As listeners, we hear a speech signal that is riddled with variation.  We are exposed to words, but a single word is produced differently each time it is uttered.  These words stream by listeners at a rate of about 5 – 7 syllables per second, further complicating the listeners’ task.   How listeners map a speech signal onto meaning despite massive variation is an issue central to linguistic theory.  One problem we currently face is that the vast majority of these different realizations of words are understood equally well by listeners.  We know that listeners have at their disposal detailed and rich lexical representations.  But, as I show here, even these cannot account for a listener’s ability to take all the individual parts of a word that often vary drastically and understand them as quickly and adeptly as they do.  Given any window of speech, listeners are presented with information about sounds, sound patterns, words, speakers and their intentions, emotions, accents and other social characteristics.

In this talk, I begin with an assumption that listeners, by default, use these ever-present cues together.  I examine the perception of phonological variants sensitive to typically co-present phonetic and social cues.  For example, rather than comparing the perception of a frequent phonological variant (e.g., tap) to the perception of a less frequent member of a variant pair (e.g., [t]) embedded in controlled word-frame, I examine the perception of these variants in phonetic and social contexts in which they are typically heard by listeners. In doing so, I show that many effects attributed to phonological variants are illusory and suggest that the emphasis on the variants hides important patterns linking acoustic variation and social representations.

This emphasis also results in an interesting paradox in the literature that I call the representation paradox: both frequent and less frequent (but canonical) phonological variants facilitate perception and both frequent and less frequent phonological variants hinder perception. I briefly discuss three notions that contribute to the paradox and show that the paradox dissolves once we consider effects of acoustic variation that indexes either a particular variant or a broader speech situation.  These notions are: (1) variation is an obstacle listeners must overcome, (2) phonological variation is categorical, and (3) phonological variant effects reflect the nature of lexical representations.

I present data from three phoneme-categorization experiments designed to address this apparent paradox and show that phonological variant effects are acoustic variation effects in disguise and listeners perceive various phonological variants equally well.  Additionally, I suggest that data from these experiments supports a view in which acoustic patterns are stored along with social representations, and these representations in turn influence speech perception at a low level. This work helps explain effects of phonological variants in speech perception that are oftentimes conflicting, highlights the vast amount of processing that occurs independent of the lexicon and has implications for how linguistic units are stored and recalled by listeners.