LI06CH02_Kiparsky ARjats.

cls November 26, 2019 9:2

Annual Review of Linguistics

Metered Verse
Paul Kiparsky
Department of Linguistics, Stanford University, Stanford, California 94305, USA;
email: [email protected]
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Annu. Rev. Linguist. 2020. 6:25–44 Keywords

First published as a Review in Advance on
meter, verse, rhythm, stress, prosody, text-setting
November 5, 2019

The Annual Review of Linguistics is online at Abstract

linguistics.annualreviews.org
Generative metrics studies versification as a stylization of phonological form.
https://doi.org/10.1146/annurev-linguistics-011619-
This review article outlines its main goals, hypotheses, and findings and
030620
presents a template-matching version of it, which models meters as map-
Copyright © 2020 by Annual Reviews.
pings of abstract verse patterns to their permissible linguistic instantiations.
All rights reserved
It lays out the key features that verse rhythm shares with rhythm in other
cognitive domains and distinguishes it from biological rhythms, offering ev-
idence suggesting that these features are rooted in the language faculty. Af-
ter a review of the typology of stress-, weight-, and tone-based meters, the
predictions of the theory are illustrated in more detail with an analysis of
Shakespeare’s blank verse. The theory is extended by modeling conventions
of setting poems to music as an interface between composition and deliv-
ery. English text-setting, which privileges natural phonological stress and
phrasing, even at the expense of the poem’s meter, lineation, and caesuras,
is contrasted with other traditions in which text-setting is more faithful to
meter. The negotiation of phonology and meter in song provides a sensitive
probe into the prosodic organization of language.

25
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

1. RHYTHM AND STRESS

Meter has been called the heartbeat of poetry. But like language itself, and music and dance, it
pulsates more intricately than anything in the biological or physical world. While each beat of a
healthy heart is identical to the next, meter is based on the regular alternation of prominent and
nonprominent beats (stressed/unstressed, strong/weak, downbeat/upbeat).1 The prominent beats
in turn alternate between more prominent and less prominent beats, and so on, up to a unique
culmination, generating a prominence hierarchy within a domain (Liberman & Prince 1977). In
addition, the prominence peaks at each level of such a hierarchy are grouped with preceding or fol-
lowing less prominent units into measures, usually resolvable into binary constituents in language
and meter, but often irreducibly ternary in music and dance. The mind imposes periodicity and
constituency even on objectively undifferentiated sequences of beats: We hear the “tick-tick-tick”
of a watch as a sequence of “tick-tock” units.
These aspects of rhythmic organization shared by meter and language—alternation, hierar-
chy, and grouping—can be represented by trees in which each nonterminal node has a unique
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

strong immediate constituent or, equivalently, by bracketed grids where relative prominence is
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

represented by column height (Prince 1983, Hayes 1989):

(1a)

(1b)

The tree representation (example 1a) and the bracketed grid representation (example 1b) con-
vey exactly the same information and are interconvertible. However, they differ somewhat in the
formal operations and constraints that can be perspicuously expressed in them.
Certain constituents in the hierarchy have substantive properties of their own. In phonology,
these constituents make up the prosodic hierarchy; its stylized counterpart in verse is the met-
rical hierarchy, which appears to be strictly binary and hence does not map onto the prosodic
one:

1 However, certain pathologies are manifested by binary rhythms between or within heartbeats, as in pulsus
alternans, an alternation between strong and weak heartbeats, and dicrotic pulse, with two beats per cardiac
cycle, the first on the systole, the second on the diastole. Even the single pulse of a normal heartbeat is actually
generated by two contractions—of the atria and ventricles, in that order—which in turn form part of a phys-
iologically complex cardiac cycle. Nature abounds in multiphasic events, but there seems to be no reason to
consider them to be alternations between prominent and nonprominent beats. In any case, nothing in nature
comes close to the depth of hierarchical rhythmic organization found in the cognitively structured domains
of dance, music, verse, and language itself.

26 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

(2a) Prosodic hierarchy (in language) (2b) Metrical hierarchy (in verse)
Poem
|
Utterance Quatrain
| |
Intonation group Couplet (distich)
| |
Phonological phrase Line
| |
Word Half-line (hemistich)
| |
Foot Dipod
| |
Syllable Foot
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

| |
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Mora Beat
The prosodic and metrical hierarchies are subject to Strict Layering, a violable constraint which
militates against skipping levels and against improper bracketing [Itô & Mester 2003 (1992)]:
(3) Strict Layering
(3a) A nonterminal unit of the prosodic hierarchy, Xp , is composed of one or more
complete units of the immediately lower category, Xp−1 .
(3b) A nonroot unit of a prosodic hierarchy, Xp , is completely contained in a unit of the
immediately superordinate category, Xp+1 .
A consequence of the grounding of meter in language is that lineation and length restrictions
are not defined by counting feet or any other units in example 2b—for phonology cannot count
past two. They emerge from constraints on the hierarchical rhythmic structure that versification
imposes on texts (Chen 1980, Kiparsky 2006). Accordingly, an iambic pentameter is not simply
a sequence of five iambs: It is built from cola that are themselves built from feet. Some of the
evidence that confirms this prediction is summarized in Section 3. It follows that the line has no
privileged status over the couplet and the stanza, which are often defined by additional metrical
constraints on lines, and that lineation is in part conventional, in that any place where an obligatory
major prosodic break in the verse regularly divides equivalent units may be considered a line break,
by convenience or tradition. 4-3-4-3 ballad quatrains can be printed as fourteener distichs and
vice versa, and 3-3-4-3 quatrains are interchangeable with poulter’s measure. Some editions of
the Kalevala print its 8-syllable parallel couplets as single 16-syllable lines, and nobody minds.
These considerations argue against the contrary view (Fabb & Halle 2008, p. 242) that lineation
and length restrictions are the primary property of verse, and that meter arises as a by-product of
counting syllables to fix the length of lines, with rhythm an epiphenomenal “property of the way
a sequence of words is read or performed.”
In addition to its constitutive structural properties of hierarchy and constituency, meter is en-
livened by excursions such as syncopation, skipped beats, doubled beats, and extrametrical beats,
followed by a return to the regularly alternating baseline rhythm. A heart that pulsates like verse
would have its owner sent, with sirens screaming, to the nearest cardiological ward.
But the complexity is regimented. A meter is defined by a set of constraints on the distribution
of prominence-defining phonological features and on the misalignment of linguistic constituents
(words, phrases, sentences) with metrical constituents (feet, dipods, cola, lines, stanzas). These

