Versification: an interdisciplinary journal of literary prosody
Filename: Holder_997_BurnsCooperRev.html
4 September 1997, Vol. 1, No. 1

A Review of Rethinking Meter: A New Approach to the Verse Line by Alan Holder

G. Burns Cooper
University of Alaska, Fairbanks

Copyright © G. Burns Cooper 1997
Received: 24 September 1997; Published: 12 October 1997

KEYWORDS: meter, foot, iambic, phrase, phrasalism, intonation .

REVIEW OF: Rethinking Meter: A New Approach to the Verse Line by Alan Holder. 1995.
Bucknell University Press, Lewisburg (London: Associated University Presses). pp. 298.
ISBN: 0-8387-5292-6. $42.50 hb.

I. Introduction

This book is a significant critique of metrical criticism and theory, and offers some useful suggestions for how that criticism and theorizing should proceed in the future. Because Versification is a new journal, it seemed worthwhile to review this book even though the copyright date is 1995. Rethinking Meter will be worth a look for anyone interested in English poetic prosody.

Among the book's chief points of interest:

  • Holder reviews and critiques much of the classic literature in the field.
  • He argues against the usefulness of the concept of the poetic "foot."
  • He insists explicitly on "natural speech stress" and implicitly on an experiential standard for analysis—that is, if an effect is too subtle or abstract for a normal reader to perceive it, it should not form part of the analysis.
  • He vigorously defends "phrasalism" as an alternative to foot theory.
  • He defends the importance of "performances," as opposed to abstract and soundless "texts," in scansion and in poetic analysis generally.
  • He concludes with an original chapter on the importance of intonation in poetry, with examples of how intonation theory could be applied to criticism.
Among its chief flaws, as I see them:
  • Holder is sometimes unfair to the authors he criticizes; rather than give them the benefit of the doubt, he tends to read others' claims in the least plausible way.
  • Partly because of these ungenerous readings, he tends to overstate the radicality and freshness of his own theory and insights.
  • Though he refers to some of the key linguistic works on prosody and intonation, he does not seems to fully understand or fully integrate their insights; he also seems much less familiar with post-1980 linguistics.
  • He has an almost automatic negative reaction to elements of complexity and especially of abstraction in the theories and analyses of others. Yet ultimately he has to allow some of these same kinds of complexity into his own theory, in order to make it work.
  • A few of his own scansions and explications seem just as forced and implausible as some of those he ridicules.
A more detailed critique is given below. First, though, readers may find it useful to have an overview of the book's contents in the order that Holder presents them.

II. Overview

Rethinking Meter begins, as is common in books on prosody, with a dismissal of much of what has already been written on the subject; the first sentence of Chapter 1 reads: "Perusing the critical literature of prosody is useful for defining the issues regarding poetry's sound patterns, but the sad fact is that the bulk of that literature is wrong-headed or irrelevant, forever warming up dubious pieties"(19). Despite its wrongheadedness and irrelevancy, however, Holder devotes most of the book (the first four chapters, plus major parts of the other three) to a review of this literature.

The first two chapters, "In the Muddled Kingdom of Meter" and "A Further Look at the Foot," discuss what one might call the Tradition in English prosody. Holder brings up many of the classic 19th and 20th century works on meter, including Saintsbury, Omond, Stewart, Wimsatt and Beardsley, Brooks and Warren, and others, and more recent scholars and critics who retain some allegiance to these traditional approaches. He also reviews some of the scholarship on pre-19th century metrical theorizing, and presents a persuasive case that the foot-substitution model's claim to being the only Tradition that early English poets were following is by no means as solid as some critics have supposed. In fact, he points out that the foot model has difficulties even for Classical poetry, which is presumably the source from which English adopted it. Here and in chapter six he argues that "phrasalism," the model he advocates, has equally deep historical roots, going back to Anglo-Saxon half-lines, virgules in Chaucer manuscripts, and some (though not all) Renaissance commentaries on poetics. He tries to show where the foot-substitution model goes wrong, partly by pointing out questionable scansions produced by scholars who use that model.

"Foot-substitution" here is my term, not Holder's.  By it I mean the idea that verse forms are best described as a sequence of a certain number of a certain type of feet, and that any variation from that form is best described as the substitution of a different type of foot; for example, the iambic pentameter is a sequence of five "iambs," and a line like King Lear's "Why should a dog, a horse, a rat, have life" is an iambic pentameter with a trochee substituted for the first iamb (assuming that "Why," "dog," "horse," "rat," and "life" are the stressed words). Holder actually refers to this model just as "the foot," or even as simply "meter," but I avoid this usage because both "foot" and "meter" mean something rather different in linguistics, and I believe the alternative usages are easier to defend.

Throughout the literature review, Holder criticizes models which he considers overly abstract or overly complex, and argues stridently against some traditional prosodic notions such as "the Platonic ideal of the iambic line," "tension" or "counterpoint," "isochrony," and "meter as expectation". He also rejects the idea that all metrical variations should be considered "expressive" commentaries on the meaning of the poem.

Chapter three, "Recent Prosodic Commentary," is a critique of the work of more recent metrists, including Derek Attridge, Frederick Turner and Ernst Pöppel, Susanne Woods, various authors from the collection Phonetics and Phonology 1: Rhythm and Meter (ed. Youmans and Kiparsky), Richard Cureton, and, at greater length, George T. Wright. The critique is along similar lines to the earlier chapters: Holder praises these authors where he agrees with them, but criticizes them wherever they incorporate the concept of feet, or where he feels they are excessively abstract, or give scansions that do not match "natural" (that is, prose) stress patterns, or offer explications of meaning that he finds implausible.

Chapter four, "The Haunting of Free Verse," focuses on the idea, advanced by several critics, that ostensibly unmetered verse has a kind of "ghost meter" lurking behind it, and that the best way to understand free verse rhythm is to find these covert meters. Holder links this idea to the equating of meter with foot-substitution. He presents some examples of critics "discovering" particular kinds of feet in apparently "free" verse. Some of these scansions are more plausible than others, but Holder's argument is that the whole approach is misguided. Instead, he wants to insist on an approach that analyzes poems into phrases, rather than feet, as their primary rhythmic units. He also argues against the practice of ignoring line divisions in putting together sequences of feet, so that the critic is actually scanning lines created by that critic rather than by the poet. Finally, he reiterates his healthy skepticism about claims that poetic rhythms are closely tied to body functions such as heartbeats or neural pulses.

