Latent consonant harmony in Russian: Experimental evidence for agreement by correspondence

It has been recently proposed that phonological constraints enforcing consonant harmony (long-distance consonant assimilation) are grounded in functional exigencies of speech production. Specifically, the Agreement by Correspondence approach (Walker 2000; Hansson 2001, 2007; Rose & Walker 2004) hypothesizes that patterns of consonant harmony originate in difficulties at the level of phonological planning and phonetic implementation of featurally similar consonants. One interesting prediction of this approach is that harmony-like patterns may arise spontaneously, under certain conditions, even in languages that do not exhibit consonant harmony as a phonological process. In this study we test this prediction experimentally, by examining patterns of errors involving sibilant fricatives in Russian, a language that does not exhibit consonant harmony as a phonological process