Publishing Partner: Cambridge University Press CUP Extra Wiley-Blackwell Publisher Login
amazon logo
More Info


New from Oxford University Press!

ad

Language Planning as a Sociolinguistic Experiment

By: Ernst Jahr

Provides richly detailed insight into the uniqueness of the Norwegian language development. Marks the 200th anniversary of the birth of the Norwegian nation following centuries of Danish rule


New from Cambridge University Press!

ad

Acquiring Phonology: A Cross-Generational Case-Study

By Neil Smith

The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.


New from Brill!

ad

Language Production and Interpretation: Linguistics meets Cognition

By Henk Zeevat

The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin


Email this page
E-mail this page

Review of  Language and Literary Structure


Reviewer: Lev Blumenfeld
Book Title: Language and Literary Structure
Book Author: Nigel Fabb
Publisher: Cambridge University Press
Linguistic Field(s): Ling & Literature
Subject Language(s): English
Book Announcement: 14.3536

Buy
Discuss this Review
Help on Posting
Review:
Fabb, Nigel (2002) Language and Literary Structure: The Linguistic
Analysis of Form in Verse and Narrative, Cambridge University Press.

Announced at http://linguistlist.org/issues/13/13-2396.html

Lev Blumenfeld, Stanford University.


SUMMARY

In _Language and literary structure_ Nigel Fabb undertakes a detailed
exposition of a generative approach to the study of literary
form. Throughout the book, Fabb defends the basic thesis that there
are three kinds of literary form: explicit form, which has to do with
relatively overt properties of particular performances of literary
texts; implied form, which holds of a text as by virtue of a reader's
inference about the text; and generated metrical form, which is
derived from the linguistic representation of the text by an
autonomous system of rules. Crucially, for Fabb, invariant facts about
form are treated by the generative component, while gradient
properties which can hold to a degree are treated as implied form.

In Chapter 1, Fabb outlines his views on literary form, introducing
the distinction between the three types of form and showing the basic
analysis of the English iambic pentameter and other meters, including
the complexities involved in non-projection of syllables and in
resolution. Fabb shows which rules are needed to derive the generated
metrical form, and which correspondence rules between the linguistic
representation and the generated metrical form are needed to
characterize metricality. Throughout, Fabb emphasizes the crucial
distinction between linguistic and metrical form, arguing that
accounting for metricality amounts to stating the correspondence
between the two representations and not merely constraints on the
linguistic form. Implicit in Fabb's exposition is a typology of
possible meters, since metrical rules and rules of correspondence can
formally vary only to a limited extent.

Chapter 2 continues the discussion of complexity in generated metrical
form. The analysis of trochaic inversion is shown to make a crucial
distinction between inversion involving polysyllables and monosyllabic
inversion which imitates the initial rhythm of polysyllabic trochaic
inversion. Fabb then goes on to discuss non-projection rules and the
complexity they impart to literary form. The chapter concludes with
two arguments in favor of the approach proposed by Fabb, as opposed to
other accounts of meter that take ''rhythm ... to be at heart of the
meter, rather than counting'' (p. 54). First, Fabb shows that counting
out the positions in a line, which is the fundamental task of rules
deriving generated metrical form, makes two predictions about possible
meters: that there should not be meters with regular rhythm but
irregular line length, and that there should not be systematically
rhythmic prose. Fabb's second argument in favor of his approach has to
do with accounting for aperiodic meters which, he argues, only his
counting-based approach can treat.

Chapter 3 moves from rigid aspects of meter accounted for by the
generative metrical component to gradient properties of meter, which,
Fabb argues, constitute 'communicated form' in the sense that these
aspects hold not as invariant properties of texts but as inferred
thoughts about texts. Fabb claims that communicated properties of
texts, such as 'being a sonnet', do not hold invariantly but only to a
degree, because they are inferred from a large set of potentially
conflicting conditionals of varying strength. Properties such as
'being a sonnet' only hold insofar as a reader derives them as
thoughts from conditionals relating different aspects of form to each
other, such as 'being a sonnet', 'having fourteen lines', and 'being
written in iambic pentameter'.

