One of the fundamental issues in the study of speech perception is to model the cognitive process of phonetic judgment and to infer the underlying mechanism. Experimental results on categorical judgment as well as on judgment of typicality do not attest to the existence of a single prototype for each phonetic category, but rather to the existence of a continuum of exemplars, viz., a region of high typicality, whose shape is not necessarily convex, in the multidimensional space of acoustic/phonetic parameters. The neural mechanism for separating such a region from others is based on thresholds (i.e., boundaries), but not on distances from a single point of reference. On the other hand, the production of a specific phone in a given context by a single speaker displays a very sharp distribution, which indicates the existence of a speaker-specific prototype in speech production. Such an apparent asymmetry of production and perception is shown to exist also in the use of language, and can be explained in the light of a general underlying mechanism for coding and decoding the information. Implications of these results to spoken language processing will also be discussed.