Siptár, Péter and Mikl�s Törkenczy (2000) The Phonology of Hungarian.
Oxford University Press, hardback ISBN 0-19-823841-X, $99.00, xiv+319pp
(The Phonology of the World's Languages).
Charles Reiss, Concordia University, Montreal
This review follows up on Stefan Frisch's thorough review of this
book on the Linguist List
(http://linguistlist.org/issues/12/12-1866.html#1). I will give only the
briefest sketch of the book's contents and refer the reader to Frisch
for more detail. Instead I will concentrate on several points of
theoretical interest in Hungarian phonology and S+T's treatment of these
S+T have definitely succeeded in their stated purpose of providing an
accessible overview of the phonology of one of the few non-Indoeuropean
languages of Europe. The book benefits from their careful scholarship,
their expertise as linguists, and their native-speaker intimacy with
both the Educated Colloquial Hungarian and various dialect phenomena.
The first section, Background, consists of two chapters. Chapter 1,
Introduction, lays out the plan of the book, provides a sketch of
previous literature and statement of the framework and theoretical
assumptions adopted-a version of derivational phonology. Chapter 2,
contains an overview of Hungarian grammar, including interesting
phenomena such as the definite/indefinite conjugation distinction: the
former is used with direct objects that are third person and definite;
the latter is used elsewhere--with other objects and and for
intransitives. The occasional collapse of this contrast (in the first
person singular past and in the second person plural conditional) offers
a nice puzzle for morphologists. The chapter also has a guide to the
orthography, which is often used in phonological discussion instead of
The second section, Systems, contains three chapters. Chapter 3
discusses the Vowel System, including the vowel harmony for which
Hungarian is well-known. The Consonant System is the topic of Chapter 4,
including discussion of clusters and voicing assimilation processes.
Phonotactics: Syllable Structure is the title of Chapter 5, where the
patterns found within and across syllables are described.
The third section, Processes, builds on the earlier discussions of the
'systems'. Chapter 6, Processes Involving Vowels examines vowel harmony
and lengthening and shortening phenomena in greater depth. Chapter 7,
Processes Involving Consonants, treats phenomena such as palatalization
(lexical and postlexical) and assimilation and gemination rules
involving sibilants. Chapter 8, Processes Conditioned by Syllable
Structure, primarily handles vowel-zero alternations. The final Chapter
9 discusses Surface Processes, including fast speech effects. A list of
references and an index follow.
While quite modest in tone, the first paragraph of the book, in which
S+T state their decision to use a derivational approach rather than a
non-derivational model such as 'Declarative Phonology, Government
Phonology or Optimality Theory', could also be read as a challenge to
the proponents of such theories. S+T motivate their decision as a
reflection of their aim of providing an accessible description of
Hungarian phonology, rather than from 'theoretical preferences', but
they do point out that a `comprehensive analysis of the sound pattern of
a single language' is hard to come by in such frameworks--
non-derivational theories have yet produce a work to match
SPE in either the scope of theoretical problems and solutions it
presents, or in the descriptive breadth it attains. However, it must be
noted that S+T do not provide a straight 'rule-based' phonology--as is
common in the derivational literature, they appeal to constraints and
parameters as well as traditional rules.
Much of the material in the book assumes some degree of
morphology-phonology interaction. This is treated under the heading of
an 'analytic/synthetic' distinction in morphological domains which
parallels the distinction between + and #, or Level 1 and Level 2
affixation. While there may be evidence that such distinctions are
necessary in Hungarian, I would like to argue that S+T's presentation of
the distinction is not convincing in at least two cases.
S+T note that clusters such as -kp- occur across the boundary of a
compound word like _ker�kp�r_ 'bicycle'(_ker�k_ 'wheel' + _p�r_ 'pair').
However, such clusters never appear morpheme internally or between
morphemes that span an synthetic juncture. The absence of such clusters
internally or in close juncture is attributed to the existence of a
'transsyllabic constraint' (128). This transsyllabic constraint does not
apply across an compound boundary, so the cluster of __ker�kp�r__ can
surface intact (131). Here is the problem: the existence of the
constraint is not supported by any alternations--the cluster -kp- does
not occur within morphemes and it never arises by morpheme concatenation
other than compounding. The 'constraint' could never be violated since
no grammatical morphemes in the language begin with /p/. Therefore,
there is no reason whatsoever to posit a transsyllabic constraint against
-kp- since the nature of the Hungarian lexicon assures that it never
arises, except in compounds.
Once we reject the motivation for the transsyllabic constraint, we
cannot use its failure to be honored between compound members to
motivate a distinction of phonological domains. S+T state that ''it must
be pointed out that clusters that straddle analytic boundaries do not
reveal the constraints governing interconstituent [transsyllabic]
sequences'' (131). More accurately, clusters straddling such boundaries
provide evidence that the interconstituent constraints do not exist at
all. This somewhat negative view is supported by the fact that,
according to S+T, /kp/ is actually found in ten monomorphemic words
(page 129: Table 17).
What kind of evidence could be found for the existence of the
transsyllabic constraint? The use of loan 'phonology' is a common source
of external evidence. In the present case, if Hungarian speakers truly
had constraints against monomorphemic voiceless velar-labial clusters
they should, while pronouncing _ker�kp�r_ with ease, have difficulty
borrowing and pronouncing as a monomorphemic word the English
'backpack'. They do not. It should also be noted that S+T suggest that
there is a constraint against /kp/ clusters within a morpheme or across
a close juncture, but there is no constraint against -gb-.
S+T provide a descriptive overview of vowel harmony in Hungarian
(Chapter 3) as well as a novel analysis based on the Clements and Hume
(1995) feature system (Chapter 6). Following tradition in the field they
claim that vowel harmony also provides evidence for phonological
domains: ''All vowel harmony rules apply in Block 2 ... This means ...
that analytic suffixes undergo [vowel harmony] just like synthetic
suffixes do...The domain of application of vowel harmony is the
phonological word. In other words, it applies morpheme internally,
across a synthetic suffix boundary, and across an analytic suffix
boundary...but not across compound boundaries or word boundaries'' (158,
fn.3). It also does not occur between a prefix and root. Rather than
viewing such facts as evidence for phonological domains, it is probably
more economical to view it as a reflection of the fact that prefix and
root vowels are fully specified, whereas the harmonic suffix vowels
lexically underspecified (an assumption accepted by S+T).
Other arguments for phonological domains in Hungarian remain to be
examined. However, those based on vowel harmony and phonotactic
constraints are unconvincing, but unfortunately reflect widespread
practice in the literature. To conclude, this book makes available a
rich body of data on a fascinating language and it attempts some new
analyses. However, some traditional assumptions in phonological theory
generally, and Hungarian phonology specifically, could be treated more
Clements, G. N. and Elizabeth V. Hume. 1995. The Internal Organization
of Speech Sounds. In J. Goldsmith (ed.) The Handbook of Phonological
Theory. Oxford: Blackwell, 245-307.
[The reviewer submitted a small correction to this review in issue http://linguistlist.org/issues/12.2784.html. -- Eds.]
ABOUT THE REVIEWER
Charles Reiss is Associate Professor of Linguistics at Concordia
University in Montreal. He is interested in Phonology, Morphology,
Cognitive Science, Learnability Theory and Historical Linguistics.
He is co-organizer with Mark Hale of the Second North American Phonology