www.annualreviews.org • Metered Verse 27

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

constraints can depend on genre and period, and on the level of the hierarchy at which the mis-
match occurs. The iambic pentameter of Shakespeare’s dramatic verse offers an extraordinarily
abundant metrical palette (described in Section 3, below), that of his sonnets less so, but both
are metrically richer than Pope’s neoclassical heroic couplets, and even that tight form still offers
some options denied to most German or Russian verse.
In virtue of being subject to well-defined constraints, metrical verse is the simplest, unmarked
form of literary language—indeed, in many unwritten literatures, the first or only form. Prose is
a more complex form of literature [as already pointed out by Herder 1960 (1768); see Hanson &
Kiparsky 1997], distinguished by rhythmic variety and often by the avoidance of metrical cadences
(antimetricality; Borgeson et al. forthcoming).
Metrical verse, lyric as well as epic, is common in preliterate cultures.2 Even in written lit-
eratures, all metrical verse is in principle oral, for writing systems reflect prominence-defining
phonological features imperfectly, if at all, and the graphic properties of a text, including spacing
and line breaks, are not in themselves constitutive elements of verse but rather are cues to phono-
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

logical phrasing and lineation. A metrical system not only is a cultural artifact but also is molded
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

within the limits of cognitive and linguistic constraints by functional pressures, of which the most
important are that the metrical repertoire should be as expressive as necessary to accommodate
the language (fit) and that it should be as restrictive as possible within those bounds (interest)
(Hanson & Kiparsky 1996).
The goal of generative metrics is a “grammar” of meter, rather than a “dictionary.” It aims to
characterize metrical systems as points or regions in a typological space, with linguistic theory the
main theoretical model and explanatory engine. Its fundamental hypothesis is that the metrical
form of verse is a stylization of phonological form. Appropriately formalized, this hypothesis pre-
dicts a typology of possible verse meters which is more narrowly constrained than musical meters
in several respects. In particular, verse meter inherits from phonology a strict binarity. Ternary
measures, ubiquitous in music, are in verse constructed from binary ones by beat splitting. While
polymeters and nonisochronous (additive) meters such as 2-2-3-2-3 or 2-2-2-2-1 21 -1 12 are used in
musics of Africa and South Asia, they are not found not in their poetries (Deo 2007).
Another consequence of the grounding of meter in language is that lineation and length re-
strictions are defined not by counting feet or any other units in example 2b—for phonology cannot
count past two—but rather by their abstract hierarchical organization. As noted above, an iambic
pentameter is not merely a sequence of five iambs: It is built from cola that are themselves built
from feet. Some of the evidence that confirms this prediction is summarized in Section 3.
Several approaches to generative metrics are currently being explored (Blumenfeld 2015). One
line of investigation which has proved productive treats meter as a mapping of abstract verse pat-
terns (“templates”) to their permissible linguistic instantiations. Both templates and their instan-
tiations can be represented by bracketed grids or labeled trees, as in example 1. The abstract verse
patterns can be characterized by systems of ranked Optimality-Theoretic markedness constraints,
of the same type as the constraints that govern the prosody of language itself. Evidence is mount-
ing that they are very simple and uniform.3 All the complexity is located in the constraints that
delimit the permissible mismatches between abstract metrical patterns and the linguistic patterns
of prominence that realize them. These can be formalized by Optimality-Theoretic correspon-
dence constraints. The output of this constraint evaluation is a metrical analysis, which specifies

2 See Beissinger (2012) for a concise review of the major traditions of oral and oral-derived poetry, with
bibliography.
3 For instance, there is no such thing as “trochaic substitution” in iambs: An inverted initial foot is still an iamb,

its inversion licensed by the correspondence rules (Kiparsky 2007).

28 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

Syllabification,
Metrical patterns Correspondence
stress, weight,
(verse design) rules/constraints
phrasing...

Metrical analysis:
scansions, complexity
(verse instances)

Performance conventions
(recitation, text-setting)

Delivery instances
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Figure 1
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Structure of a metrical theory.

whether a given text instantiates (i.e., legitimately corresponds to) a given abstract metrical pattern
and, if so, how complex the instantiation is, as measured by the licensed violations it incurs.
Vocal performance in song, recitation, and drama can profitably be included in the theory,
since it is shaped both by meter and by linguistic form, and since prevailing styles of performance
historically influence meter. We can model conventions of text-setting as an interface between
composition and delivery. Figure 1 shows the structure of a metrical theory. The constraints
not only define the space of metrical variation but also predict the relative frequencies of the
permissible variant types (Kiparsky 2006).
According to Figure 1, composing and performing a song require matching three hierarchies
of alternating prominence: an intrinsic prosodic form assigned by the language’s phonology, inde-
pendently of how it is versified; a metrical parse, independent of how the text is set to music; and
a musical rhythm, independent of the words that may be sung to it. The metrical structure of a
text is invariant and does not change with the way it is set to music or performed. Each rhythmic
tier is representable by a hierarchical tree or grid structure but is subject to its own constraints.
The required correspondences and permissible mismatches between them are regulated by con-
ventions that evolve historically within limits grounded in the faculty of language. Predominant
metrical systems and recitation, singing, and text-setting practices in a poetic tradition are mutu-
ally accommodated, and in time mutually optimized. Performance can be “tilted” to reflect meter,
and metrical forms must be compatible with prevailing text-setting and recitation practices. For
example, the eighteenth-century practice of reciting verse by highlighting the meter made it im-
possible to read the work of poets like Wyatt and Donne in a natural way, and demanded poetry
that could be metrically parsed with a minimum of mismatches.
The model in Figure 1 allows for cases in which a literary tradition accesses a different gram-
mar than the ordinary language. For example, metrical practice can be based on the phonology of
an earlier stage of the language. In traditional French versification, consonants that are deleted in
word-final position count for purposes of rhyme, except that homorganic final voiced and voice-
less obstruents are treated as equivalent. For example, long and tronc rhyme, but neither of them
rhymes with rond or pont, which, however, rhyme with each other; none of them rhymes with son.
Phonemically and phonetically, all five words end alike: /tKõ/ [tKõ], /lõ/ [lõ], /mõ/ [mõ], /Kõ/ [Kõ],
/sõ/ [sõ]. Morphophonologically, they all end differently: {tronk}, {long}, {pont}, {rond}, {son}—the
consonant shows up before suffixes, as in tronquer, longue, ponter, ronde, sonner. The two rhyming

www.annualreviews.org • Metered Verse 29

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

pairs match correctly at the output of the lexical phonology, which Kiparsky (2018a) argues to
be a significant level or representation. By final devoicing, both long and tronc end in /-k/ and
both pont and rond end in /-t/. The evidence that final devoicing takes effect at the word level
whereas final deletion is postlexical is that, in the classical liaison system (now as old-fashioned
as the rhyming convention that reflects it), final voiced stops appear in devoiced form before a
following vocalic word in close contact, as in long hiver [lÕ.ki.vEK] ‘long winter’ or grand homme
[grã.tɔm] ‘great man’. Therefore, they must enter the postlexical phonology with the final conso-
nant present in devoiced form. In sum, traditional French versification conventions crucially refer
to the lexical representation that is computed by the word phonology and forms the input to the
sentence phonology:
(4) Underlying (morphophonemic representation): {tKonk}, {long}, {pont}, {Kond}, {son}
Lexical representation: /tKõk/, /lõk/, /põt/, /Kõt/, /son/
Phonetic (and structuralist phonemic) representation: [tKõ], [lõ], [põ], [Kõ], [sõ]
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

That such fine points of rhyming outlived the sixteenth-century pronunciation they reflected for
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

several centuries is presumably due to the fact that the living morphophonology of the language
kept them intelligible.

