both these proposed syllabifications is very scanty indeed (thus, for example,
the claim that Sievers’ Law, whereby glides (G) become vocalic after
super-heavy syllables, does not apply to VOO.GV sequences, on which see Barber
2013: 36-7, 30-32, 82-3). Much more interesting, to my mind, is his comment on
“the significant theoretical hurdles inherent in accepting the two treatments
V.ROOV, V.OO.RV as holding within one and the same language”. This failure to
model the supposed PIE situation within the theory might then be seen as a
useful hint that scholars should think again about such a reconstruction where
the evidence is not strong, providing a different sort of typological
guideline for reconstruction. I would have liked to have heard more about this
possibility.

REFERENCES

Barber, Peter (2013). Sievers’ Law and the History of Semivowel Syllabicity in
Indo-European and Ancient Greek. Oxford: Oxford University Press

Beekes, Robert S. P. (1988). Laryngeal developments: a survey. In Alfred
Bammesberger (ed.), Die Laryngaltheorie und die Rekonstruktion des
indogermanischen Laut- und Formensystems, 59-105. Heidelberg: Winter

Clackson, James (2007). Indo-European Linguistics. An Introduction. Cambridge:
Cambridge University Press

Kümmel, Martin Joachim (2007). Konsonantenwandel. Bausteine zu einer Typologie
des Lautwandels und ihre Konsequenzen für die vergleichende Rekonstruktion.
Wiesbaden: Dr Ludwig Reichert Verlag

Lass, Roger (1975). Internal reconstruction and generative phonology.
Transactions of the Philological Society 74, 1-26

Lass, Roger (1993). How real(ist) are reconstructions?. In Charles Jones
(ed.), Historical Linguistics: Problems and Perspectives, 156-89. New York:
Longman

Mayrhofer, Manfred (1986). Indogermanische Grammatik I/2: Lautlehre.
Heidelberg: C. Winter

McMahon, April (2003). On not explaining language change: Optimality Theory
and the Great Vowel Shift. In Raymond Hickey (ed.), Motives for Language
Change, 82-96. Cambridge: Cambridge University Press


ABOUT THE REVIEWER

Nicholas Zair is a Research Associate in the Faculty of Classics at Cambridge
University, working on the project 'Greek in Italy', which is funded by the
Arts and Humanities Research Council. He is interested in Proto-Indo-European
historical phonology and morphology, particularly with regard to the Italic
and Celtic language families. He is currently working on his second book, on
the orthography and phonology of Oscan, a language spoken in the south of
Italy in the first millennium BC.

Page Updated: 26-Jun-2015