Nigel Fabb, University of Strathclyde

These very interesting comments raise some difficult issues, to which I offer partial responses as follows.

1. The definition of “poem” is taken from Fabb (What is Poetry?). It covers metrical poetry, nonmetrical but parallelistic poetry, and free verse. Its goal is to differentiate most poetry from most prose, but it has some blurred boundaries: for example it does not fit prose poems, and it holds at a small scale – ordinary prose can be divided into paragraphs, chapters etc. which are larger sections not determined by syntactic or prosodic structure.  So it is intended as a useful but not strict (or universal) definition; its purpose is to draw attention to the fact that where poetry has regular added forms such as metre, rhyme, alliteration or parallelism, these added forms are dependent on poetic sections.  Because prose lacks these kinds of section it also lacks the poetic forms.  (Furthermore, as I argue in another entry, the poetic sections on which the forms depend are short enough to fit into working memory.)   Poetic sections – such as metrical lines – can have boundaries which coincide with linguistic constituent boundaries; in some traditions, this is strictly enforced, while in others it is not.  That is, there is definitely the possibility of a regular (perhaps statistically predictable) relation between poetic sections and linguistic constituents.  The point of the definition is that the reverse is not true; that is, there is no generalization that a specific type of linguistic (syntactic or prosodic) constituent is always coincident with a poetic section; if that was the case, then people would be talking poetry whenever they spoke and there would be no prose.

2. As noted in the comment, there are apparently meters which enumerate words. Hebrew (Carmi 60-62) offers examples, and Rumsey argues that the PNG language Ku Waru has word-counting meters. If words, rather than some prosodic element inside words, are being counted in these meters then they fit the Fabb and Halle approach, which is not committed to phonology determining meter, better than the Kiparsky approach.  But also if these are just counting meters and there is no control of rhythm, then these ‘word counting’ meters are like syllable counting meters, or mora-counting meters, or Korean sijo in which accentual phrases are counted.  As such, these purely counting meters fall outside the scope of the universal, which look not at counting as such, but at the relation between counting and rhythm (in the sense of a pattern of prosodically differentiated units).

3. Lerdahl and Jackendoff showed that music has various kinds of hierarchical structure, of which (musical) metrical structure is closest to the rhythmic organization of metrical verse. However musical metrical structures are not subject to an upper bounding in length; for example, if the music is in 3/4 time, we do not in general find a kind of requirement relating to the metrical structure itself that every four bars form a distinct section.  The organization of the musical sequence into sections is Lerdahl and Jackendoff’s ‘grouping structure’, a hierarchical organization into motives, phrases and sections which is distinct from but can be systematically related to the metrical structure: they are ‘independent but interactive’. Grouping structure is one of a range of kinds of hierarchical grouping structure in cognition (as Lerdahl and Jackendoff point out, and an area in which they were pioneers; for general discussion see Cohen).   Other kinds of grouping include event segmentation (Radvansky and Zacks).  One of the characteristics of all these kinds of grouping is that there is no inherent limit on their maximal size; they are not measured or counted out.  (Though Lerdahl and Jackendoff do have a rule against very small groups.)  But metrical structures in poetry are made from a controlled number of elements with a lower and upper bound; in this, they are unique.   Note that nothing prevents musical grouping being organized numerically, or prose being organized numerically (and Hymes argued that some folk tales were organised into counted groups); but this is optional not obligatory. Grouping structure in music is a bit like division into paragraphs and chapters in a novel; it is a kind of hierarchical grouping with no inherent limit on maximal size. It is unlike lineation in metrical verse.

4. Finally, a comment on the Fabb and Halle theory of meter. This shares with all theories of meter the view that a metrical line is attached to a hierarchical metrical representation (whether tree or grid, close variants of one another), with rules which relate the prosodic phonology of the line to the metrical representation.  Theories differ in the extent to which the metrical representation is determined by the phonology of the line, or is autonomous of the phonology; the Fabb-Halle theory tends more (but not completely) to the latter.  We focus more than other theories on how the representation is built (and in our theory, also transformed). Our specific contribution is to note that the metrical representation is controlled in its size by non-linguistic principles; a meter such as iambic pentameter counts ten syllables (or five times two syllables) – but no linguistic rule can count higher then two or perhaps three.  This is the basis of the universal which I propose, which means that there are upper limits on the size of metrical sequences.  This is not inherent to the Kiparsky theories which focus on local relations between small metrical constituents such as feet and short sequences of syllables, though the setting of a limit is derived (in a different way) by Golston and Riad.