2. PROMINENCE
A meter may require prominence in Strong positions, or prohibit prominence in Weak positions,
or both. Prominence is expressed by stress, syllable weight, and/or pitch. None of these features
is intrinsically binary: Languages can have degrees of stress and weight, and distinctive levels and
contours of pitch. The continua are exploited to give texture to verse (Ryan 2017) and, indeed,
to literary prose. But remarkably, in categorical constraints they are all binarized—another
manifestation of the pervasive binarity of verse structure. Stress meters (syllabotonic meters)
work with a binary distinction between stress peaks and nonpeaks. It may be defined on different
units in the hierarchy (example 2), each potentially constrained by meter, but always binarily.
Quantitative meters, found in classical Greek, Arabic, Sanskrit, Persian, Urdu, Hausa, Somali, and
Hungarian, among other languages, establish a binary distinction between light syllables, with
one unit of length, or mora, and heavy syllables, with two or more moras. Tone-based meters,
such as Chinese regulated verse and Vietnamese Luc Bat, are based on a binary division between
the language’s lexical tone categories.
Some metrical systems involve interactions of two or even three prominence features: Latin
and Skaldic verse obey separate constraints on quantity and stress, and Finnish and Tamil impose
special weight conditions on stressed syllables (Ryan 2017). In the deseterac meter of Serbian epic
songs, stress, weight, and pitch accent all seem to be relevant (Zec 2009), again as binary features.
It appears that the prosodic features that play a role in a language’s verse are limited to those
which are active in its phonology, either phonemic (distinctive/contrastive) or conditioning
processes of word phonology (Kiparsky 2018a). For example, in the phonological systems of
Hindi-Urdu, classical Arabic, Finnish, Czech, Latin, and Hungarian, vowel length is distinctive
and primary word stress is predictable. Yet only the first two languages rely exclusively on syllable
weight in their versification; the others exploit both of these prosodic features to varying extents.
This is because word stress plays a crucial role in their respective lexical phonologies, and a
negligible one in those of Hindi-Urdu and Arabic. Turkish offers intralinguistic confirmation:
The Persian-derived classical meters of Ottoman Turkish are based on syllable weight, while the
popular modern meters are based on stress. Correspondingly, vowel length is distinctive only in
the Persian-derived (ultimately Arabic) vocabulary that the classical Turkish meters use. English,
German, and Swedish meters are almost exclusively stress based, though syllable weight interacts
30 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

with their lexical phonologies in ways that are reflected in some of their meters. Russian has only
stress, and consequently only stress-based meters, with syllable weight playing no discernible
role. Mordvin has neither lexical stress nor vowel quantity nor distinctive pitch, and its meters
are consequently syllable counting with variable line divisions as the principal rhythmic device
(and tend to compensate by rich nonmetrical devices such as rhyme, phrasing, and parallelism).
French stress is basically a phrase-level feature, and its meters are basically syllable counting, with
phrase-final stresses matched to strong beats in some genres (Dell & Halle 2009).
Since syllable weight plays an important role in English phonology and versification,4 it may
seem surprising that all attempts to create wholly or partly weight-based meters in English have
failed. Not that these solo efforts produced bad poetry: On the contrary, Philip Sidney’s quan-
titative hexameters and elegiac distichs in Old Arcadia are quite attractive, and Gerard Manley
Hopkins’s Sprung Rhythm, a meter of his own devising which is based on both stress and weight,
is the vehicle of some of the most gorgeous verse the language has to offer.5 The real reason
that Sidney’s and Hopkins’s innovative weight-sensitive meters failed to gain traction is probably
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

the phonological opacity of the English syllable weight distinction, due to its dual role in condi-
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

tioning stress in the lexical phonology (Chomsky & Halle 1968) and being, in turn, modified by
stress-conditioned resyllabification in the postlexical phonology (Kahn 1976). Generative phonol-
ogy reveals the phonological rationale behind the quantitative experiments, and vindicates them
against complacent critics who denied the phonological status of syllable quantity in English6 and
blamed the poets for being confused.7
Most quantitative meters select one or more constraints from the following menu:8
(5a) Strong positions
1. Must be a bimoraic foot (less restrictive)
2. Must be a bimoraic syllable – (more restrictive)
(5b) Weak positions
1. Cannot be a bimoraic foot (more restrictive)
2. Cannot be a bimoraic syllable – (less restrictive)
Note that a meter in which both examples 5a2 and 5b1 are enforced is isosyllabic.

4 See below on W-resolution, which makes the moraic trochee the maximal occupant of a metrical position, a

quantitative principle that remained alive from Beowulf through the twentieth century.
5 The rest of this paragraph draws on Kiparsky (1989) for Sprung Rhythm and Hanson (2001) for Sidney’s

quantitative verse.
6 “Since the would-be quantitative poet was obliged to remember constantly the arbitrarily assigned ‘quanti-

ties’ of the English syllables he chose to use, quantitative composition was a laborious academic-theoretical
business, like all such nonempirical enterprises more gratifying to the self-congratulating practitioner than
to the perplexed reader” (Fussell 1979, p. 68). “The rules that the poets used for determining syllable quan-
tity applied to the written text, but not necessarily to what one would hear. . . . We must conclude that the
quantitative experiment is somewhat like a written code—one needs to count and measure letters in order
to determine the system” (Hollander 1985, pp. 65–66). “Length is Modern English is phonetic rather than
phonemic” (Steele 1999, p. 267).
7 “This atmosphere of confusion and muddlement enwraps almost all those who commit whoredom with this

enchantress [quantitative meter],” “a curious measles or distemper which, dangerously but not by any means
without beneficial results, affected English poetry. . .for more than half a century” (Saintsbury 1908, p. 168),
“unnatural, arbitrary, inconsistent, and dependent on spelling” (Attridge 1974). To his credit, Attridge rec-
ognized that he tried to base his quantitative verse on the actual pronunciation of English, understood the
distinction between weight and stress, was aware of the role of quantity in English, and tried to align stress
with the ictus.
8 In examples 5 and 6, stands for a light syllable, – for a heavy syllable, and for a heavy syllable or two light
syllables.