Crucially, Fabb takes gradience to be the defining property that
distinguishes communicated form from generated form. Other
properties, such as accessibility to introspection and unrestricted
formulation, are claimed to correlated with gradience. Thus,
sonnethood is gradient in that evidence for it may be incomplete or
contradictory, it is available to conscious introspection, and it is
formulated in unrestricted, ordinary-language terms. On the other
hand, the property of having ten projected syllables is absolute,
generally inaccessible (at least not easily) to introspection, and
formulated in terms specific to the metrical module (p. 59-60). This
leads Fabb to call the former 'communicated form' and assign it to the
pragmatics, and the latter 'generated metrical form' and assign it to
the metrical module of grammar. Fabb then illustrates in detail how
Relevance theory is applied to deriving inferences about communicated
form, emphasizing the ways in which complexity derived from
conflicting or weak implicatures is a source of aesthetic pleasure.

Chapter 4 moves from a general argument for communicated form to an
application of the theory to a more complex case, communicated
meter. Fabb argues that the explanatory gaps that remain from the
theory of generated metrical form are covered assuming that metrical
form can also be a kind of communicated form. In keeping with the
general assumption that gradience is the key property that separates
the two components, Fabb discusses gradient aspects of iambic
pentameter, such as the relative similarity of lines to the basic
weak-strong rhythm. Fabb claims that metrical form, in addition to
being generated by metrical rules, is also inferred by a pragmatic
mechanism similar to the one that accounts for properties like
'sonnethood'. Trochaic inversion involving monosyllables, which
remained outside the scope of the theory in Chapter 2, is accounted
for by means of conditionals that relate surface rhythms of lines to
their metrical forms. Importantly, Fabb argues that there is no
relation between generated and communicated meter, and thus the two
independent aspects of form exist alongside each other and bring
complexity to verse that contributes to aesthetic experience. The
chapter moves on to discuss Fabb's insight with respect to metrical
ambiguities and metrical imitation. Fabb concludes with a brief
discussion of 'metrical tension', arguing for an inference-based
approach to the conflict between explicit and implied form present in
metrical tension.

Expanding his Relevance-based account of communicated form, Fabb
applies it to lineation in Chapter 5. He argues that line boundaries
are not absolute or invariant properties of texts but hold only as
inferences about texts, and thus lineation is a kind of implied rather
than generated form. Fabb argues that lineation shows all of the
properties of communicated form, such as gradience, conflicting and
weak inferences, &c. Among other conditionals, Fabb argues that those
pertaining to statistical preference for word placement in the
beginning and end of lines constitute weak evidence for lineation. The
interaction of line boundaries with meter and other aspects of form
such as rhyme receive detailed attention, as do lineation
ambiguities. The Chapter closes with a discussion of asymmetries in
strictness between the beginnings and ends of lines. Because the
asymmetries typically involve constraints on specific positions within
the line, Fabb argues that they are best accounted for as part of the
metrical rules, but, as all metrical rules, they may be supplemented
by a pragmatic account.

Chapter 6 moves to yet larger constituents, line groups. Fabb uses by
now familiar arguments in favor of treating line grouping as
communicated form, not generated form. In his account of rhyme
schemes, however, Fabb proposes to extend the theory of counting and
grids to line groups, where grids are constructed on asterisks
projected by individual lines and rhyme scheme requirements are then
stated on the grids thus constructed. The chapter concludes with a
brief discussion of line grouping and other aspects of structure in
oral narrative.

The brief concluding Chapter 7 brings the arguments in the book
together by suggesting that the various types of complexity due to the
ways in which ''a text is thought'' -- tension between different
representations, ambiguity, conflicting conditionals, weak
implicatures -- contribute to aesthetic experience, since verbal art
''exploits to the full every option for making verbal behaviour
difficult'' (p. 217).


EVALUATION

1. TRIVIAL MATTERS
Fabb presents his argument with remarkable clarity. The Chapters of
the book are carefully organized and supplied with helpful
summaries. The theoretical points are richly illustrated with examples
from mostly, but not exclusively, English neo-classical verse. My only
complaint on presentation has to do with the unnecessary and cluttered
repetition of explicit modus ponens inferences throughout the book:
the steps are simple enough to be omitted after a few introductory
examples.

2. GRADIENCE OR NATURALNESS?
Fabb's work is an eloquent and convincing defense of a generative
approach to literary form. This approach can be defined by two
important aims: first, the explanation of literary form is modular in
the sense that not all aspects of a work are to be explained at the
same time, and verbal art owes its complexity, in part, to the
multiplicity of modules required for its processing. Second, the
general research project is, as in generative grammar, to categorize
limitations on diversity of literary forms across traditions and time
periods, and to appeal to the properties of the human language faculty
for an explanation of gaps in the data. Fabb's theory of poetics is an
integral part of linguistics, and insightfully shows how the data and
theory of each field can inform the other.

