On the relevance of locus equations for production and perception of stop consonants

Abstract

Weexamined the possible relevance of locus equations to human production and perception of stop consonants. The orderly output constraint (OOC) of Sussman, Fruchter, and Cable (1995) claims that humans have evolved to produce speech such that F2 at consonant release and F2 at vowel midpoint are linearly related for consonants so that developing perceptual systems can form representations in an F2 0 ns -by-F2vowel space. The theory claims that this relationship described by locus equations can distinguish consonants, and that the linearity oflocus equations is captured in neural representations and is thus perceptually relevant. Weinvestigated these claims by testing how closely locus equations reflect the production and perception of stop consonants. In Experiment 1, we induced speakers to change their locus equation slope and intercept parameters systematically, but found that consonants remained distinctive in slope-by-intercept space. In Experiment 2, we presented stop-consonant syllables with their bursts removed to listeners, and compared their classification error matrices with the predictions of a model using locus equation prototypes and with those of an exemplar-based model that uses F2 0 ns and F2vowelJ but not locus equations. Both models failed to account for a large proportion of the variance in listeners' responses; the locus equation model was no better in its predictions than the exemplar model. These findings are discussed in the context of the OOC.It is well known among speech scientists that formant transitions provide listeners with important information for identifying the stop consonants Ib/, Id/, and 191 in syllable-initial position. However, it is a matter of debate how such transition information is used. In particular, the importance of these highly context-sensitive transitions to perceivers provides a challenge for theorists who claim that listeners use invariant information to identify stops (e.g., Fowler, 1986;Stevens & Blumstein, 1981). One early hypothesized invariant was a locus to which all second-formant transitions of a given consonant pointed back in time by extrapolation (Delattre, Liberman, & Cooper, 1955). Locus points were found for Ibl and Id/, but 191 required two loci; and even the loci for Ibl and Idl proved controversial. For example, Ohman (1966) found that the locus points were not invariant, and that they changed, depending on the identity ofa preceding vowel. Later, Kewley-Port (1982) made extensive acoustic measurements of formant transitions for the stop consonants in various vocalic contexts and concluded against any invariant in the transition. Transitions, therefore, appear to be inherently context sensitive. As a result, theories assuming both the importance of invariance for perception and a role for formant transitions have taken different directions. One is to assume, as motor theorists have, that This research was supported by NICHD Grant HD-OI994 to Haskins Laboratories. We would like to thank Harvey Sussman, Bjorn Lindblom, Doug Whalen, and an anonym...

Locus equations are an acoustic expression of articulator synergy

“…As for intercepts, velars in the front vowel context have a very high intercept, close to or higher than for /d/, whereas in the back vowel context, they have an intercept that is lower than /d/. These generalizations tend to be robust despite a great deal of subject variability in the exact placement of slopes and intercepts for a given idiolect, which can be seen in examining the data in several studies ͑Sussman et al, 1991;Brancazio and Fowler, 1998;Sussman et al, 1993;Sussman, 1999;Guion, 1998;Modarresi et al, 2004͒. As was done in these studies, the EMMA velars were split by whether they were in front ͑/i/ and /ei/͒ or back ͑/u/ and /a/͒.…”

The Effect of Speaking Style on a Locus Equation Characterization of Stop Place of Articulation

“…Following a procedure first described by Crowther [1994], Brancazio and Fowler [1998] artificially manipulated coarticulation in 3 speakers in an attempt to test stability of locus equation slopes/y intercepts as generalized phonetic descriptors of stop place of articulation. To increase coarticulatory overlap they encouraged speakers 'to be as far into the vowel gesture as possible at the release of the consonantal closure' [Brancazio and Fowler, 1998, p. 27].…”

“…If coarticulatory influences are minimal, then F 2 onset frequencies will show little variation across vowel contexts. Using slightly different terminology, this scenerio occurs when a consonant exhibits maximal resistance to coarticulatory effects of the following vowel [Brancazio and Fowler, 1998;Fowler, 1994;Recasens, 1985]. A locus equation plot with F 2 onsets that remain stable across vowel contexts yields a flat slope, typifying an alveolar/dental place of articulation.…”

The coarticulation/invariance scale: Mutual information as a measure of coarticulation resistance, motor synergy, and articulatory invariance

“…One method of quantifying of coarticulation resistance, defined in the acoustic domain, is provided by locus equations (LE) (Lindblom, 1963;Sussman et al, 1991;Brancazio and Fowler, 1998), which are regression lines estimated by predicting F2 of a consonant from F2 of a vowel, whereby the consonant is fixed and the vowel varies over the possible vowels of the language. The slopes and intercepts of these lines vary systematically as place of articulation of the consonant changes.…”