www.annualreviews.org • Metered Verse 31

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

The various binary meters of Greek and Latin exploit all these options (West 1982, pp. 88–93;
Kiparsky 2018b):
(6a) Strict iambic verse, with feet of the form – (examples 5a2, 5b1)
(6b) Iambic with resolution in Strong, with feet of the form (examples 5a1, 5b1)
(6c) Iambic with resolution in Strong and split Weak, with feet of the form
(examples 5a1, 5b2)
In ternary quantitative meters, both Strong and Weak are bimoraic (moraic trochees), and the
correspondence constraints on positions determine the distribution of their monosyllabic and
disyllabic realizations.
Tonal meters are widespread in Southeast Asia. A typical form is the popular Vietnamese
Luc Bat “six–eight” stanza, with alternating six-syllable and eight-syllable lines. Words fall into
two tonal classes, flat (bằng) and sharp (trắc). Flat words have either no tone or a low falling
tone. Sharp words can have one of the tones that have a high component: sắc, hỏi, ngã, or nặng.
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Example 7 is from Kim Vân Kiều ‘The Tale of by Kiều’ by Nguyễn Du (translated by Lê Xuân
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Thuy):

(7)

The end of an eight-syllable line rhymes with the end of the next six-syllable line and with the
sixth syllable of the next eight-syllable line (rhymes underlined in example 7). Two words rhyme
if both have sharp or flat tone, and identical or similar codas (rhymes with identical nuclei and
codas are considered perfect). Even-numbered syllables, which I assume to be the heads of binary
feet, are tonally restricted as shown. The schema is shown in example 8, where  stands for flat, 
for sharp, and × for unspecified:9

(8) ×  ×  × a
×  ×  × a  b
×  ×  × b
×  ×  × b  c
I posit the following structure for a distich (D), with the rule that the head of a Strong foot is
 and the head of a Weak foot is , and that the Strongest positions (6 and 8) rhyme:

9 The schema is adopted from http://en.wikipedia.org/wiki/Lu.c_bat, where more references are given. An
exception mentioned there is that the second syllable in odd lines is free when there is a break after the third.

32 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

(9)

3. ENGLISH METER
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

The meter of English is based on stress (some varieties of it also on syllable weight, e.g., Hopkins’s
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Sprung Rhythm; Kiparsky 1989). Stress affects the temporal organization of the utterance and
provides the inflection points for intonation contours. The local cues for stress in English are
pitch, duration, loudness, and the distribution of certain segmental phonological features, such
as vowel reduction in unstressed syllables and tapping of coronal stops before them (e.g., atom
[ǽɾəm], atomic [əth ámɨk]).
Like quantitative meters, stress-based meters can require either prominence in Strong posi-
tions or nonprominence in Weak positions, or both. Prominence in stress-based meters is usually
defined by stress peaks, syllables bearing the strongest stress in some domain—a foot, a word,
or a phrase. For instance, philósophy and philósophìze have one peak (underlined) in each domain,
philósophìzing has two foot peaks but still just one word peak, and tìme pást has only a phrasal
peak:
(10a) Constraint on Strong Positions: A Strong position must contain a stress peak.
(10b) Constraint on Weak Positions: A Weak position cannot contain a stress peak.
Verse-initial and phrase-initial positions may be exempt (“inversion”).
Popular meters of English obey example 10a. The standard literary binary meters, notably iambic
pentameter, do not obey it. Instead, they impose example 10b, varying in what counts as a peak. For
example, Milton’s verse restricts phrasal peaks, while Shakespeare’s restricts word peaks (Kiparsky
1977).
Microvariation in the correspondence constraints and in the prosodic phonology of poetic
language generates a vast repertoire of metrical “dialects” or “styles,” especially in major meters
like iambic pentameter. One of the richest is the blank verse of Shakespeare’s plays.10
The basic iambic pentameter template is example 11:
(11) Shakespeare’s iambic pentameter
(11a) The basic pattern includes five iambic (Weak–Strong) feet, hierarchically organized
as in example 19a.
(11b) At the end of a verse and before a phrase boundary, there can be an additional Weak
position (an extrametrical syllable), shown by a .

10 On the stricter, simpler form of iambic pentameter of his sonnets and narrative poems, see Hanson
(2006).

www.annualreviews.org • Metered Verse 33

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

The principal correspondence constraints are those in example 12:

(12a) Constraint on Positions: A position contains one syllable (obeyed in standard

binary meters of English, except under the conditions specified below).
(12b) Constraint on Weak Positions: A Weak position cannot contain a stress peak
(obeyed in standard binary meters of English, except for verse-initial and phrase-
initial positions).

Compare the passage from Shakespeare in example 13a with its constructed unmetrical counter-
part in example 13b, obtained by replacing monosyllabic stresses by polysyllabic ones (underlined)
to create violations of the Constraint on Weak Positions:
(13a) And thou, / thrice-crow/nèd Queen/ of Night, / survey
With thy / chaste eye, / from thy / pale sphere / above,
Thy hun/tress’ name / that my / full life / doth sway.
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

As You Like It, 3.2.2–4

Access provided by 90.44.193.167 on 02/02/21. For personal use only.

∗
(13b) And thou, / well-main/tain’d Queen / of Night, / survey
∗
With con/cealed eyes, / from re/mote /spheres above,
∗
Thy hun/tress’ name / that en/tire lives / doth sway.
(construct)

In the blank verse of Shakespeare’s plays, optional elision and resolution allow a sequence of
syllables to occupy a single position. Elision is the phonological reduction of two syllables to one,
either by syncope or by contraction:

(14) Syncope elides a short vowel before a sonorant between a stressed syllable and an
unstressed syllable within the same word.
(14a) I speak / not as / in ab/ fear / of you.
(Macbeth, 4.3.39)
(14b) And see/ing ig/ is / the curse / of God
(Henry VI, Part II, 4.7.72)
(14c) Upon / thy eye-/balls mur/ ty/ranny
(Henry VI, Part II, 3.2.55)
(14d) That good / Duke Humph/rey trai/ / is
(Henry VI, Part II, 3.2.124)

Syncope usually does not apply at the end of a line or before a break. It remains a live process in
modern English, as in infinite → inf’nite or gén(e)rative versus géneràte:

(15) Contraction elides an unstressed vowel after a stressed one.11

(15a) Then, / ’twas he / that made / you to / depose

(Henry VI, Part III, 1.2.26)

11 A few words allow contraction across v: ne’er, o’er, e’en, but not ∗ cle’er, ∗ clo’er, ∗ bea’er.

34 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

(15b) Thou / to Co/ventry, / there to / behold

(Richard II, 1.2.258)
(15c) A sooth/ bids you / beware / the Ides / of March
(Julius Caesar, 1.2.19)
While these two elision rules are optional processes of the poetic language itself, resolution is a
metrical license located among the correspondence constraints, which allows the placement of two
syllables in one position. Being a metrical license, (a) it is not reflected in actual pronunciation, (b) it
is sensitive to metrical structure, applying to Strong but not Weak positions, and (c) it is restricted
to specific meters. Shakespeare’s dramatic verse uses two types of resolution: W-resolution and
F-resolution (unmetrical constructs added to show the limits of the operation):
(16) W-resolution: A light stressed syllable followed by another syllable within the same
word can occupy a single Strong metrical position.
 
