How Much Phonology in ‘Laryngeal Phonology’?

Phonetically observed voicing phenomena cannot be unambiguously and exhaustively identified with laryngeal phonology, in which phonetic properties such as voicedness or voicelessness are directly translated into phonological categories, and in which phenomena such as voicing assimilation are necessarily viewed as phonological spreading. This much became clear with the advent of privative representations, which cannot capture, for example, phonetically observable symmetrical voicing assimilations as symmetrical spreading. This paper attempts to demonstrate that the laryngeal phonology of Polish per se is very small and can be reduced to strictly privative representation with substance free categories and positional licensing as the only instance of phonological computation. This minimal phonological component allows us to understand a complex set of voicing phenomena in Polish, including positional and dialectal variation. What is required, however, is an increased role of the language specific phonetic knowledge in the process of acquisition and in phonetic interpretation. Additionally, the relation between phonetics and phonology must be viewed as arbitrary.