I will now discuss some specific proposals made in the book, focusing
on Fabb's distinction between generated and communicated form and its
implications for imperfect rhyme, gradient metricality, and lineation.

Fabb's choice of gradience as the criterial property separating
aspects of literary form accounted for by a modular system of rules
from those accounted for by pragmatic inference mechanisms stems from
assumptions about grammars of natural languages that exclude
non-categorical phenomena. Adopting a different view of grammar, such
as Optimality Theory, which makes no distinction between categorical
and gradient phenomena and treats grammaticality as a matter of
degree, would have led Fabb to a radically different view of literary
form as well. Such an OT-based view of metrics, espoused, for
example, by Golston (1998), Hayes (to appear), Friedberg (2002) and
Kiparsky (2003), leads one to treat metricality of lines in terms of
their similarity (or faithfulness) to a metrical template, where
greater deviations produce greater complexity of a line, and absolute
metricality is defined simply as a cutoff point on a gradient of
violations. What is meant by 'closeness' and 'deviation' can be made
precise in the formulation of the Optimality-theoretic constraints.

Although the choice between different views of literary form hinges on
one's view of grammar, it is possible to distinguish the two theories
on grounds internal to poetics. While accepting Fabb's bimodular view
of literary form, with some aspects of it treated 'in the grammar' and
others 'in the pragmatics', I will argue that Fabb's choice of
gradience as the criterial property is incorrect: there are gradient
aspects of meter that behave like aspects of grammar. (The converse is
not true: due to the nature of the inferential process, no aspects of
form that belong to the pragmatic component are rigid in the sense of
being all-or-none properties.)

Instead, I believe the correct criterion for assigning aspects of
literary form to one component or another is NATURALNESS
vs. ARTIFICIALITY of literary form.

The study of literary form from a perspective that emphasizes its
groundedness in the human language faculty must face a recurring
difficulty: unlike properties of grammars of natural languages,
properties of 'poetic grammars' are subject to conscious control and
manipulation on the part of the artist. The reason for this is, in
part, that natural grammars are shared by communities of speakers and
constrained by functional pressures of communicative efficiency, so
any unilateral manipulation on the part of an individual speaker would
jeopardize the grammar's function. This is why language change is slow
and gradual. Literary grammars, on the other hand, belong as much to
individuals as to communities, and, crucially, the functional
communicative pressures do not apply to literary form. If nothing
else, an opposite pressure to stand out from one's peers and
predecessors pushes the artist to manipulate his 'grammar' in a
self-conscious way, at least in some traditions. (This point by itself
highlights the importance of poetry for linguistics: because
communicative constraints are largely inapplicable, certain aspects of
the language faculty that are invisible in natural language are
exposed in literary form. For example, rhyme and alliteration are
marginal in language (cf. Jakobson 1960, Zuraw 2002), but hold full
sway in poetry.)

Furthermore, institutionalized learning of literature explicitly
trains artists to pay attention to form, creating literary traditions
based on the conscious molding of formal constraints. Similarly,
prescriptive literary norms, much like linguistic prescriptivism, can
have the effect of sustaining linguistically unnatural structural
features.* The fact that it is most often adults past the critical age
who learn poetic grammars -- 'speakers' of literary grammars are
typically 'L2 speakers' -- further exposes formal constraints to
conscious sculpting.

This manipulability of literary form in part undercuts its rootedness
in the language faculty, for any ostensible universal proposed today
may be subverted, consciously and deliberately, by an adventurous poet
tomorrow. For example, in the face of claims that rhyming strings must
be prosodic constituents, I could write and, with enough funds and
good luck, publish, a whole book of poems where the rhyming scheme
calls for the identity of the last two segments of a line -- an
allegedly impossible system. It is easy to cite real examples from
history: the medieval Russian poet Smirnitsky who borrowed the
orthographic Greek distinction between long and short vowels, creating
a monstrously unnatural system doomed to failure (Gasparov 1984:
28-29), or modern poets like Dylan Thomas writing syllabic verse with
a large line length (e.g. 17 syllables) without any hint of internal
constituency. Orthography-based forms such as acrostics, popular in
many traditions and time periods, also belong to this category.