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Tyranni/cal
(16a) power: / if he / evade / us there
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

∗
Identi/cal
(Coriolanus, 3.3.2)
 
their prodi/gal
(16b) Stoop with / oppres/sion of / ∗ weight
majesti/cal
(Richard II, 3.4.34)
 
ami/ty
(16c) Of god/like ∗ ; which / appears / most
chasti/ty
(The Merchant of Venice, 3.4.3)

(17) F-resolution: Two function words (most often a preposition + the) can occupy
a single Weak position.
(17a) His fel/lowship / in the cause / against / your
(Titus Andronicus, 5.2.13)
(17b) Ill in / myself / to see, / and in thee / ill
(Richard II, 2.1.94)
Elision, resolution, and extrametricality may be combined to yield feet of four syllables, or even
five, as in examples 18a and 18b, where the last foot has syncope, W-resolution, and an extramet-
rical syllable:

(18a) And take / my milk / for gall, / you mur/ mini

(Macbeth, 1.5.49)
(18b) With a / of / the in/ flatte
(Timon of Athens, 5.1.36)
(18c) What’s He/cuba / to him, / or he / to
(Hamlet, 2.2.585)
(18d) Some griefs / are medi/cina/ble; that / is one
(Cymbeline, 3.2.33)
(18e) To call / for . / Appear / it to / your mind
(Troilus and Cressida, 3.3.3)

www.annualreviews.org • Metered Verse 35

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

1 0.3

2 2.8

3 5.3

Metrical position
4 32.7

5 20.1

6 22.9

7 12.0

8 3.6

9 0.3
0 5 10 15 20 25 30 35
Percentages of caesuras
Figure 2
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Percentages of caesuras after each metrical position in early iambic pentameter, N = 358. Based on samples
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

from Shakespeare, Milton, Pope, and Wordsworth. Data from Keppel-Jones (2001, p. 234).

In Shakespeare’s iambic pentameter, the feet are most commonly grouped as 2+3, with a
caesura after the fourth syllable. Figure 2 shows the percentages of caesuras after each position
in early iambic pentameter.
In the later plays, the caesura after the sixth syllable becomes more common (Tarlinskaja 2014,
p. 164). This suggests that the basic structure of Shakespeare’s line shifts from example 19a to
19b:12

(19a)

(19b)

12 The right branching in example 19a might reflect the general stylistic long-last preference, as in friends,

Romans, countrymen; let her rot, and perish, and be damned to-night; lands and revenues; drawn and ready; soft and
delicate (Ryan 2019).

36 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

1 1.9

2 7.1
Metrical position 3 9.3

4 18.7

5 11.7

6 20.7

7 16.0

8 11.2

9 4.7
0 5 10 15 20 25
Percentages of caesuras
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Figure 3
Percentages of caesuras after each metrical position in Browning’s blank verse, N = 321. Based on “My Last
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Duchess” and “Return of the Druses” (beginning).

Varying the placement of the caesura in successive lines is an important feature of verse. In
neoclassical iambic pentameter distichs (heroic couplets), a caesura after the fifth position becomes
more common. Dividing the line into two constituents of equal size privileges the Parallelism
Constraint, ubiquitous in meter and music (Lerdahl & Jackendoff 1983, p. 51), over Strict
Layering and Alignment. In the nineteenth century, the caesura shifts still further rightward;
Browning favors it after the sixth beat in his early work (Figure 3) and after the seventh beat in
his later work.
The descriptive generalizations outlined here beg for explanations. Why does W-resolution
occur only in Strong positions and F-resolution only in Weak positions? The answer is that
W-resolution in Weak positions would violate the Constraint on Weak Positions, which
is otherwise strictly obeyed. Example 20 presents a near-minimal pair that shows the contrast
between elision and resolution:
 
A sooth/
(20) ∗ bids / beware / the Ides / of March
A truth/teller
(Julius Caesar, 1.2.19)

Conversely, F-resolution in Strong positions, as in construct 21b, would introduce a joint violation
of both examples 10a and 12a:

(21a) Whom lep/rosy / o’ertake! — / i’ the midst / o’ the fight,

(Antony and Cleopatra, 3.10.11)
∗
(21b) Whom lep/rosy / take i’ the / midst o’ the / campaign
(construct)

Why are lexical stresses of polysyllabic words more restricted than stressed monosyllables? In
particular, what makes them less suitable to appear in Weak positions? Lexical stresses of polysyl-
labic words are more salient than those of monosyllabic words, since they contrast with lexically
unstressed syllables within the same word. By the same token, unstressed syllables of polysyllabic
words are more saliently unstressed than those of monosyllabic words, and hence are even less
suited to Strong positions, for they contrast with lexically stressed syllables within the same word.
These considerations predict a hierarchy of prominence:

www.annualreviews.org • Metered Verse 37

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

(22) Ṕ a stressed syllable in a polysyllabic word (most prominent)

|
Ḿ a stressed syllable in a monosyllabic word
|
×
M an unstressed syllable in a monosyllabic word (necessarily a function word)
|
×
P an unstressed syllable in a polysyllabic word (least prominent)
The following implicational relations are therefore predicted for constraints on metrical Strong
and Weak positions (> means “implies”):
× ×
(23a) Constraints requiring stress in Strong positions: P > M > Ḿ > Ṕ
× ×
(23b) Constraints requiring absence of stress in Weak positions: Ṕ > Ḿ > M > P
Variation in English meter provides abundant support for this hierarchy. One example must
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

suffice here: the types of extrametrical syllable permitted in different varieties of English iambic
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

pentameter. The most restrictive system, found in early Marlowe (Schlerman 1989, p. 200), allows
only unstressed syllables of polysyllabic words to be extrametrical, as in example 24a. In his later
work, Marlowe adopts the commoner, less restrictive practice of allowing any unstressed syllable
to be extrametrical, also observed by Shakespeare (see example 24b; Schlerman 1989, p. 202).
Shakespeare’s plays (but not his sonnets) extend it to secondary stresses of compound words, as in
example 24c (Kiparsky 1977). The most unbuttoned of the Jacobean dramatists, such as Fletcher,
put stressed full words, occasionally even phrasal peaks, into extrametrical positions (example 24d):

(24a) And sit / with Tam/burlaine / in all / his

(Tamburlaine, 1.2.209)
(24b) I come / to bu/ry Cae/sar, not / to praise
Julius Caesar, 3.2.73
(24c) Quite o/verca/nopied / with lus/cious wood
(A Midsummer Night’s Dream, 2.1.251)
(24d) But I / would reach you, / and bring / you to / your trot .
(The Tamer Tamed, 1.3.9)

4. TEXT-SETTING
How is a metrical text fitted to the rhythmic pattern of a song or chant? The expectation is that
prosodic features play a role in text-setting to the extent that they are important in meter and
phonology. As mentioned above, these prosodic features are not necessarily phonemic, but they
must function in the language’s lexical phonology. Before proceeding to explore this conjecture,
let us consider the relation between silence and phrasing in meter and music.
Structure above the line is organized primarily by two antagonistic constraints, Parallelism,
which requires dividing a unit into two constituents of equal size, and Closure, which requires
making the final unit salient in some way. The normal way of achieving metrical closure is by
making the final beat silent, which in song is realized by elongating the last overt one (Blumenfeld
2016).13 But silent beats have other functions as well, which are of particular interest for the theory
of meter.

