Pharao Frequency of use and the Apparent Time Paradigm Nicolai Pharao University Department of Nordic Studies and Linguistics, University of Copenhagen, Denmark; nicolaip@hum.ku.dk On-going sound change has almost exclusively been studied through the Apparent Time Paradigm, in which sound change is charted by analyzing the phonetic variation of speakers of different age groups and inferring from such differences the course of an on-going change in the realization of particular phonemes. Most of these changes have been studied at the level of individual segments, e.g. /t,d/ deletion, /a/ raising, and /u/ fronting, regardless of which word they occurred in, while still factoring in phonological context and phonotactic or morphological constraints (e.g. position in the syllable or position in stem or affix; cf. Raymond, Dautricourt & Hume, 2006). In addition to this, some phonetic variation has been observed to be specific for a sub-group of all the words containing the conditioning environment, a mechanism known as Lexical Diffusion (cf. Wang, 1977; Yaeger-Dror, 1994). In a usage based approach to sound change, e.g., (Bybee, 2001), most sound changes are seen to be gradual in the sense that they do not apply across the board to every word containing the conditioning environment, but start in high frequency words and proceed gradually through the lexicon; i.e., most sound changes are diffused lexically. This hypothesis has implications for the study of sound change through comparison of samples of speakers of different age groups. An allophone that is hypothesized to be participating in an on-going change should be attested only in words of relatively high frequency in the older generation as compared to its pattern of distribution in a sample taken from younger generations in the same speech community. That is, the new allophone should be attested in a broader range of words in the younger generation than in the older generation. Furthermore, if frequency of use is operative in the diffusion of a sound change across generations, we can also ask about the role of frequency in the diffusion of a sound change in individual lexicons over time. If individual lexicons are affected by frequency of use in the manner outlined in (Bybee, 2001), the innovative allophone should become relatively more frequent in items of lower frequency later in life. The present study investigates the role of frequency in the phonologies of individuals and groups through the use of a sub-corpus of a Real Time study currently being conducted of spoken Danish. Real Time studies involve re-recording and re-analysis of the behavior of subjects from an original Apparent Time sample (see Sankoff, 2006, for a review of the literature on Real Time studies.) This enables an investigation of the stability of phonetic behavior across the lifespan. The sub-corpus contains auditory analyses of the variation in the realizations of /ɛ/, /æ/ and /ɑ/ in different phonetic contexts in the speech of 5 different subjects who were originally interviewed in the 1980s and have been re-interviewed in 2005. The interviews, ranging from 30 minutes to 2 hours in length, have been analyzed in their entirety, including classification of every stressed realization of the phonemes. This enables a test of the role of frequency at the time of original recording through the comparison of the range of frequencies of words containing standard and innovative allophones of the phonemes hypothesized to be changing. The role of frequency in possible lifespan changes can also be studied by investigating whether incipient sound changes at the time of the original recording have spread only to words of lower frequency in the speech of individual speakers. Should individual speakers prove to have adopted innovative allophones in lexical items of lower frequency later in their lives, this would lend support for the role of frequency in the continual shaping of the individual lexicon. If the occurrence of the innovative allophones do not spread to words of lower frequency in the re-recordings of the informants, this would suggest that frequency and hence the mechanism of lexical diffusion is operative at the level of groups of speakers only, and that, while the phonological grammars of individuals may be gradient, lexical conditions on the distribution of phonetic innovations would seem to fossilize some time before adulthood. Laboratory Phonology 11 111 LabPhon11 abstracts edited by Paul Warren Wellington, New Zealand 30 June - 2 July 2008 Abstract accepted after review