The sound strings of all spoken languages alternate between consonants and vowels (CVCV…), so it is natural for the naive observer to assume that all languages also have syllables. But phonologists do not want to posit superfluous categories, so they need to be convinced that there are some generalizations that need to be stated in terms of a category “syllable”. Chomsky & Halle (1968) thought that they could describe English without positing a syllable, but its seems that since [Bybee] Hooper 1972, everyone has assumed that syllables play a role in phonology – not only in English, but probably in all languages. Especially phonotactic constraints such as consonant cooccurrence are much easier to state if reference can be made to a syllable constituent. But are all languages like English? Well, most probably are, but Hyman (1983) noted that in the Nigerian language Gokana, syllable constituent is not required for any kind of phonological generalization. “Prosodic stems” have the shape CV(V)(C)(V)(V) (e.g. té, bèè, búl, kávà, bùùrù, kùmìè, ʔoòà, bèèàɛ̀), and no reference to the syllable is necessary to describe the phonotactically possible patterns.
So does this mean that we have an exception to the universal that “all languages have syllables”? In a recent paper revisiting the 1983 issue, Hyman (2011) starts by noting that the question of syllable universality is not straightforward to answer. Does the universality claim mean that no language will be found with certain observable properties (i.e. is it a descriptive-empirical claim)? Or is it an analytic claim, meaning that all languages must be analyzed with a syllable, even if such an analysis is not the only possible (or even the most natural) one?
Taking a close look at Gokana again, Hyman finds that there is limited evidence for syllable-based distributional constraints after all: In particular, if one assumes that prosodic stems can have maximally two syllables, each of which has maximally two moras, one can explain why *CVCVVV and *CVVVCV are not possible prosodic stems. Without a syllable notion, this constraint is harder to state.
One could happily conclude that a counterexample to a nice universal has been explained away, but Hyman is not so complacent: After all, the syllable-based account of Gokana is not necessary, and all that has been found is some very limited evidence. Gokana is still quite different from many other languages, in that it makes so little use of the syllable. And introducing the syllable, a very powerful concept, to do very little work is not particularly attractive, unless one assumes to begin with that syllables exist in all languages (i.e. unless one interprets the universality claim as an analytic claim). As also noted in the introductory section of Hyman (2008), linguists often have ways of fitting language data into their preferred systems, by tweaking the analysis in directions that do not seem natural, but are justified by claims about representational universals. (For example, Lowenstamm 1996 claims that the only syllable type in all languages is CV.) Hyman is not happy with this method, and says that an overarching goal of his paper is
“to demonstrate that universal claims of the sort ‘all languages have syllables’ are necessarily theory-dependent and hence less testable than often assumed. This not only explains the numerous metamorphoses and instability of the syllable concept throughout the history of phonology, but also suggests that questions of (absolute) universality should not dominate the agenda.” (p. 56)
This resonates with what has been observed in morphosyntax: Claims of the sort ‘all languages have adjectives’ (Dixon 2004) are not testable, unless one specifies the properties of adjectives in a way that makes clear claims about observable phenomena. The easiest way to do this for adjectives would be to say that they are words that need a copula in predicative function and no special relative clause marking in attributive function, but with such a restrictive definition, adjectives would clearly be non-universal.
Hyman likewise wants “to convert what might be interpreted as an analytic claim into a descriptive one”. For this, according to Hyman,
“one has to “normalize” the data according to some general standard that minimizes the differences between the interpretations that different theories accord to the data.” (Hyman 2008: 85)
But how is this done? Isn’t such a “normalization” just another theory? I would say no, but this presupposes that one distinguishes between language-particular theories and cross-linguistic generalizations. Language-particular theories (=analyses) often make use of abstract concepts (such as the syllable or the adjective) because language-internal generalizations cannot be captured otherwise. But these abstract concepts need not be the same across languages – different languages generalize in different directions. For this reason, language comparison must be based primarily on fairly concrete concepts (comparative concepts), which are of a different sort and need not play any role in language description/analysis. So even though Hyman doesn’t say so explicitly, I think that by invoking a concept “normalization” he implicitly endorses the view that language comparison requires a separate set of comparative concepts that are different from language-particular analytical categories. It’s only with such comparative concepts that one can make testable, “descriptive” (and hence interesting) claims about universals.
Dixon, R.M.W. 2004. Adjective classes in typological perspective. In R.M.W Dixon & Alexandra Y. Aikhenvald (eds.), Adjective classes: A cross-linguistic typology, 1–49. Oxford: Oxford University Press.
Hooper, Joan B[ybee]. 1972. The syllable in phonological theory. Language 48(3). 525–540.
Hyman, Larry M. 1983. Are there syllables in Gokana? In Jonathan Kaye, Hilda Koopman, Dominique Sportiche, & André Dugas (eds.), Current approaches to African linguistics, vol. 2, 171–179. Dordrecht: Foris.
Hyman, Larry M. 2008. Universals in phonology. Linguistic Review 25(1/2). 83–137.
Hyman, Larry M. 2011. Does Gokana really have no syllables? Or: what’s so great about being universal? Phonology 28(01). 55–85.
Lowenstamm, Jean. 1996. CV as the only syllable type. In Jacques Durand & Bernard Laks (eds.), Current trends in Phonology: Models and Methods, 419–441. Salford: ESRI.