13 This is consistent with the closure effects mentioned in footnote 12, for there is no reason to posit silent feet

in pentameters; in any case, the highest layer of constituency in blank verse is the line, so putting the shorter
hemistich last gains no cohesion over putting the longer one last.
38 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

Let us begin with a simple example of a Hausa verse narrative documented by Schuh (1995).14
The performer, the popular artist Ɗan Maraya Jos, accompanying himself on a plucked string
instrument, uses three voices, a narrator (N) and two enacted characters, a woman (W, in boldface)
and a fraudulent herbalist (H, in italics). As the extract in example 25 shows, the lines are of variable
length, and no regular meter is discernible at first glance (macrons show vowel length; the other
diacritics show tone):
(25)
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Once we understand the composition’s short lines as having empty metrical positions, its meter is
instantly recognizable as regular anapestic tetrameter, a standard meter of Hausa:15

(26)

The performance of the artist is the key to recognizing the meter. The empty positions in the meter
correspond exactly to empty beats in the performance. After each sequence of missing beats, the
singer picks up the regular metrical pattern. The empty beats are obvious from the accompaniment
but cannot be parsed from the text alone, so the meter becomes intelligible only through the
performance. The theoretical interest of this example is that it shows the metrical relevance of
phonologically unrealized structure. It supports template-matching approaches over the parsing
approach of Fabb & Halle (2008) and the “holistic” prosodic approach of Golston & Riad (2000)
and Riad (2016). These theories operate on language directly and posit no independent pattern
to supply the expectation of the missing beats.

14 See also http://aflang.humanities.ucla.edu/language-materials/chadic-languages/hausa/hausa-

poetry-song/ and Hayes & Schuh (2018).
15 Here I diverge from Schuh’s (1995) analysis.

www.annualreviews.org • Metered Verse 39

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

Empty beats challenge bottom-up parsing theories such as that of Fabb & Halle (2008), since
one cannot parse what is not there. Their directional parsing algorithm also runs into difficulties
in that the lines of this text would have to be scanned both left to right (for lines ending with empty
beats) and right to left (for lines beginning with empty beats).16 Moreover, the operations on grids
that are their major means of negotiating mismatches cannot generate the short lines because the
sequences of empty beats need not be constituents. Lines can begin or end with an odd number of
empty beats, including sequences of three and five beats, which cannot be characterized as groups
at any level, whether feet, dipods, or half-lines. But they are easily recognized in performance.
Empty beats also provide compelling arguments for including performance in the province of
metrical theory, which Fabb & Halle (2008) explicitly reject. It turns out that line-medial empty
beats occur only when the singer changes character—a salient event in performance but not a
formal property of the text.
The prosodic metrical theory of Golston & Riad (2000) would probably have to deal with these
data by different constraint rankings that output lines of various lengths, in this short sample
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

alone of three, five, six, and seven metrical positions, both left-aligned and right-aligned. This
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

would miss the generalization that lines of any length are admissible provided that they consist of
a subsequence of the eight-syllable anapestic patterns.
Empty beats occur in Western art poetry as well.17 Matthew Arnold’s “Dover Beach” drama-
tizes a speaker expressing his thoughts to a companion as they view the Channel at night. Amidst
the poem’s subtly rhymed regular iambic pentameters are occasional shorter lines of four and three
feet, placed to mark significant silences in the represented discourse. Example 27 reproduces lines
1–14, with ellipses to show where Samuel Barber’s composition (Opus 3) puts a corresponding
break in the short lines. Two such short lines begin the poem, conveying a sense of thoughts
emerging from a contemplative silence. At line 9, there is again a missing beat, as the participants
listen to the surf:
(27) … The sea is calm tonight. …
… The tide is full, the moon lies fair
Upon the straits; on the French coast, the light
Gleams and is gone; the cliffs of England stand,
Glimmering and vast, out in the tranquil bay.
Come to the window, sweet is the night-air!
Only, from the long line of spray
Where the sea meets the moon-blanched land,
Listen! … you hear the grating roar
Of pebbles which the waves draw back, and fling,
At their return, up the high strand,
Begin, and cease, and then again begin,
With tremulous cadence slow, and bring
The eternal note of sadness in. …
A purely formal analysis of the text as having a meter with variable line length misses the dramatic
function of the short lines.
In setting English poems to music, most composers privilege the natural phonological phras-
ing, even at the expense of the poem’s lineation and caesuras where they conflict. Phonological
prominence also tends to trump metrical prominence, in that stressed syllables, especially of
16 Note that it has what Fabb & Halle (2008) classify as a strict meter. Loose meters by definition are those

which have at least one noniterative “parenthesis insertion” rule, which this meter does not have.
17 There is a famous one in Hopkins’s “Spelt from Sibyl’s Leaves.”

40 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

Figure 4
Sure On This Shining Night (from FOUR SONGS, OP. 13). Music by Samuel Barber, words by James Agee. Copyright 1941
(Renewed) by G. Schirmer, Inc. International Copyright Secured. All Rights Reserved. Figure adapted with permission from
G. Schirmer.

polysyllabic words, are preferentially aligned with Strong musical beats, regardless of whether
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

they fall in Strong or Weak positions in the verse. Weak positions in iambic verse can therefore
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

correspond to Strong musical positions, and conversely, especially if the phrasing and natural
stress pattern require it (Halle & Lerdahl 1993, Hayes 2009).
A simple example is the first quatrain of James Agee’s iambic “Sure on This Shining Night,”
which features a rhythmic parallelism of the couplets, with initial inversion in the odd lines:
(28) Sure on / this shi/ning night
Of star/-made sha/dows round,
Kindness / must watch / for me
This side / the ground.
Samuel Barber’s composition Four Songs (Figure 4) highlights this parallelism, without attempting
a close rendering of the metrical foot structure.
Dell & Halle (2009) show that while English matches stresses to strong musical beats across
the board, French does so only at the ends of lines, and conversely that French traditional songs
require a parallel pairing of syllables to beats in each stanza, which is not the case in English.
They propose to derive both differences from the fact that stress in English is perceptually salient
throughout the utterance, and only before major breaks in French, and that French meter is
basically syllable counting.
For the rendering of stress and quantity it is difficult to give any general rules. It varies by style
and genre even within a language and period. Finnish provides an illustration of this variability.
In this language, the primary word stress falls on the initial syllable. Vowel and consonant length
is contrastive throughout the word, as in tule ‘come!’, tulee ‘comes’, tuule ‘blow!’ (wind), tuulee
‘blows’, tullee ‘probably comes’, and tuullee ‘it is probably windy’. But in singing, vowel length can
be neutralized by lengthening, and stress can be displaced onto syllables that fall in prominent
positions. In the following rendition of a traditional Finnish ballad (the counterpart of “Edward,
Edward”), the singers lengthen the stressed open syllables, pronouncing, for instance, túlet ‘you
come’ as túulet:18
(29)