In Chapter five, "Preliminaries to Revision," Holder sets out his position on what he rightly calls "basic questions" about the pronunciation and organization of poems. He insists again that the words of a poem should be pronounced with only the stress they would receive in ordinary speech. Thus he is against "promoting" or "demoting" syllables because of their position in the line. He argues that the line is the primary unit of organization of poetry (both in sound and in sense), and that line integrity must be respected; that respect includes putting a pause at the end of every line. The importance of the line and of the line-end pause applies to both metrical and free verse. He briefly addresses the issue of how a poem's lineation affects the speed with which it is read. More importantly, he discusses the interaction of lineation with syntax, and how enjambment relates to line integrity and to phrasing. Finally, he discusses the nature of phrasing, and in particular the relationship of syntactic phrasing to phonological phrasing (which he sees as essentially identical).

Chapter six, "Phrasalism," is really the philosophical heart of the book. In it, Holder develops his thesis that phrases and lines (which sometimes amount to the same thing) are the most important organizational and rhythmic units in poetry, and he gives examples of possible applications of this theory to actual poems, usually comparing a non-phrasalist analysis by some other critic to a phrasalist analysis of his own. He begins with a long section informally defining what he means by "phrase." Though he does not deal with the issue explicitly, it becomes clear that he does not believe in multiple levels of phrasing, but equates "phrase" with what some linguists would call a "pause-phrase" (that is, a group of words bounded at both ends by silence)(1) and sees this as coterminous with the syntactic phrase (which he also sees as essentially one-dimensional), the intonational phrase, and the "sense-unit."

He reviews a number of other authors who could be regarded as "phrasalists," especially Charles L. Stevenson, Roger Mitchell, and Richard Cureton, but he again spends the most time of all on George Wright, who apparently coined the term "phrasalist" but is very much a traditionalist. Holder's criticisms primarily take three forms: 1) instances of these authors dividing phrases in ways he disagrees with; 2) instances of these authors scanning stress patterns in ways he disagrees with; or 3) authors either failing to provide explications of "expressive effects" of phrasing and stress or providing explications which Holder finds unconvincing. Finally, in the section on Wright's book Shakespeare's Metrical Art, Holder offers what he calls "a reformulation" of the iambic pentameter, in which he does not refer to the notions of the foot or the iamb. For that reason he suggests that it should really be called "decasyllabic verse". (This reformulation will be detailed below.) He proceeds to offer his own readings of poems discussed by these authors: Shakespeare's Sonnet 116, Pound's "In a Station of the Metro," and Plath's "Ariel."

Chapter seven, "Tuning In: Making a Place for Intonation in Prosodic Analysis," argues that intonation has been an unjustly neglected aspect of poetic prosody, and offers examples of ways it could be incorporated into prosodic criticism. The chapter begins by discussing, at some length, just what intonation is and how it works in non-poetic language: its relation to stress, accent, and phrasing, and its semantic and pragmatic functions. He argues that intonation patterns, while variable, are not indeterminate or impossibly complex, and that they can be inferred from written poems. Holder also reviews some of the literature on poetic intonation, but finds it sparse and inadequate. Though he mentions several linguists who have worked on intonation, he focuses on just two, David Crystal and Dwight Bolinger, and settles on Bolinger's system of notation and, by and large, on Bolinger's claims about the functions of intonational contours. It is these that he incorporates in the later part of the chapter in offering his own intonational analyses of poetic lines. He brings in some new examples, but mostly uses the same ones from chapter six and before (Shakespeare, Pound, Stevens, Plath), which allows a helpful comparison of approaches.

The Epilogue, "Epireading, Graphireading, and the Matter of Voicing," is a brief attempt to relate the previous chapters to a large issue in recent literary theory and philosophy, the nature of the text. Holder discusses the difference between "epireading," which sees writing as a record of past speech, and "graphireading," which takes the written text as a thing in itself rather than the transcription of a vocal/auditory event. This issue inevitably evokes the names of Jacques Derrida and other postmodernists, and much of Holder's discussion centers on their ideas, but it is worth pointing out (as Holder does) that the question is much older than that; the chapter mentions traditional Christian theology and Greek philosophy, and in more recent times, Mallarme, Wellek and Warren, and Kenneth Burke, among others. Holder reviews the debate and comes down firmly on the side of epireading. Interestingly, he argues not that epireading is the way reading is always done, but that it is the most rewarding way to read—because all the effects of sound patterning that he has been discussing can only be perceived if we experience writing as speech: "Everything I have said in this book is predicated on the assumption that, ideally at least, a poem is to be voiced, not subliminally, but out loud...."

III. Critique

Style and editing

This book reviews a lot of scholarship, some of it in depth and some only in passing. Though the breadth of Holder's research is impressive, the coherence of the book might have been improved if he included only those sources he considered most relevant and important. Many seem to be included only to show their inadequacy. I happen to have read the great majority of these sources at one time or another, yet I still sometimes had trouble following what Holder was saying that another author claimed, which made it difficult to evaluate Holder's opinion of these claims. This problem is at least partly due to his summarizing many complex ideas in a very abridged form.

Holder is highly opinionated, and that accounts for some of the pleasure the book can give (for example in his witty put-downs of "foot-fetishists," "metrical good old boys," "prosodic inquisitors," "the curious sport of foot-spotting," etc.). However, this aggressive attitude can also become abrasive, especially when he occasionally combines it with a nominalized, pedantic style: "There is, in fact, a domination of the pages on prosody by the application of that system of scansion" (22); "While allowing for rhetorical overkill in the remarks of John Nist that I quoted at the start of my introduction, one can only applaud his attributing a stultifying certainty and rigidity to traditional metrics and their pedagogic perpetuation" (23).

There are not an unusual number of editing errors, but a couple may cause real problems for readers who wish to pursue the subject further. Robert Hass (recently poet laureate of the United States) is consistently referred to as Robert Haas. Also, there is a typographical error in a scansion of a line by Walt Whitman, quoted from Richard Cureton (on p. 169), that makes both Cureton's commentary and Holder's discussion of that commentary almost impossible to follow, until we figure out, purely by inference, what the scansion is actually supposed to be.

