The Enigma Variations: Response to Sampson

Language has…manifold devices for carrying on its business of communication [and] distinctiveness lost at the phonological level might be assumed without interruption of communication by higher-level markers in morphology and syntax.

Robert King, 1967

The Chinese pattern discussed by Geoff Sampson is surely remarkable, but it’s not enigmatic, and it’s certainly not paradoxical. Morphological responses to natural phonetic tendencies are encountered quite frequently in language change. The linguistic system, with its myriad phonetic and semantic pressures effecting changes simultaneously and at times antagonistically, always emerges functionally unscathed, its semantic clarity intact. The crux of the matter is this: Sampson is casting his net too narrowly, focusing on the morpheme (very roughly, ; ), rather than the lexeme (very roughly, ; ). In the ongoing history of Mandarin, it is predominantly the lexeme, rather than the morpheme, over which the phonetic and semantic pressures on language use and structure demonstrably exert their influence.

Consider a few comparable cases:

1. 1. In certain southern American English speech communities (and increasingly, elsewhere), the lax non-low front nasal vowels ĩ, and have merged, rendering homophonous the previously heterophonic “pin” and “pen”, for example. There are well-understood phonetic reasons for a general lack of discriminability of such vowels, and so their merger makes phonetic sense. But does it make semantic sense? After all, mightn’t their merger culminate in lexical ambiguity hence listener confusion? Given the ultimate trajectory of this change, the answer is “no: listener confusion was averted.” How can we conclude this? Well, in many of these same communities, “pin” has been replaced by “stickpin” and/or “pen” has been replaced by “inkpen”. There would seem little motivation for these innovated compounds apart from their heterophonemaintaining character; otherwise, it would have to be attributed to lucky chance that these dialects innovated the compounds, whereas in other dialects, “pin” and “pen” remain contentedly ensconced in the lexicon as phonetically distinct entries. And while chance does indeed play a role in the particular sorts of speech variation that are ever-present as language is being used and re-used, it plays a far lesser role on the selectional pressures acting on this variation: the very spoken variants (chance variants) that are successfully communicated to listeners—in this case, the compounds—are also the very variants (selected variants) that are likely to be reproduced as these listeners become speakers. Successful speech propagates; failed speech dies out. The most plausible scenario, then, is this: just as the vowels in question began their natural tendency to merge, chance spoken variants that served a disambiguating function (compounds like “stickpin” and/or “inkpen”) emerged as successful, and thus began to take hold, eventually supplanting the occasionally confusing monomorphemic words. Indeed, as the compounds began to gain ground, they likely served to “free up” the sound change, as there were fewer functional barriers to its phonetic and lexical progression. None of this is likely to have happened in sequence. The present day pattern would be inexplicable if first, the vowels merged, second, rampant confusion resulted, and finally, compounding came to the rescue: languages do not suffer confusion easily. Rather, the compounding innovation likely co-evolved with the vowel merger: the semantically clear compounds were naturally selected and conventionalized, and the language maintained its semantically unambiguous structure, as it always had, and as it always will.

2. 2. In a small region of southern France, the lateral has merged towards the voiceless alveolar stop in final position. Where Standard French has bεl “pretty”, this dialect has bεt; where the standard Southern French has gαl “cock” (“chicken”), this dialect has gαt. However, these southern speakers don’t use gαt anymore. Instead, they use a variety of other local terms, including vicaire, and the word for “chick” (pul in standard Southern French, but put here). Bloomfield (1933) considers an explanation suggested by Gilliéron (1910) for this lexical shift: gαt meaning “chicken” was now homophonous with the local pronunciation for “cat”. As Bloomfield notes, the isogloss for the sound change tellingly coincides exactly with the vocabulary change. Again, due to selectional pressures on language use...