I believe, however, that although these problems may raise serious
methodological concerns, they do not challenge the theoretical claims
of Fabb's or any other mentalist approach to literary form: there
still is a crucial difference between, on the one hand, natural,
grounded aspects of form that are acquired and applied unconsciously,
and are handled by a module of grammar, and, on the other hand,
artificial aspects of form that must be communicated and understood
consciously and explicitly (Kiparsky 1987; Hanson 1991; Hanson and
Kiparsky 1996). In other words, despite all the obscurities, there
still is a difference between 'generated form' and 'communicated form'
in Fabb's sense.

In other words, my main objection to Fabb's proposal is that the right
criterion separating the two types of form is not gradience at
all^×there is no logical connection between artificiality and
gradience. Rather, to claim that an aspect of form is handled by
pragmatics is to claim that it is not NATURAL FORM, i.e., grammar has
no tools to deal with it.

3. ILLUSTRATION
Let me turn for illustration to imperfect rhyme. This is a topic Fabb
devotes little attention to in the book, but he makes clear how its
analysis might proceed in the approach he defends. Because the
identity of the phonetic material involved in imperfect rhymes is
partial, for Fabb, deciding whether two words rhyme is a matter of
judgment and inference, of explicitly comparing two 'explicatures'
(representations derived from particular performances of literary
texts) (p.67). Rhyme then is not an inherent property of a text, but a
matter of 'how the text is thought'. On the other hand, 'perfect'
rhyme where the identity of the rhyming constituents is complete is a
matter of grammar.

I believe that imperfect rhyme must be treated in the same component
as full rhyme: in the grammar. First, it is well- known that perfect
rhyme is constrained by the language faculty in a specific way:
prosodic constituent structure limits potential rhyming
constituents. For example, there are systems where the identity of the
nucleus and coda of the final syllable is required. Other systems
require to rhyme only the nucleus (assonance), only the coda (slant
rhyme), or only the onset (alliteration). More complex systems such as
the classical Russian rhyme can also be analyzed in terms of prosodic
constituency. Here, the nucleus and the coda are required to be
identical, but if the line-final syllable is open, then the onsets
also must match. Thus what rhymes is the minimal branching constituent
containing the nucleus: the nucleus and coda if there is one, and
otherwise the whole syllable. However, there are no systems where,
e.g., the last three segments of the line are required to be
identical, regardless of their constituent status.

Crucially, the same constraints apply to imperfect rhyme: partial
identity is also required of units in a way that is constrained by
prosodic constituent structure. This fact is a surprise given Fabb's
proposal to treat imperfect rhyme outside of the phonological
component. If all there is to judging phonetic similarity is pragmatic
inference, there is no reason to expect abstract phonological
properties such a syllable constituency to play any role.

The second argument against separating perfect and imperfect rhyme
into different components is that some recent approaches to
phonological theory have argued for the synchronic theoretical status
of phonetic similarity IN THE GRAMMAR. Most radically, this is the
approach of Steriade's P-map, which explicitly relies on phonetic
similarity as a grounding for the choice of a phonological repair for
marked structures. Also, the standard ideas about correspondence and
faithfulness in OT are a formalization of partial similarity and
degrees of similarity in the synchronic grammar. As correspondence
constraints compute degrees of deviation between levels of
representation, imperfect rhyme can be analyzed with correspondence
constraints to determine the degrees of acceptable deviation. An
immediate empirical consequence of this approach (but not of Fabb's)
is that the types of similarity relevant in imperfect rhyme are the
same types that play a role in correspondence elsewhere in grammars:
they should be formulable in terms of phonological features. If one
subscribes to Steriade's views, stronger consequences can be derived:
the same fixed similarity relations, encoded in the P-map, should
apply in the grammar and in imperfect rhyme.

Furthermore, the claim that imperfect rhyme is inferred is
inconsistent with Fabb's assertion (pp. 59-60) that communicated form
should have the properties characteristic of pragmatic aspects of
form: accessibility to introspection and formulability in natural
language terms (rather than in terms specific to the phonological
module of grammar). Much like metrical rules, detailed properties of
imperfect rhyme are only now being discovered with the tools of
linguistics (see Zwicky's (1976) work on rhyme in rock lyrics and
recent work by Steriade), indicating that they are far from conscious
introspection. And in each case, the constraints are formulable in
purely phonological terms.