One more production problem is probably not Holder's fault, but still makes the book less enlightening than it might otherwise be: that is the slow pace of academic publishing and distribution. The book is copyrighted in 1995, but clearly most of it was written long before that date. A good example of the difficulties this causes is in Holder's criticism of Richard Cureton's work. He finds inconsistencies among Cureton's articles and finds at least one of them too dense and "compressed" to follow easily. He remarks in an endnote that Cureton refers to "the impending publication of a long manuscript. . ." that might make some of these things clearer, but "as of the time of my [Holder's] writing this, such publication has not yet occurred" (257). The manuscript in question was surely Cureton's Rhythmic Phrasing in English Verse, which was published in 1992. It is too bad that Holder did not have it available, as this book is surely one of the most systematic and rigorously theoretical accounts of meter and rhythm that has been written. Undoubtedly, Holder would still find much to disagree with in Cureton's book, but at least he would be criticizing a unified and contextualized account. This is especially unfortunate as one would expect Holder and Cureton, both extremely interested in phrasing, to be natural allies—yet Holder is actually more respectful of "traditionalists" like Wright and Roger Mitchell.

Feet and stress

As I mentioned above, a large part of the book is devoted to a dissection of traditional metrical theory, in the hopes of replacing it with something more suitable. Holder makes his point successfully, but presses it too far.

One reason he doesn't like the foot is that it is arbitrary: foot boundaries, at least as they have traditionally been applied, do not correspond to any intuitive division of language into units of sound or sense. I have long agreed with Holder's argument here, especially after trying to teach traditional prosody to an undergraduate class. Many students have trouble deciding where to divide one foot from another; some seem to give up and do it almost at random. They do this not because they are insensitive to something that is "there" in the text, but because, unlike experienced prosodists, they do not bring preconceived categories with them and insert those categories into the text. For example, take this line from Wordsworth's The Prelude: "Make rigourous inquisition, the report" (l. 148). Other than to make it fit the paradigm of iambic pentameter, would we have any reason to divide it into these units: "Make ri |gourous in | quisi | tion, the | report"? The divisions cut across word boundaries, phrase boundaries, punctuation, sense units—almost any other grouping of language one can think of. Is "gourous in" even a linguistic entity? Or "tion, the"?

Yet this criticism is fair only if we acknowledge that the more intelligent proponents of foot-substitution never claim that feet are the same type of entity as words or phrases. No prosodist I know of has claimed that we pause after each foot. If we try to make the paradigm work as well as possible, we have to make the foot a separate dimension of measurement. The traditional foot, insofar as it makes sense at all, only makes sense as part of a sequence of feet, as a way of describing the length and general stress pattern of a line. Perhaps an analogy will help: The front wall of my log house is 14 feet high. I would argue that the preceding statement is a meaningful one. Yet no "foot" of those 14 really stands out as discrete from the others. Each log is an entity that would have some existence if the wall were dismantled, but each foot is not. Yet the foot measurement is still useful as an abstraction, because actual logs are irregular in thickness. To say a wall is "14 logs high" is precise in one way but imprecise in another, just as it is to say a line is "two phrases long".

Advocates of the traditional foot generally do acknowledge that word-boundaries, phrase-boundaries, and even pauses are part of our experience of a poem, though different writers give that experience different amounts of weight. The venerable concept of the "caesura" is one acknowledgement of phrasing. But a pure traditionalist would not consider that phrasing to be part of the meter; instead, he would consider it a separate level of structure that interacts with the meter. Holder does offer a couple of examples of well-known prosodists apparently dismissing the importance of phrasing, but more of his examples seem to be arguing merely that phrasing and meter are not the same thing, as in C.S. Lewis' "[A pause] is a rhetorical and syntactical fact, not a metrical fact" (cited in Holder, 150).

Despite my reservations on this point, much of this part of the review is insightful; Holder can adduce respected authorities for each opinion he offers, and he is adept at pointing out contradictions and gaps in the arguments and wording of those he disagrees with. He is at his best when he presents and discusses other authors' actual analyses of poems; as Holder says early on, "By their scansions shall ye know them." Some of these scansions seem very implausible indeed, which is usually Holder's point. On the other hand, he is at his worst when he seems to be more eager to find weaknesses than he is to give authors credit for what they do handle well.

Linguistic Lapses

Holder presents himself as a literary scholar, not a linguist. That being the case, he has done a pretty good job of educating himself on the linguistic theories most relevant to his arguments. However, his reading in linguistics seems to be heavily dependent on what is listed in T.V.F. Brogan's bibliography, English Versification, 1570-1980. This is, of course, the definitive work of its kind, so he is wise to start with it—but there is an awful lot of linguistic work from the '90s, '80s and even the late '70s that it does not include. What's more, even where Holder does cite the relevant texts, he sometimes does not really integrate what they say into his own views.

For example, the model of the foot Holder attacks, an arbitrary unit that crosses morpheme boundaries, word boundaries, and phrase boundaries with impunity, and has anywhere from zero to two or more stressed syllables, does seem to be motivated only by the need to have names for different verse-forms. However, many theorists—mainly linguists—have tried to redefine the foot as a unit that is closely tied to the morphology (internal structure) of words and the syntax of small phrases. The theory of metrical phonology has become well-established in the last two decades, and is supported by evidence not only from English but from various other languages. Holder never really examines whether a foot defined in a less arbitrary way could have a place in his theory.

Metrical phonology could also be brought to bear on another of Holder's pet peeves: promotion and demotion. In arguing for "natural stress," Holder (quite rightly) argues that prosodists should not violate the phonological integrity of words and phrases by changing their pronunciation just to make them fit the prosodists' ideas of  meter. He assumes that this is what we are doing any time we take a word that in another context might be unstressed and mark it as stressed, or vice versa. For him, only the dictionary-sanctioned stress on any word is "natural." What's more, he usually only recognizes one lexical stress; he doesn't seem to believe in secondary stress. In fact, he doesn't seem to believe that stress is relative; he sees it as binary, either yes or no.

But metrical phonologists have made it clear that in actual speech stress is context sensitive, and though much of the context that matters is semantic, the rhythmic context is also relevant. The famous "English rhythm rule" is one example of this principle in effect: in a phrase like "Mississippi legislators" or "Tennessee legislators," the primary stress is likely to be moved to the first syllable of each state name in order to make the phrase more regular in rhythm, even though the dictionary pronuciation would put primary stress on the third syllable of both "Mississippi" and "Tennessee." Because the first syllable of "legislators" is stressed, the stress in the preceding word moves away from it. Furthermore, in the first example, both "-sip-" and "-lat-" will receive secondary stress, that is they will be stronger than the syllables immediately adjacent to them, but not as strong as the primary stress in each word. This has the effect of creating a still more regular alternation: STRONG-weak-strong-weak-STRONG-weak-strong-weak.