18 https://www.youtube.com/watch?v=NwAcHcuYzL4.

www.annualreviews.org • Metered Verse 41

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

When an unstressed syllable has to fall on a strong beat, they sometimes shift the word stress onto
it, as in line 3 of this verse:
(30)

Native Finnish meters are ingeniously adapted to the coexistence of phonemic vowel length
and initial stress so as to jointly optimize variety and phonological faithfulness. The Kalevala me-
ter consists of eight-syllable lines made up of four trochaic (Strong–Weak) feet, with obligatory
alliteration and parallelism. The basic metrical rule is that a stressed (i.e., word-initial) syllable
must be < in Weak positions and – in Strong positions (< = CV̆ ). This avoids lengthening of
vowels in the sensitive first syllable, which would neutralize the most important site of the length
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

contrast. The resulting mismatches between stress and metrical positions in words with an odd
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

number of syllables create the main rhythmic excitement of this meter.

In the rendition by Niekku, the singers foreground the tension between phonology and meter
by resolving it differently in repetitions of lines 1 and 3. They sing them once with the original
word stress, and the second time they shift the word stress to the syllable that falls on the metrical
stress, in each case with lengthening of some of the stressed vowels:19

(31)

In general, distinctive quantity seems to be reflected more faithfully in text-setting than stress
is (Hayes 2016). Since syllable weight is the only active prosodic feature in the lexical phonology
of Urdu (or, at least, by far the most important one), and its meters are also based on weight, good
singers take care to preserve it. Begum Akhtar’s ghazal renderings respect syllable weight as far
as possible while maintaining Urdu’s phonemic vowel length contrast, even at the cost of some
compromises in the syllable count. Metrically prominent positions are sung as bimoraic. In words
like gul ‘rose’, this is done by extending the note over the vowel and the sonorant coda. Words like
səb ‘all’ are bimoraic and occupy heavy positions in the meter, but their closing obstruent cannot
carry a note, nor can the phonemically short vowel be lengthened lest the length contrast is lost.
In positions that require a heavy syllable, such words are sung with final -ə inserted as a last resort,
which allows their heavy syllables that have only one singable mora to fill the musical space allotted
to them. In this system, faithfulness to syllable weight trumps faithfulness to the syllable count.20
The investigation of text-to-tune alignment in crosslinguistic perspective is only beginning, but
preliminary results already indicate that the negotiation of phonology and meter in song provides
a sensitive probe into the prosodic organization of language.

19 Suomen kansanmusiikki 1 (Folk Music of Finland 1), Kansanmusiikki-instituutti 1988.

20 Afine example is Begum Akhtar’s rendering of Ghalib’s Sab Kahan Kuch Lala-o-Gul Mein (https://www.
youtube.com/watch?v=0wggu6H8Qxo).

42 Kiparsky
LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

DISCLOSURE STATEMENT
The author is not aware of any affiliations, memberships, funding, or financial holdings that might
be perceived as affecting the objectivity of this review.

LITERATURE CITED
Attridge D. 1974. Well-Weighed Syllables: Elizabethan Verse in Classical Metres. Cambridge, UK: Cambridge Univ.
Press
Beissinger MH. 2012. Oral poetry. In The Princeton Encyclopedia of Poetry and Poetics, ed. G Roland, S Cushman,
pp. 978–81. Princeton, NJ/Oxford, UK: Oxford Univ. Press
Blumenfeld L. 2015. Meter as faithfulness. Nat. Lang. Linguist. Theory 33:79–125
Blumenfeld L. 2016. End-weight effects in verse and language. Stud. Metr. Poet. 3:7–32
Borgeson S, Anttila A, Heuser R, Kiparsky P. The rise and fall of antimetricality. Forthcoming. https://web.
stanford.edu/∼kiparsky/Papers/Antimetricality-Draft-2018-04-20.pdf
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Dell F, Halle J. 2009. Comparing musical textsetting in French and in English songs. In Towards a Typology
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

of Poetic Forms: From Language to Metrics and Beyond, ed. J-L Aroui, A Arleo, pp. 63–78. Amsterdam:
Benjamins
Deo AS. 2007. The metrical organization of classical Sanskrit verse. J. Linguist. 43:63–114
Fabb N, Halle M. 2008. Meter in Poetry: A New Theory. Cambridge, UK: Cambridge Univ. Press
Fussell P. 1979. Poetic Meter and Poetic Form. New York: Random House
Golston C, Riad T. 2000. The phonology of Classical Greek meter. Linguistics 38:99–167
Halle J, Lerdahl F. 1993. A generative textsetting model. Curr. Musicol. 55:3–23
Hanson K. 2001. Quantitative meter in English: the lesson of Sir Philip Sidney. Engl. Lang. Linguist. 5:41–91
Hanson K. 2006. Shakespeare’s lyric and dramatic metrical styles. In Formal Approaches to Poetry, ed.
BE Dresher, N Friedberg, pp. 111–33. Berlin: Mouton de Gruyter
Hanson K, Kiparsky P. 1996. A parametric theory of poetic meter. Language 72:287–335
Hanson K, Kiparsky P. 1997. The nature of verse and its consequences for the mixed form. In Prosimetrum:
Cross-Cultural Perspectives on Narrative in Prose and Verse, ed. J Harris, T Ziolkowski, pp. 17–44. Cambridge,
UK: Brewer
Hayes B. 1989. The prosodic hierarchy in meter. In Phonetics and Phonology, Vol. 1: Rhythm and Meter, ed.
P Kiparsky, G Youmans, pp. 201–60. San Diego: Academic
Hayes B. 2009. Textsetting as constraint conflict. In Towards a Typology of Poetic Forms: From Language to Metrics
and Beyond, ed. J-L Aroui, A Arleo, pp. 43–61. Amsterdam: Benjamins
Hayes B. 2016. Stochastic constraint-based grammars for Hausa verse and song. Paper presented at Old World Conf.
Phonol., Budapest, Jan. 16
Hayes B, Schuh R. 2018. Metrical structure and sung rhythm of the Hausa Rajaz. Phonol. Anal. Forthcoming.
https://roa.rutgers.edu/content/article/files/1768_hayes_1.pdf
Herder JG. 1960 (1768). Fragmente. In Sprachphilosophische Schriften, ed. E Heintel, pp. 91–154. Hamburg,
Ger.: Meiner
Hollander J. 1985. Observations in the art of English quantity. In Vision and Resonance: Two Senses of Poetic Form,
pp. 59–70. New Haven: Yale Univ. Press. 2nd ed.
Itô J, Mester A. 2003 (1992). Weak layering and word binarity. In A New Century of Phonology and Phonological
Theory: A Festschrift for Professor Shosuke Haraguchi on the Occasion of His Sixtieth Birthday, ed. T Honma,
M Okazaki, T Tabata, S Tanaka, pp. 25–65. Tokyo: Kaitakusha
Kahn D. 1976. Syllable-based generalizations in English phonology. PhD Thesis, MIT, Cambridge, MA. http://
hdl.handle.net/1721.1/16397
Keppel-Jones D. 2001. The Strict Metrical Tradition: Variations in the Literary Iambic Pentameter from Sidney and
Spenser to Matthew Arnold. Montreal, Can.: McGill/Queen’s Univ. Press
Kiparsky P. 1977. The rhythmic structure of English verse. Linguist. Inq. 8:189–247
Kiparsky P. 1989. Sprung rhythm. In Phonetics and Phonology, Vol. 1: Rhythm and Meter, ed. P Kiparsky,
G Youmans, pp. 305–40. San Diego: Academic

