A phonetic trait of some contemporary speakers of American English is the pronunciation of initial [s] in the triconsonantal cluster –str- as a palatal [∫], so that in this change-in-progress words like strength and street sound like [∫tréŋθ] and [∫tríyt]. One correspondent has noticed it in the pronunciation of some North Carolinians and therefore tends to think of it as a dialectal feature of South/South Midland speech. Another is cautious about giving it a regional label and has taken it to be an individualism, more likely to be Southern than Northern. This trait occurs commonly in Jamaica, West Indies, when speakers of the creole patois are going up the social scale, and has also been reported in New Zealand English.

What is interesting about this phonetic development is its particularity as regards the presence of /tr/ in the cluster. Other consonantal clusters involving initial /s/ do not undergo the change, so that, for instance, words like spleen, spring, etc., are unaffected. Since /r/ and /l/ are both liquids, it would seem to be some property of /r/ and /t/ as compared to /l/ and /p/ that would likely account for the change.

In order to solve this problem, one needs to stop thinking of sound changes as narrowly grounded in the physical substance of (articulatory) phonetics, particularly when it comes to assimilation. There is no straightforward physical similarity between [∫] and [r] in American English, so even if the change of [s] to [∫] “across” [t] before [r] were reckoned to be a kind of assimilation “at a distance,” that would still leave unexplained the absence of a change in the context of [l] and [p].

A semiotic view of phonetic variation, by contrast, does provide an explanatory framework, and does so in a unitary way that fits the epistemological requirement of always seeking to reveal the isomorphism of structure between phonology and grammar. The sounds /∫/, /t/, and/r/, in English as in other languages, are marked for certain phonological features, meaning a superordinate semiotic value––relative restrictedness in conceptual scope––which is absent in /s/, /p/, and /l/, making the latter unmarked for the same features.

What the change at issue demonstrates, therefore, is the interpretative mapping of semiotic value into physical substance at the core of language. The phonetic sequence /str/ changes in the pronunciation of some speakers to /∫tr/ as a process by which the identity of /r/ and /t/ as marked sounds is instantiated. The effects of linguistic rules are, here as elsewhere, invariably the way in which language manifests––to speakers, learners, and analysts alike––the system of values that informs linguistic structure. Phonetic effects, in other words, are a diagram of phonological values. Nothing could more succinctly encapsulate the idea of phonology as semiotic.

MICHAEL SHAPIRO