Not only English, but many unrelated languages allow manipulation of stress to create more regular metrical patterns; Yup'ik Eskimo, for example, has a rhythmic stress rule that tends to put stress on alternating syllables. Most, if not all, theorists of metrical phonology have some provision for this kind of promotion and demotion—including Elisabeth Selkirk,(2) whose views on phrasing and pausing Holder relies on heavily. We could speculate that in metrical poetry this rhythmic effect might be even more pronounced than in ordinary speech. But he seems not to have taken in this aspect of metrical phonology—probably because it contradicts a view of stress that he doesn't want to give up.

One could argue, of course, that demotion or promotion of this sort is an aspect of performance, not an inherent part of the text. It is true that stress shift and promotion/demotion rules are typically optional. Yet this argument is not available to Holder, because he repeatedly insists that it is the performance, whether silent or (preferably) aloud, that is the proper object of study, the written text being merely a script for performance.

He does add a hedge at one point; on p. 129 he asserts that "the words of a poem should be read in such a manner as to receive their lexical stress, or only such modification of that stress as is required by the rules of English phonology." This is the first and only real acknowledgement that stress can be modified by phonological context. Unfortunately, he seldom integrates this insight into his scansions.

There are several other places where Holder's lack of familiarity with contemporary linguistics creates problems. For example, when he criticizes traditional prosodists for ignoring phrasing in favor of foot-divisions he merely overstates what is basically a valid point—but he also entirely ignores phonologists like Bruce Hayes and Paul Kiparsky, who depend heavily on phrase boundaries and pauses in their formulation of rules for what counts as "metrical" in the verse of Shakespeare and other poets. I believe their accounts are weak in other ways, but they certainly cannot be accused of ignoring phrasing—even though they also hang on to the concepts of the foot and of metrical positions in the line, both of which Holder would like to throw out.


Holder's limited sophistication in linguistics, paired with his resistance to multi-level theories, again handicaps him in his discussion of phrasing. He wants there to be one entity known as "the phrase," and for it to make identical divisions in terms of syntax, semantics, pause, and intonation. That is, for him a syntactic phrase will also be exactly one "sense-unit," pause-phrase, and intonational phrase. He does not acknowledge other types of phrasing, such as the "clitic phrase," at all. This point of view seems to be based mainly on his intuition rather than any empirical data. In my experience with recorded speech, these different types of phrasing tend to be congruent or coterminous with each other, but quite often do not match up. Other linguists who work extensively with real-life speech data have come to the same conclusion; Woodbury (1987), for example, outlines a multilevel definition of phrasing, with possible mismatches at every level, that I have found quite useful.

When confronted with evidence that different types of phrase are not necessarily congruent (for example, in cases of poetic enjambment), Holder is inconsistent. At times he seems to acknowledge these facts and hedge his stronger claims: ". . . we certainly have the possibility that syntactic entities and intonational ones will coincide within a given sentence" (207). At other times, he seems to try to rationalize the problem away. His shaky grasp of linguistics also contributes to a confusing discussion here. He cites Selkirk in claiming that apparent instances of syntactic and intonational units not being congruent are caused by a "Transformational Grammar" notion of the sentence.(3) The example he uses, taken from Selkirk, who takes it from elsewhere, is "This is the cat that chased the rat that ate the cheese." Here is his discussion:

    One of the bracketings sets off as a syntactic unit, in this case a verb phrase, the words "chased the rat". A similar bracketing gives us "ate the cheese." But in vocalizing the sentence one is likely to say "This is the cat," then pause, then say "that chased the rat," then pause, then say "that ate the cheese." Such a phonological rendering sets up sound units that do not observe the respective boundaries of the syntactic ph[r]ases, "chased the rat" and "ate the cheese." But there is nothing that compels us to describe this or any other sentence in transformational grammar terms in order to characterize its structure. A perfectly acceptable alternative would be to see the sentence in question as composed of a main clause, followed by two relative clauses, each of which begins with "that." The three clauses, designated by intonation phrases, would give us a perfect fit between intonation and syntax. (207)
What is he talking about? I know of no Transformational Grammar, or Phrase Structure Grammar, treatment that would not analyze this sentence as a main (or "matrix") clause with two relative clauses (though the terminology may vary a little). The difficulty, if there is one, is that the constituents are nested:
    This is [NPthe cat [RCthat __ [VPchased [NPthe rat [RCthat __ [VPate [NPthe cheese]]]]]]].
That is, "the cat that chased the rat that ate the cheese" is a noun phrase, "that chased the rat that ate the cheese" is a relative clause modifying "cat," "chased the rat that ate the cheese" is a verb phrase within the relative clause, and so on. The only problem with this is that "this is the cat," and "that chased the rat" are intonation units but not whole syntactic constituents. This is indeed a theoretical problem in the relation of syntax to phonology, but it does not seem like a very difficult one to solve—for example, by stating that the beginning of a clause is a possible place for an intonational phrase boundary, without reference to the syntactic structure of what comes before. (Selkirk resolves the problem in a slightly different but equally uncomplicated way.)

There are better examples Holder could have used to illustrate this point, especially if we look at examples of spontaneous speech instead of made-up textbook sentences. There are also a few other places in the book where he needlessly explains away conflicts with transformational grammar that are not really conflicts—for example, when he says that despite what the grammars say, he will include clauses in the category of phrases—apparently unaware that many contemporary grammars don't really distinguish between clauses and phrases in any way that would affect his theory (though they sometimes use substitute names that cover both, such as "projection").

Expectation and iambic pentameter

A large part of the book is devoted to discussions of the English iambic pentameter. Holder, since he does not believe in feet at all, does not consider the word "iambic" to be very useful in describing this verse form. He is also skeptical of critics who dwell overly much on reader's "expectations" that a line will match the iambic pentameter model. This skepticism encompasses not only the widely-discussed concept of "expectation" (also known as "metrical set") but also the familiar idea that readers compare each actual line to a "Platonic ideal" line as they read, and that the difference between the actual and the ideal produces "counterpoint" or "tension."

There seem to be two main reasons that Holder doesn't agree with these ideas. One is that he finds the notion of a "Platonic ideal" too complex; how, he implicitly asks, could a reader be hearing two metrical patterns at once? The second reason is that he feels that the stress patterns of poetry in this tradition vary too widely for "iambic" to be a very accurate description.

Much of the argument here seems to center on differing interpretation of the key terms. For example, when theorists use the term "ideal" to describe an absolutely regular iambic line, Holder seems to interpret it as meaning "maximally desirable". Yet, as Holder acknowledges, most critics do not see perfect regularity as desirable; they prefer some variation. He quotes Mark Liddell, "There must be something wrong with an aesthetic system whose norms exist only to be violated" (57). But if we take "ideal" in the sense of "the guiding idea," the conflict is diminished if not resolved altogether. If we followed Liddell's logic in all the arts, we would have to say that photorealism is the supreme form of painting, but that photographs are superior to any painting; that death-masks are superior to sculpture; and that tonal music should be without syncopation or dissonance—since in each case there is a background pattern that the artist "fails" to match exactly.

