The Capitalistic versus Militaristic Debate: Paradigm Uniformity Revisited

The present study examined the well-known data in which the onset of the third syllable is flapped in capitalistic whereas it is aspirated in militaristic even though both words have the same stress pattern and syllable structure. While a number of studies have considered several different approache...

Full description

Saved in:
Bibliographic Details
Published in영어학 Vol. 21; pp. 636 - 655
Main Author Jungyeon Kim
Format Journal Article
LanguageEnglish
Published 한국영어학회 2021
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:The present study examined the well-known data in which the onset of the third syllable is flapped in capitalistic whereas it is aspirated in militaristic even though both words have the same stress pattern and syllable structure. While a number of studies have considered several different approaches in this discussion including paradigm uniformity effect, foot-based analyses, optimality theoretic (OT) accounts, and analogy, there has been no research that seeks to account for the possibility that the underlying /t/’s of those two words can be realized as both aspirated and flapped by speakers of American English. This study basically follows a prosodic foot-based account to explain this phenomenon and attempts to capture the variant realizations using the audio pronunciation listed in eleven different online dictionaries within the Maximum Entropy (MaxEnt) Model, which is a probabilistic model that assigns each candidate a probability rather than picking a single winner as in standard OT. The frequency data observed from those dictionaries were fed into MaxEnt to see if the learned grammar can successfully predict the observed frequency. The current simulation results show that the frequency found in the actual linguistic data corresponds to that predicted by the training corpus data, which indicates that the learned grammar is able to accurately reproduce the observed frequency. These findings suggest that MaxEnt modeling has a more explanatory power than classical OT analyses in that it can serve to account for grammars involving free variation. KCI Citation Count: 0
ISSN:1598-1398
2586-7474
DOI:10.15738/kjell.21..202107.636