Phonetic Detail in the Perception of Ethnic Varieties of US English 293salient feature for Pattani Malay speakers (Abramson 1991). This ¿ nding sug-
gests that part of what is commonly thought to be our “phonetic knowledge”
is the learned ability to follow particular perceptual cues.
The issues involved with the non-linear and seemingly non-transparent
association between articulation, acoustics and perception can be framed by
two sets of research questions having to do with the region between speech
act and listeners’ reactions. The ¿ rst set of questions pertains to issues that are
external to this chapter but central to the ¿ eld of historical linguistics, such
as why and how this relationship developed. This is the evaluation problem,
the embedding problem and the actuation problem of Weinreich et al. (1968).
Understanding the phonetic link between utterance and reaction (Figure 13.1)
requires answers to the following set of questions:
- What acoustic cues overlap with other cues? Are there latent factors
underlying the set of measurements? - What acoustic characteristics are or are not members of the set of sig-
ni¿ cant perceptual cues? - Within the set, which of the signi¿ cant perceptual cues are stronger
than others? - Is there variation even in the acoustic-perceptual relation, i.e., does the
selection of cues, or the weighting of cues, vary by language, dialect,
accent, etc.?
The main goal of this chapter is to address these four questions and dem-
onstrate variation in the region between speech act and action. The presenta-
tion of the data sets, all previously reported, uses statistical methods, which
emphasize potential many-to-many associations between acoustic character-
istic and perceptual cue. As such, it contributes answers to questions about
variability in the assignment of acoustic characteristics to perceptual cues,
highlighting differences between a “perceptual” analysis using acoustic data
alone and a “perceptual” analysis linking the two kinds of data.
- Example 1: Wisconsin English
3.1 Background
Wisconsin English provides an interesting example of how an ethnically af¿ l-
iated dialect can inÀ uence the surrounding dominant varieties.^6 A series of
studies compared recordings of Wisconsin residents from four different eras