I conclude from this that there is no reason to separate perfect and
partial rhyme into two components and no reason to treat the latter as
outside of grammar, and that these facts are fully consistent with
treating rhyming correspondence as a gradient of similarity, with
different poets and traditions picking a different cutoff.

Now I turn to similar issues surrounding what Fabb calls 'communicated
meter', the subject of Chapter 4 of the book. Fabb's main argument in
favor of duplicating the grammatical metrical constraints in the
pragmatics is that even if two lines fare equally well on the
grammatical constraints, the line whose rhythm is more similar to the
rhythm of the metrical template may be preferred. Perhaps the most
striking example of this phenomenon is Fabb's important finding cited
on p. 39, that monosyllabic trochaic inversion, which is not regulated
by the metrical constraint formulated on 'lexical' stress (the main
stress of polysyllabic words; see p. 38-39), gravitates toward the
same positions in the line that admit polysyllabic trochaic inversion:
the first foot. To account for this remarkable coincidence of the
gradient and rigid aspects of meter, Fabb proposes that the similarity
of lines to a template be handled by the pragmatic component.

An immediate prediction of Fabb's theory that monosyllabic trochaic
inversion 'imitates' polysyllabic inversion is that the former should
only occur in systems where the latter is present. This prediction is
false: in standard Russian iambics, the prohibition against lexical
stress in weak positions is absolute and extends to line-initial feet,
i.e. there is no polysyllabic trochaic inversion. However,
monosyllabic inversion is common and statistically gravitates toward
the initial and post-pausal position in the line (Taranovsky 1971:
420). This is consistent with a view that treats metricality as a
cutoff point, chosen differently by different poets and traditions, on
a gradient given by the linguistic system.

Let me now turn to lineation, a topic where Fabb's claims are,
perhaps, most controversial. Recall that Fabb treats line boundaries
as a kind of communicated form, negotiated in the pragmatic
component. There is no doubt that lineation can be subject to a great
deal of subversion and play on the part of an author, as documented by
Fabb in Chapter 5 and especially on pp. 143-144. He lists parallelism,
alliteration, rhyme, meter, linguistic boundaries, and line length as
sources of possibly conflicting evidence for lines, as well as some
non- linguistic sources of evidence such as orthographic patterns in
acrostics and typographical layout. Authors play with line boundaries
more than with other aspects of form (e.g. more than with meter),
perhaps, because lines are so conspicuous in the typographical
appearance of a text and can thus easily enter into an artist's
conscious manipulation. However, my position differs from Fabb's in
that this manipulability by itself is not sufficient evidence to deny
a purely linguistic basis to larger constituents such as lines. In
fact, I argue below that line boundaries are just as rigid as the
invariant aspects of meter, but for independent reasons are more
easily obscured by artificial devices.

There is an important paradox to which Fabb alludes on pp. 141-142:
pragmatically, and hence incompletely, determined line boundaries
serve as domains for purely grammatical metrical rules. Rejecting on
general grounds the solution of ordering metrical component after the
pragmatics, Fabb proposes that the two components be kept separate,
but correct scansion, relative to line boundaries, is achieved by
trying all possible derivations: ''[A]ll options are simultaneously
pursued, with scansions rejected when they fail, until the text can be
coherently scanned'' (p. 142).

This approach is squarely inconsistent with Fabb's view of line
boundaries as purely communicated form. Generating a scansion involves
deriving the metrical grid and assigning constituent structure to it,
through bracketing rules or whatever other formalism one wishes to
apply. The metrical grid crucially 'counts out the line' (see
pp. 53-54), i.e., DETERMINES line boundaries. Thus, even if the system
is constructed to prevent the metrical rules from being sensitive to
inferred line boundaries, line boundaries must necessarily be present
in the OUTPUT of the metrical rules: you cannot have one without the
other. Thus even on Fabb's account lineation turns out to be not only
communicated form but also generated form.

If this view is correct, then any metrical form (but not free verse)
must have determinate lineation. This lineation determines the domain
of metricality constraints, as well as any other metrical rules or
constraints sensitive to line boundaries, such as 'line-initial
looseness' or 'line- final strictness'. As with other aspects of form,
a poet may or may not choose to subvert this grammatical constituent
structure by introducing conflicting evidence from any of the sources
Fabb mentions, but this does not undercut the presence of line
boundaries in the generated metrical form. An immediate consequence is
that meter can only be sensitive to the generated and not 'artificial'
line boundaries: e.g., there should be no line-initial metrical
looseness based on purely typographic aspects of lineation not
corroborated by the meter.