Perhaps "framework" would be a less loaded term than "ideal" here. Or the terms "marked case" and unmarked case," though they may smack of linguistic jargon, would be really useful for this discussion.(4)

Similarly, the word "expectation" can be taken in a general or a specific way. A number of critics have argued that part of the effect of a known form like the iambic pentameter derives from the reader expecting a certain pattern, and having that expectation satisfied or frustrated. Holder, in opposing this view, cites a number of estimates showing that allegedly "iambic" poetry has a large proportion of lines that are not composed of all iambs. How, he asks, can readers be led to "expect" iambic lines if the lines they have already read are so inconsistent? But of course we do not expect perfect regularity. All that is necessary to produce an expectation is that we recognize a recurring pattern—we start to look for its next occurrence. Since in the estimates Holder cites, between 40% and 75% of all lines in "iambic" poetry are perfectly regular (or at least have five full stresses), and we can assume that many of the remainder are imperfectly but primarily iambic, it seems to me it would take a pretty oblivious reader not to notice this pattern and start looking for it.

Holder's redefinition of the iambic pentameter, which he renames "decasyllabic verse," tacitly acknowledges these facts. Here is my summary of it (summarized from pp. 174-5):

  • It "commits itself fairly strictly to generating ten syllables per line, though lines of nine or eleven syllables should not astonish us."
  • It allows for the "convention of syllable elision."
  • It "allocates stresses to four or five of the ten syllables."
  • It "normally" starts with an unstressed syllable.
  • It "normally" uses "at least one unstressed syllable to separate stressed syllables." (Presumably he means one between each pair of stresses, not one per line.)
And here is how we are to scan and analyze decasyllabic verse:
  • One syllable words are "assigned a stress or not in accordance with rhetorical and semantic considerations."
  • Two or more syllable words are accorded "natural emphasis". (As near as I can figure it out, this usually means one lexical stress per major category word—noun, adjective, verb, or adverb.)
  • "Deviations from the norm as hitherto defined are to be examined for possible expressive significance."
  • We also identify phrases within the line and compare this phrasing with other lines. (Holder typographically represents these with extra spaces between phrases.)
For the most part this is a sensible and accurate description. As Holder points out, it recategorizes many common types of lines as "regular" that would have to be considered "variants" in a more traditional definition. However, despite renaming it "decasyllabic," he's really had to keep a covert iambic model in here anyway: "[In decasyllabic verse] we would compare each line's pattern of stress distribution with the metrical paradigm's distribution norm, looking for possibly significant deviations" (175). So there is a "distribution norm" that we hear deviations from, but somehow that is not the same as an "ideal iambic line" that we compare actual lines to. Holder repeatedly refers to prosodists' "penchant for having it both ways;" he should include himself in this group.

"By their scansions shall ye know them"

As I mentioned earlier, one of Holder's frequent rhetorical strategies is to show a scansion and/or interpretation by another critic, then to criticize it and, sometimes to offer his own alternative analysis. This strategy can be persuasive, and certainly makes the book more interesting to read. However, it also invites us to examine both versions for ourselves, and when we do, we find that some of Holder's own analyses are just as forced as the ones he criticizes. I found far more nitpicky examples than anyone would want to read in this review, but I will offer a few of the most telling.

For one example, after having picked apart Wright's scansion of Shakespeare's Sonnet 116, Holder offers his own. But if Wright is guilty of "distorting" natural stress patterns to make it fit the meter, Holder may be guilty of doing the same thing to avoid looking metrical. Here are his scansions of three (nonconsecutive) lines:

The first of these examples is plausible, but "Let" could also easily be stressed, which would give it the regulation five stresses instead of four. (However, he needs the first two words to be unstressed for a later point, when he contrasts them with "O no" in line 5.) The third example (line 8) follows Holder's rules, but misses an important point: the second syllable of "although" is clearly stronger than the first, and also stronger than the word that follows it. And if we mark "-though" with a stress, the line is perfectly iambic. Finally, the middle example is the most bizarre; I can't see any justification either for stressing the word "it" (which, as Holder points out, is one of the most inherently unstressable words in the language) or for putting a pause between "it" and "alteration". Holder's long rationalization of his decision, based on the fact that the syntax (word order) of the clause is inverted, is unconvincing. My guess is that, consciously or unconsciously, Holder puts the stress on "it" to avoid having a secondary stress on the first syllable of "alteration," which would make the line much more regular (even more so, if we also promote "when"):
In fact, we might say that Holder "alters when he alternation finds"  (alternation of stressed and unstressed syllables, that is).

He goes on to base an interpretive claim on this odd scansion: "the hitch in line 3 might be said to register a sense of psychological disturbance on the part of the 'love'. . . ." The psychological disturbance may be there (though the lines seem to me much more an expression of conventional wisdom than of strong emotion), but object-verb inversion is so common in Renaissance poetry that the claim seems weak. It's just not true that inverted syntax always signals psychological disturbance. Also, we should remember why inversions are so common: overwhelmingly, they operate to make the meter more regular (assuming we believe in meter) or to get a rhyme to the end of a line.

Another example of a problem scansion comes from the chapter on "The Haunting of Free Verse" (111-112). Holder is rebutting a claim by Paul Fussell about Walt Whitman's poem "When I Heard the Learn'd Astronomer."  Fussell points out that the last line, "looked up in perfect silence at the stars," is regularly iambic and says that it reveals "the ghost meter which the poem has been concealing all along." Holder does not question Fussell's scansion of the last line, but comments sarcastically: "I might note that the poem's third line—'When I was shown the charts and diagrams, to add, divide, and measure them'—with its eighteen syllables and, by my count, seven or eight stresses, does a supremely successful job of such concealment."

However, this line can be used against Holder's position as easily as for it. If you allow secondary stress and metrical promotions of weaker syllables, you get a pretty regular line:

Most clearly, "-grams" should have secondary stress, since it does not have lexical stress but is also unreduced, unlike the middle syllable of "diagrams." "I" is naturally weak, but is a good candidate for promotion; like "grams" it has an unreduced vowel sound and comes between other syllables that are weaker than it is. In any case, most commentators—both literary and linguistic—have acknowledged that sequences of three unstressed syllables, with the second of the three in an even-numbered position in the line (where we would expect a stress, if we believe in expectation) fit easily into iambic lines. Some readers would even give metrical stress to "them" because of its place in the line, but because it is likely to be encliticized (weakly pronounced, with a 'schwa' vowel, and attached to the end of the verb), I would not. Thus, with the probable exception of the last word or foot, we have an almost perfectly iambic line, and the part before the comma is a plausible pentameter. (In fact, the line is even more patterned than that.)