www.annualreviews.org • Metered Verse 43

LI06CH02_Kiparsky ARjats.cls November 26, 2019 9:2

Kiparsky P. 2006. A modular metrics for folk verse. In Formal Approaches to Poetry, ed. BE Dresher, N Friedberg,
pp. 7–49. The Hague: Mouton
Kiparsky P. 2007. Iambic inversion in Finnish. In A Man of Measure: Festschrift in Honour of Fred Karlsson on
His 60th Birthday, ed. M Suominen, A Arppe, A Airola, O Heinämäki, pp. 138–48. Turku, Finl.: Linguist.
Assoc. Finl.
Kiparsky P. 2018a. Formal and empirical issues in phonological typology. In Phonological Typology, ed. L Hyman,
F Plank, pp. 54–106. Berlin: Mouton de Gruyter
Kiparsky P. 2018b. Indo-European origins of the Greek hexameter. In Sprache und Metrik, ed. D Gunkel,
O Hackstein, pp. 77–127. Leiden, Neth./Boston: Brill
Lerdahl F, Jackendoff R. 1983. A Generative Theory of Tonal Music. Cambridge, MA: MIT Press
Liberman M, Prince A. 1977. On stress and linguistic rhythm. Linguist. Inq. 8:249–336
Prince AS. 1983. Relating to the grid. Linguist. Inq. 14:19–100
Riad T. 2016. The meter of Tashlhiyt Berber songs. Nat. Lang. Linguist. Theory 35:499–548
Ryan KM. 2017. The stress–weight interface in metre. Phonology 34:581–613
Ryan KM. 2019. Prosodic end-weight reflects phrasal stress. Nat. Lang. Linguist. Theory 37:315–56
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Saintsbury G. 1908. A History of English Prosody from the Twelfth Century to the Present Day, Vol. 2: From Shake-
Access provided by 90.44.193.167 on 02/02/21. For personal use only.

speare to Crabbe. London: Macmillan

Schlerman BJ. 1989. The Meters of John Webster. New York: Lang
Schuh RG. 1995. The metrics of three Hausa songs on marriage by Ɗan Maraya Jos. Unpubl. ms. http://citeseerx.
ist.psu.edu/viewdoc/download?doi=10.1.1.584.6468&rep=rep1&type=pdf
Steele T. 1999. All the Fun’s in How You Say a Thing. Athens: Ohio Univ. Press
Tarlinskaja M. 2014. Shakespeare and the Versification of English Drama, 1561–1622. Farnham, UK: Ashgate
West ML. 1982. Greek Metre. Oxford, UK: Clarendon
Zec D. 2009. The prosodic word as a unit in poetic meter. In The Nature of the Word: Essays in Honor of Paul
Kiparsky, ed. K Hanson, S Inkelas, pp. 63–94. Cambridge, MA: MIT Press

44 Kiparsky
 LI06_TOC ARI 27 November 2019 16:56

Annual Review
of Linguistics

Volume 6, 2020

Contents

On Becoming a Physicist of Mind

Willem J.M. Levelt p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 1
Metered Verse
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Paul Kiparsky p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p25

Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Grammatical Gender: A Close Look at Gender Assignment

Across Languages
Ruth Kramer p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p45
The Role of the Lexicon in the Syntax–Semantics Interface
Stephen Wechsler p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p67
The Syntax of Adverbials
Thomas Ernst p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p89
Successive Cyclicity and the Syntax of Long-Distance Dependencies
Coppe van Urk p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 111
Antipassives in Crosslinguistic Perspective
Raina Heaton p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 131
Lexical-Functional Grammar: An Overview
Kersti Börjars p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 155
Determiners and Bare Nouns
Veneeta Dayal and Yağmur Sağ p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 173
Treebanks in Historical Syntax
Ann Taylor p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 195
Distributional Semantics and Linguistic Theory
Gemma Boleda p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 213
The Grammar of Degree: Gradability Across Languages
Vera Hohaus and M. Ryan Bochnak p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 235
Techniques in Complex Semantic Fieldwork
M. Ryan Bochnak and Lisa Matthewson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 261
 LI06_TOC ARI 27 November 2019 16:56

From African American Vernacular English to African American

Language: Rethinking the Study of Race and Language in African
Americans’ Speech
Sharese King p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 285
The Status of Endangered Contact Languages of the World
Nala H. Lee p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 301
Individual Differences in First Language Acquisition
Evan Kidd and Seamus Donnelly p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 319
Language Variation and Social Networks
Devyani Sharma and Robin Dodsworth p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 341
Sociolinguistics of the Spanish-Speaking World
Annu. Rev. Linguist. 2020.6:25-44. Downloaded from www.annualreviews.org

Manuel Dı́az-Campos, Juan M. Escalona Torres, and Valentyna Filimonova p p p p p p p p p p 363

Access provided by 90.44.193.167 on 02/02/21. For personal use only.

Language and Discrimination: Generating Meaning, Perceiving

Identities, and Discriminating Outcomes
Justin T. Craft, Kelly E. Wright, Rachel Elizabeth Weissler,
and Robin M. Queen p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 389
Language and Masculinities: History, Development, and Future
Robert Lawson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 409
Linguistic Perspectives on Register
Larissa Goulart, Bethany Gray, Shelley Staples, Amanda Black,
Aisha Shelton, Douglas Biber, Jesse Egbert, and Stacey Wizner p p p p p p p p p p p p p p p p p p p p p p 435
Fantastic Linguistics
Sarah Thomason and William Poser p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 457
Errata
An online log of corrections to Annual Review of Linguistics articles may be found at
http://www.annualreviews.org/errata/linguistics