The arguments given so far apply just as well to constituents above
and below the line: cola and stanzas. There is always the difficulty,
though, that the larger the constituent, the more amenable it is to
manipulation.

What I have said so far does not deny the existence of purely
grammatical ambiguities in lineation and line grouping, but they
appear to be of a very limited type. One convincing example of a
non-pragmatic ambiguity in line grouping, in my opinion, is the
structure of the Italian terzina, whose rhyming scheme ABA BCB CDC
&c. suggests that beginning with line 4, each third line doubles as
the last line of the previous quatrain and the first line of the
following one, illustrated below (where the brackets above and below
the string of lines indicate overlapping constituent structure.)

(1) [ Q1 ] [ Q3 ]
A B A B C B C D C D E D E
[ Q2 ] [ Q4 ]

This chronic 'ambistanzaicity' strictly parallels what is
grammatically necessary anyway for smaller units -- ambisyllabicity
(e.g. Myers 1987) and the recently proposed 'ambipodicity' (Hyde 2002)
-- and is of a very different character than the ambiguities listed by
Fabb, because grammar has the tools to deal with it.


CONCLUSION

Let me reiterate the main points of my discussion. I have argued that
gradience is not criterial in separating generated from implied
form. Rather, aspects of form rooted in the language faculty are
processed in a modularized component, while artificial aspects of form
are processed in an inferential-pragmatic way.

In support of this view I have argued that such non-rigid properties
as imperfect rhyme and gradient metricality belong IN THE
GRAMMAR. Also, despite the fact that evidence for line boundaries is
often conflicting, there is no reason to think that they are not
present in the output of the generative metrical component.

Although I disagree with some of the proposals put forth in _Language
and literary structure_, I would argue in favor of Fabb's general
approach: first, the modular treatment of literary form, and second,
the project of explaining limitations on literary diversity by
appealing to the human language faculty. Fabb's work shows that the
very possibility of gaining insight about literary form INDEPENDENTLY
of the content of a text is already in itself a vindication of a
mentalist approach to the study of verbal art.


NOTE

*I am grateful to Boris Maslov for bringing this point to
my attention.


REFERENCES

Friedberg, Nila (2002). Metrical complexity in Russian verse: A study
of form and meaning. PhD Thesis, University of Toronto.

Gasparov, M.L. (1984). Ocherk istorii russkogo stikha. Moscow: Nauka.

Golston, Chris (1998). Constraint-based metrics. NLLT 16: 719-770.

Hayes, Bruce (to appear). Faithfulness and componentiality in
metrics. To appear in Nature of the word: essays in honor of Paul
Kiparsky, ed. by Kristin Hanson and Sharon Inkelas.

Hyde, Brett (2003). A restrictive theory of metrical stress. Phonology
19: 313-359.

Jakobson, Roman (1960). Linguistics and poetics. Style in language,
ed. by T.A. Sebeok, 130-144.

Kiparsky, Paul (1987). On theory and interpretation. Linguistics of
writing, ed. by D. Attridge et al. NY: Methuen.

Kiparsky, Paul (2003). A modular metrics for folk verse. Ms.,
Stanford University.

Myers, Scott (1987). Vowel shortening in English. NLLT 5: 485-518.

Taranovsky, K.F. (1971). O ritmicheskoj strukture russkih dvuslozhnyh
razmerov. Poetika i stilistika russkoj literatury, ed. by
M.P. Alekseev et al., 420-429. Leningrad: Nauka.

Zhirmunsky, V.M. (1923). Rifma, ee istoriia i teoriia. Saint
Petersburg: Academia.

Zuraw, Kie (2002). Aggressive reduplication. Phonology 19: 395-439.

Zwicky, Arnold (1976). Well, this rock-and-roll has got to stop.
Junior's head is hard as rock. CLS 12.
 
ABOUT THE REVIEWER:
ABOUT THE REVIEWER


Lev Blumenfeld is a third-year graduate student at the
Linguistics Department in Stanford University. He is
interested in phonological theory and metrics.


Versions:
Format: Hardback
ISBN: 0521792940
ISBN-13: N/A
Pages: 240
Prices: U.S. $ 65
U.K. £ 47.5