David Eddington, Brigham Young University
Capitalistic versus Militaristic: Paradigm Uniformity and Analogy
In American English, the first /t/ in capitalistic is generally flapped while in militaristic it is not. This has been explained as due to the influence of capi[ ]al and mili[t ]ary. Steriade (2000) discusses the influence a base word has on the phonetic realization of a word derived from that base as Paradigm Uniformity. In her experiment, speakers who produced a flap in the base form also pronounced a flap in the derived neologism.
Riehl (2003) replicated Steriade's experiment, but found some variability between the tap and untapped varieties of /t/ between the base and derived form, which Riehl adduces as evidence against Paradigm Uniformity. However, using Riehl's data I found a very significant correlation between the phone in the base word and the phone in the derived form, which actually supports Steriade's analysis contra Riehl.
According to Steriade, the default allophone of /t/ in the capitalistic/militaristic context is the flap while the occurrence of [t ] in mili[t ]aristic is due to analogy. Davis (2003), on the other hand argues that [t ] is the default in this context and that the flap in capitalistic is caused by analogy. To support his assertion he points to the lack of flap in Mediterranean and Navratilova.
My contention is that all allophonic distribution may be explained in terms of analogy to stored linguistic experiences. This contrasts to Steriade and Davis' idea of analogy as a process that interferes with the application of general rules. In my simulations, the pronunciation of 3,738 instances of the allophones of /t/ was taken from natural speech and is taken to approximate a speaker's experience with the phoneme /t/. The pronunciation of a word is determined based on analogy to the instances in this database rather than on rules. Analogical outcomes were calculated using the Analogical Modeling algorithm (Skousen 1989).
Given this database and computer model I demonstrate how the allophones of /t/ may be assigned by reference to items stored in the mental lexicon. The influence of base forms such as capital on derived forms such as capitalistic is explained as due to the same process of analogy. Moreover, the differences between flapping and lack thereof in words such as militaristic, capitalistic, Mediterranean, and Navratilova may be attributed to analogy without assuming that analogy overrides a more general rule of flapping as David and Steriade assume. These findings are especially interesting because the simulations included only surface-apparent phonetic variables; there was no representation of morphological relationships (capital / capitalistic), nor of abstract features such as prosodic licensing, syllable boundaries, foot structure, constraints, or derivations as are commonly assumed in rule-accounts.
Davis, Stuart. 2003. Capitalistic v. militaristic: The paradigm uniformity effect reconsidered. Paradigms in phonological theory, ed. by Laura J. Downing, T. A. Hall, and Renate Raffelsiefen, 107-121.
Riehl, Anastasia K. 2003. American English flapping: Perceptual and acoustic evidence against paradigm uniformity with phonetic features. Working Papers of the Cornell Phonetics Laboratory 15.271-337.
Skousen, Royal. 1989. Analogical modeling of language. Dordrecht: Kluwer.
Steriade, Donca. 2000. Paradigm uniformity and the phonetics-phonology boundary. Papers in laboratory phonology 5, ed. by Michael Broe and Janet Pierrehumbert, 313-334. Cambridge: Cambridge UP.