On the other hand, I can't see how Holder could get his count of "seven or eight stresses" without violating his own principles. The only automatic, primary lexical stresses are on "shown," "charts," "di-," "add," "-vide," and "mea-." This only adds up to six. Holder must be implicitly allowing himself at least to recognize a secondary stress (to make 7), and also to promote a weak syllable (to make 8)—even though he accuses other critics of "distorting" the natural pattern when they do the same thing. And if he does allow eight stresses, assuming they are the ones I have identified, then the line would strike me as overwhelmingly iambic.

Another possible approach to this line would avoid promotion and demotion, but Holder apparently does not sanction this approach either. That would be simply to discount as "extrametrical" any extra unstressed syllables at the beginning and end of the line, making them "anacrusis," and "hypercatalexis" respectively. Holder never mentions these possibilities, so I assume he disagrees with them. In fact, he does not even appear to believe in the traditional "feminine ending." From one point of view these devices do seem like a way of fudging a line to make it seem more metrical than it is; however, phoneticians have found these concepts (especially anacrusis) useful even for speech that is not poetry. Speakers do measurably give less time and weight to such syllables, which might be an argument for excluding them from the main rhythmic description. Allowing anacrusis would make the two halves of the Whitman line symmetrical except for extra unstressed syllables begining the first half: (?)/x/x/xx, (?)/x/x/xx (where "(?)" stands for any number of unstressed syllables after a pause).

To sum up this topic: Holder writes, "What we want is scansion without presupposition" (127). I am not convinced that this is possible. We inevitably bring one set of presuppositions or another with us.

"Expressive" possibilities

Along with citing dubious and not-so-dubious scansions by other critics, Holder also frequently criticizes literary interpretations based on these scansions, especially those based on the idea of "expressive variation." However, as his principles for "decasyllabic verse" above show, he is not opposed to this type of interpretation as a whole—in fact he is dismissive of theorists who deal only with form and do not offer a way to connect form to meaning. I agree with him that such links, when they are persuasive, can be very interesting and even enlightening. At the same time I must say that my own experience, both in reading and in writing prosodic commentaries, is that the attempt to link rhythmic or metrical form to meaning is fraught with the danger of arbitrary claims and self-fulfilling prophecies. The targets of Holder's criticism often fall prey to these tendencies, but unfortunately he himself does too, at times. As examples, let me take one interpretation based on stress and one based on phrasing; I will offer one based on intonation later.

In the first example, Holder criticizes Annie Finch's claims about traditional metrical feet in these lines from Whitman: "No poem proud, I chanting bring to thee, nor mastery's rapturous verse/ But a cluster containing night's darkness and blood-dripping wounds...". Finch calls the beginning of the first line "an embedded iambic pentameter," and says, "The triple feet in the second part of the [first] line reject the 'rapturous' dactylic rhythm as well, but only to pick it up at even greater length in the second line, which evokes the vagueness, darkness and mysterious power typical in Whitman's dactylic lines."

Holder sarcastically comments, "After the alleged 'iambic pentameter' in the opening of the first line, all of two "triple feet," that is, dactyls, presumably manage, at one and the same time, to establish a rapturous rhythm and to reject it." He also questions Finch's dactylic reading of the second line: "If there is a rapturous dactylic rhythm intended here, Whitman has botched the job." Instead, Whitman "uses his free verse in the second line to dramatize the notion of a 'cluster' by clustering stresses, as follows":

This is an odd argument. I agree with Holder that Finch sometimes goes overboard in scrounging up evidence to fit her notion of "the ghost of meter," but I can't see that he's improved the situation much. He seems to be involved in a very similar operation to Finch's, except without reference to the term "feet." If "all of two dactyls" is not enough to suggest a dactylic rhythm, how can "all of two" stress clusters be enough to mark the idea of clusters? And if Whitman was not, consciously or unconsciously, trying for an iambic rhythm at the beginning, why did he move "poem" before "proud" and "chanting" before "bring"? ("Proud poem" and "bring chanting" would, after all, double the number of stress clusters dramatizing the idea of "clusters.")

The second line brings up a more subtle, but even more important point. Holder correctly points out that in order to read it as a series of dactyls (or, better, anapests) we must subordinate the stress of "night's" and "drip-", both of which are likely to be stressed; he argues that this is an arbitrary distortion just to get the meter to work. But I would call it a simplification rather than a distortion. Phrasal stress and compound stress work differently: a head noun, such as "darkness," normally receives stronger stress than a possessive preceding it, while the first element of a compound like "blood-dripping" receives stronger stress than the second element. Thus both Finch and Holder simplify the line for their own purposes: Finch's scansion ignores the difference between weakly stressed syllables and unstressed ones, while Holder's ignores the difference between weaker and stronger stresses. And both attach imaginable but unconvincing meanings to those scansions.

An example of interpretation based on phrasing comes in Holder's critique of Roger Mitchell's explication of these lines from Alexander Pope's "Second Pastoral":

Where-e'er you walk,// cool Gales shall fan the Glade,
Trees,// where you sit,// shall crowd into a Shade,
Where-e'er you tread,// the blushing Flow'rs shall rise,
And all things flourish where you turn your Eyes.
The double slashes are Mitchell's, and are meant to show phrase boundaries. Holder finds no problems with these divisions, but does argue with what Mitchell makes of them. Mitchell says that the count of phrases per line, which goes 2-3-2-1, parallels "the emotion of the speaker whose agitation over the nymph rises in the second line and then slowly disappears..." . Holder counters that line 2 is "initially choppy" not to indicate the speaker's "rising agitation," but to show
something quite different, a posture that is at once genuinely complimentary and playful, producing a marvelous picture of trees responding to the presence of the "you" by bestirring themselves, pressing together to make a shady shelter..." . Also, the line's initial impedance by the two commas helps create, through contrast, a sense of quickening movement through the rest of the line, appropriate to the intense, if fanciful, action in describes: "shall crowd into a shade."
Holder's reading of the poem is persuasive enough, but his linking that reading to the phrase boundaries seems suspiciously ex post facto—that is, he decides what the line means, and then decides that the phrasing of the line supports that meaning. Line 2 may be the most original image, but it's hard to see that its action is any more fanciful than those in the other lines; all involve natural forces changing themselves to pay tribute to the nymph. Instead, it seems more likely that Pope changes the word order in lines 2 and 4 to make the rhythm livelier and less monotonous, and to lighten the rhetorical overkill that would be created by starting four lines in a row with exactly the same type of clause.