Works Cited

Carmi, T., ed. The Penguin Book of Hebrew Verse. Harmondsworth: Penguin, 1981.

Cohen, G. “Hierarchical Models in Cognition: Do They Have Psychological Reality?” European Journal of Cognitive Psychology 12 (2000): 1-36.

Fabb, Nigel. What is Poetry?  Language and Memory in the Poems of the World. Cambridge: Cambridge UP, 2015.

Fabb, Nigel and Morris Halle. Meter in Poetry: a New Theory. Cambridge: Cambridge UP, 2008.

Golston, Chris and Riad, Tomas. “The Phonology of Classical Greek Meter,”  Linguistics  38 (2000): 99–167.

Radvansky, Gabriel A. and Jeffrey M. Zacks, eds. Event Cognition. New York: Oxford University Press, 2014.

Rumsey, Alan. “Tom Yaya Kange: A metrical narrative genre from the New Guinea Highlands.”   Journal of Linguistic Anthropology  11.2 (2001):193-239.

By: Nigel Fabb, Strathclyde University, U.K.

Geoffrey Russom, Brown University

I think that the proposed universal is very much worth pursuing but I disagree with two of the related claims and would appreciate a response to my reservations by Fabb. The two claims are:

(1) A poem is a text made of language, divided into sections which are not determined by syntactic or prosodic structure. Such sections are called ‘poetic sections’.

(2) In music, regular rhythms can be sustained indefinitely over musical sequences which are not divided into sections; in other words, music can be a kind of ‘metrical prose’ (Fabb and Halle “Grouping”).

Claim #1 comes from Fabb and Halle (Meter). This publication does not confront the evidence for the reality of the metrical foot presented by Paul Kiparsky, and in more recent work based on the hypothesis “that literary language is a development of ordinary language, using the resources already available to it” (Fabb and Halle, Meter, 10). I was surprised that Fabb and Halle simply ignored this research after noting that it existed. I think excellent evidence and argumentation have been provided for the claim that metrical feet exist, that foot boundaries are aligned with word boundaries in the unmarked case, and that line boundaries are aligned in the unmarked case with the boundaries of natural syntactic constituents (sentences, clauses, and phrases, short phrases of course when a form employs short lines). An important universal proposed by Gilbert Youmans is that “higher-level metrical boundaries are progressively more significant than lower-level ones” (376). This explains, for example, why coincidence of foot boundaries with word boundaries is less strictly regulated than coincidence of line boundaries with phrase boundaries. I think the most challenging poetic material for Fabb would be Irish alliterative meters in which the enumerated constituent of the line is the word. (For a summary and references, see Russom.)

Fabb’s claim #2 is stated as if it were an unremarkable one but it seems very remarkable to me. Music in 4/4 time, as generally understood, is divided into measures, and the unmarked realization of the measure is as a group of four quarter notes with the prominence contour 1/3/2/4 (“1” being highest). The leftward boundary of the measure coincides with a prominently accented note in the unmarked case. I have not read Fabb and Halle (“Grouping”) but it is important to clarify what it meant by claim #2 for the benefit of music theorists and musicians. Fabb might want to consider the universals for musical groups in Lerdahl and Jackendoff’s Generative Theory of Tonal Music (the list of universals is on pp. 345–52).

[See also Nigel Fabb, “Response to Russom.”]

Works Cited

Fabb, Nigel and Morris Halle. “Grouping in the Stressing of Words, in Metrical Verse, and in Music.” In Language and Music as Cognitive Systems. Ed. Patrick Rebuschat, Martin Rohrmeier, John A. Hawkins, and Ian Cross. Oxford:  Oxford UP, 2012, 4–21.

Fabb, Nigel and Morris Halle. Meter in Poetry: a New Theory. Cambridge: Cambridge UP, 2008.

Kiparsky, Paul. “The Rhythmic Structure of English Verse.” Linguistic Inquiry 8 (1977): 189-247.

Lerdahl, Fred and Ray Jackendoff. A Generative Theory of Tonal Music. Cambridge, MA: MIT P, 1983.

Russom, Geoffrey. “Word Patterns and Phrase Patterns in Universalist Metrics.” In Frontiers in Comparative Prosody. Ed. Mihhail Lotman and Maria-Kristiina Lotman. Bern: Peter Lang, 2011, 337–71.

Youmans, Gilbert. “Milton’s Meter.” In Phonetics and Phonology: Rhythm and Meter, Volume 1. Ed. Paul Kiparsky and Gilbert Youmans. San Diego, CA: Academic P, 1989), 341-379.

By: Geoffrey Russom, Brown University