Phrasing and phrasalism

I have already discussed some issues in Holder's apprach to phrasing; some of these are rather technical and theoretical. Another aspect of phrasing which is of great importance but not at all hard to grasp is the relation of the phrase to the line. The basic question is this: is the end of a line automatically also the end of a phrase? There is simply no widespread consensus on this, among poets, critics, or prosodists. Some people reading poetry aloud consistently pause, others don't. However, Holder solves the problem by fiat: "[although] I generally hold that prescriptive statements are best resisted in thinking about prosody, I have to. . . insist that a pause be registered at the end of the line and that it be actual, however slight" (148).

This seems like the Gordian knot solution, but in fact it is probably the best one from Holder's point of view. It frees him to go on and discuss other aspects of the importance of lineation, which he rightly says is a crucial, even definitive, aspect of verse structure (especially in free verse). Holder is not one to let a false loyalty to descriptivism prevent him from making strong statements.


The book's final chapter, on intonation (pitch patterns) in poetry is its most original and most important contribution. As Holder points out, the subject has historically been underrepresented in prosodic theory and criticism. There are at least two reasons for this scarcity. First, English intonation has historically not been well understood by linguists (or anyone else), and linguists still have major disagreements both on how it should be represented and on how it functions in the language. Second, intonation is a feature of speech but not of writing, and as long as prosodists focus on written texts rather than performances, it is very hard to say what the intonation of a given poem is. The same poem could be performed with a number of different intonational melodies, depending on interpretation, context, and personal style. Holder acknowledges both of these problems, but argues that they are not insurmountable and proceeds to give some examples of intonational criticism of specific poems. Though I do not agree with all of his readings, they are important as models: he is lucid both about what he thinks the intonational patterns are and about how they can be used in explication.

As Holder argues, intonation is an important part of poetic language, and the fact that it is tricky to deal with does not justify ignoring it. It is important for several reasons. Holder focuses on its pragmatic and interactive properties: "Intonation is inseparable from. . . [the] speaker's attitude toward his or her subject matter and/or toward the audience;" it can also "mark focus or emphasis." He points out that attitude, in the form of "tone," is a standard literary critical issue—but is usually not discussed in terms of intonation. There are other intriguing aspects to intonation, as well: Bolinger notes that up and down pitch is often tied to up and down metaphors; Selkirk notes that English is unusually rich in the variety of its pitch accents, and therefore at least potentially has rich expressive possibilities (having many possible melodies may lead to many possible meanings).

The theory of intonation that Holder adopts for this purpose is taken more or less entire from Bolinger, Intonation and its Parts. This is not a bad choice, for reasons I'll outline in a moment. However, it is a less inevitable choice than Holder makes it seem. In his survey of the field, he again relies too heavily on Brogan's bibliography. He writes that intonation has been neglected in linguistics as a whole, just as it has in poetic prosody. The only intonation scholars he discusses in any detail at all are Bolinger and Crystal, and the emphasis is heavily on Bolinger. One gets the impression that these two are lone wolves, bucking the intellectual currents of the time. Certainly they are both respected authorities, and perhaps they were once lonelier than they now seem. However, there has been a virtual explosion of work on intonation in the last twenty years, spurred on partly by the improvements in computerized acoustical analysis and partly by input from the study of intonation in  tone langauges—just as a previous wave of research was prompted a couple of generations earlier by structural linguistic methods and by improvements in analog devices such as the sound spectrograph. Scholars like Mark Liberman, William Leben, Janet Pierrehumbert, Mary Beckman, Gail Ayres, Julia Hirschberg, Gregory Ward, Alan Cruttenden, Johan 't Hart, Rene Collier, David Brazil, Gillian Brown, Dafydd Gibbon, M.A.K. Halliday, and many others (I hope no one will take offense at having been left out) have been producing useful work in this field. Holder seems largely unaware of these developments.

Despite this lapse, though, Bolinger's notation and theory is a good choice for a book like this one, because it is, at least on the surface, simple, and it does not require much expert knowledge to understand his graphics and his claims. Intonational contours are represented with ordinary typeface, arranged on the page to show its "highness" or "lowness". The infinite variety of possible melodies is reduced to three basic pitch "profiles," labelled "A," "B," and "C." This apparent simplicity can be deceptive in some ways, but certainly presents a more inviting face to the non-specialist reader than most competing systems.

If Holder's solution to the problem of disagreements among linguists is simply to pick a linguist and stay with him, his solution to the other problem I mentioned—the fact that written texts do not specify any particular intonation—is similarly straightforward. On the one hand, he argues that intonation is not really as indeterminate as has sometimes been claimed. (He has a point, but again overstates the case; there's more individual and dialectal variation than he acknowledges.) On the other hand, his practical, rather than theoretical, solution is simply to use his own intuitive judgements of what the intonational pattern for a given passage should be. He does not provide a set of rules for translating lines into melodies; nor does he rely on experimental data; instead, he provides a melody and asserts it is the right one, or at least a right one. In effect, he is analyzing a performance—his own—rather than a text. This is entirely consistent with what he has said earlier in defense of performance.

Having set up a method, Holder goes on to present some intonation patterns in poetic pasages; he argues that these patterns add an element that metrical or semantic analysis would miss. Some of them are quite interesting. For example, he discusses King Lear's speech after Cordelia's death: "Why should a dog, a horse, a rat, have life,/ And thou no breath at all? Thou'lt come no more,/ Never, never, never, never, never" (V, iii, 306-8). Holder notes that it has been pointed out before that the last of these lines "acquires its force in part by reversing the iambic norm, giving us a series of five 'trochees'." But he finds it more striking that the line is also "a series of five A profiles," which in Bolinger's notation would be:

Holder notes that the A profile is associated with separateness (or terminality) and with assertiveness; these properties make sense here. Furthermore, this line contrasts with the first line of the passage, which has, in his reading, "three sharply defined B profiles:"
The B profile is more or less the opposite of the A profile; it is associated with connectedness, incompletion. So Holder sees a contrast between the "interchangeable animals that still live. . . [and] the irrecoverability of the unique Cordelia."

Holder is mainly interested in making semantic contrasts or parallels like this one; he does not make much of an argument for paying attention to intonation patterns purely for the pleasure of their sonic structure. I believe this aspect deserves more attention.(5) Still, he does find many more interesting examples like the one above, and this is an important contribution.

On the other hand, there are some serious drawbacks to Holder's method. For one thing, these intonation patterns seem to be based mostly or entirely on introspection, not on listening to actual performances produced for other purposes. Holder generally has a sensitive ear for intonation, but he produces several readings that I find extremely dubious. Second, the emphasis on intonational meanings instead of melodies again produces some of the forced links we saw in his interpretations of meter and phrasing. Finally, there are really more than three possible types of contour, as both Bolinger and Holder ultimately have to tacitly admit. This over-simplicity makes us label patterns that are meaningfully different as the same profile, and at the same time allows enough ambiguity that the same contour can be labelled two or more different ways.

Some context is generally necessary to judge a reading of intonation, and so I will have to limit the number and complexity of the examples that follow. I hope that one example of each type will give the general idea.

An example of the first type, implausible intonation, is again from Shakespeare's Sonnet 116: "O no, it is an ever-fixed mark." Holder calls the first phrase a "B profile" (rising), or actually an exaggerated example of it (in Crystal's terms, a "high booster")(222):

But if you try actually saying the line this way, you get an odd effect; instead of a confident assertion, you get a rhetorical question: "Oh, no? It is an ever fixed mark!" This would probably only occur in dialogue, as a rebuttal—for example, as the response to: "It's not a fixed mark, you know." A more likely reading would be as an A+A contour, with both "O" and "no" receiving high accents that tail off downwards. Having misidentified (in my view) the contour, he goes on to draw interpretive inferences from it, and these are inevitably on shaky ground.

Another example of a shaky interpretation comes from the same poem, in the phrase "Love's not Time's fool." Holder calls this a B+A (rising-falling) contour:

Though this is not the only possible reading, it seems plausible. Holder goes on to quote Bolinger that B+A contours often represent "comments that the audience is expected to take at face value" or else "talking to oneself." Holder remarks, "Both of these observations can be seen as applicable in the present instance. The speaker is declaring something that he apparently expects to go unchallenged, and at the same time, . . . may be engaged in a form of self-address. . .". This all sounds possible, but these are remarks that apply to the whole poem, not just these four words—they do not explain why this pattern on these four words. Furthermore, since more than one contour is possible here, it is necessary for the performer to decide what he or she means before assigning the words an intonational pattern—so using the intonation to help with interpretation seems a bit circular here. (This circularity would, however, vanish if the performer and the prosodist were not the same person.)

Finally, the labelling problem. Of course if two readers disagree about the actual pitches of a line they are also likely to disagree about which pitch "profiles" it contains, but it also happens that we can agree on the pitch and yet disagree on the labelling. This is one of the basic problems of intonational analysis: pitch comes to us in a nearly continuous and infinitely variable stream; how do we sort it into similar and dissimilar units? An example comes from Robert Frost's "Acquainted with the Night."

I wouldn't read the line with this pitch outline, but let's say this is how Holder would actually perform it. He labels it a B+B contour and goes on to discuss its interpretation. But it looks like an A+A or perhaps A+B contour to me; "me back" has the characteristic high accent followed by a jump down that defines the A profile, and though "by" as diagrammed here looks like a B profile (rise), that is an artifact of the fact that printing is done in straight lines; the most plausible readings I could produce all tailed off in pitch after the initial rise on "by." Another possibility would be that it would be a high level tone, but that is very different from high and rising, and the transcription does not show that difference. In any case, these radical differences in labelling the same pattern underscore the vagueness of these labels, and therefore of any interpretive inferences we can draw from them.

Many of these problems are inherent in the subject, but they could be lessened. For one thing, tapes are available of actors or the poets themselves reading all of these poems. If the investigator and the performer were two different people, a lot of circularity could be avoided. With live readers, experimentation would even be possible. Second, though we must resist the temptation to trust our tools instead of our minds, the technology to analyze pitch electronically is now very sophisticated. It would be a good idea for anyone entering the field of intonation at least to look at some actual F0 graphs to get an idea of the physical (as opposed to psychological) reality of pitch in speech. Computers also offer possibilities for testing impressions and even for synthesizing experiments. Some experience of this sort could help avoid producing unlikely contours.


Rethinking Meter is flawed in several ways. These flaws, however, are not so serious as to negate the value of the book. Anyone looking for a critical review of the literature on meter, or for examples of how criticism based on phrasing and intonation can work, would do well to take a look at it.


Bolinger, Dwight. Intonation and its Parts:Melody in Spoken English. Palo Alto: Stanford University Press, 1986.

Brogan, T.V.F. English Versification, 1570-1980: A Reference Guide with a Global Appendix. Baltimore: Johns Hopkins UP, 1981.

Kiparsky, Paul, and Gilbert Youmans, eds. Phonetics and Phonology, Vol. 1. Rhythm and Meter. San Diego: Academic P, 1989.

Selkirk, Elisabeth O. Phonology and Syntax: The Relation Between Sound and Structure. Cambridge, Mass: MIT Press, 1984.

Woodbury, Anthony. "Rhetorical Structure in a Central Alaskan Yup'ik Eskimo Traditional Narrative."
In Joel Sherzer and Anthony Woodbury, eds., Native American Discourse: Poetics and Rhetoric. Cambridge: Cambridge University Press, 1987.

1. Holder gives his approval to this quote from Mitchell: "The most reliable way of identifying groups [i.e. phrases] is through the pauses which sensible reading forces upon us."

2. See, for example, her "Principle of Rhythmic Alternation" and her rules for "demibeat addition."

3. He really means a "Phrase Structure" notion; a representation of internal sentence structure by trees or brackets. Transformations do not enter into this argument.

4. For non-linguists: the "unmarked case" of any phenomenon is the most basic, widely applicable, unsurprising version. "Marked cases" are the less common or less predictable ones. Simple examples would be lion (unmarked: all lions) vs. lioness (marked: only adult females), or subject-verb-object word order (unmarked in English) vs. subject-object-verb word order (marked in English: used only in special cases).

5. Holder's reluctance to talk about sound in itself goes along with an extended argument he is making against analogies between music and poetry, and in particular against identifying intonation with melody. Even though intonation is quite literally the shape of pitch in time, Holder emphasizes the differences between intonation in speech and melody in music (by which he clearly means tonal, classical, Western music). Here again he greatly overstates the case. Speaking is not the same thing as singing, but that does not mean that speech does not have melodic properties that can be interesting (and pleasing, or displeasing). Nor is his citation of neurological evidence convincing; the evidence in question does not directly address the question of whether speech has melody, or whether the brain can process musical features and linguistic features simultaneously.