AUGUST 11 2005

The story of /r/

Mona Lindau

J. Acoust. Soc. Am. 67, S27 (1980)

https://doi.org/10.1121/1.2018134

CrossMark

View Export
Online Citation

01 December 2023 08:22:35

 The effects of monotic CV competition on perception of place chusetts Institute of Technology, Cambridge, MA 02139) and
were determined in identification tasks using synthetic speech Stefanie Shattuck-Hufnagel (Department of Psychology, Cornell
stimuli. Three continua of stimuli were used: ba-da-ga, bi-di-gi, University, Ithaca, NY 14853)
and be-de-ge, with stimuli in a series varying only in F2 starting
Acoustic measurements show that when/s/or/z/are followed by
frequency. Standard three-choiceidentificationfunctions were ob-
the palatal fricative /•/, they are often produced as /•/ or /H in
tained for each continuum in quiet, and in the presence of a
casual speech, even when the palatal segmentis on the other side of a
monotic "masker" competing at a signal-to-competition ratio of
word boundary. Thus "gas shortage" can be rendered [gae•3rro•].
0 dB. The competition consistedof a clearly identified bV, dV, or
A similar phenomenon might be expected in a different palatalizing
gV exemplarfrom the test continuum.Masked identificationfunc-
environment, when word-final/s/or/z/are followed by the palatal
tions demonstrated that maskers with rising F2 transitions had no
effect, while maskers with falling F2 transitions markedly affected glide /y/, as in "miss you" or "use your". From the same
stimulus identification. The current results are similar to those ob- corpus and using the same measurement procedures reported
earlier [S. Shattuck-Hufnagel, V. W. Zue, and J. Bernstein, J.
tained previously [C. Berlin et al., J. Acoust. Soc. Am. 53, 699-709
Acoust. Soc. Am. 64, S92(A) (1978)], we examined the acoustic
(1973)] with dichotically-presented,natural CV stimuli. The present
data are consistent with the hypothesis that complex, time- consequencesof this second palatalization process. Our measure-
varying stimuli with falling F2 transitions are more "salient" ments suggestthat: (a) palataliz.ation of word-final/s, z/preceding
a /y/ occurs quite frequently; (b) when such palatalization oc-
than those with rising F2 transitions.
curs, it results in acoustic characteristics which are uniform across
the entire segment, with no residual alveolar fricative; and (c) for
3:24 some speakers, the palatalization process results in acoustic proper-
ties significantly different from/•/, suggestingan articulatory target
K8. Perceptual competitionof CV syllablesand of isolatedF2. Ann somewhere between the /s/ and /•/ in American English. [Work
K. Syrdal-Lasky(Callier Center, University of Texas at Dallas, 1966 supported by NINCDS.]
Inwood Road, Dallas, TX 75235)

Using three experimentalparadigmsmsingle-responseidentifica-

4:00
tion of dichotic and of electronically mixed monaural stimuli with
simultaneous onsets and recognition of 50 ms leading stimulus of Kll. The story of /r/. Mona Lindau (Phonetics Laboratory,
dichoticpairmperceptual competitionwas studiedbetween pairings University of California, Los Angeles, CA 90024)
within three sets of stimuli•two-formant synthetic /bae/, /dae/,
/gae/syllables, identical/pae/, /tae/,/kae/ syllables with 50 ms F1 In describing sound changes, sound patterns and alternations in
cutback, and identical isolated second formants. Tests were per- languagesone clearly needs to refer to a natural class of r sounds, or
formed at two intensitylevels suchthat the extendedvowel measured rhotics. This phonological class encompassessounds with a wide
70 dB or 90 dB SPL. Stimuli containing frequency modulated variety of manners and places of articulation, and it has been

01 December 2023 08:22:35

F2 transitions generally dominated those with F2 of unchanging suggested that its "phonetic correlate" is acoustic in nature,
frequency. Stimuli having lower initial F2 frequency generally namely a lowered F3. Acoustic properties of phonetically differ-
dominated those with higher initial F2 frequency, although this ent /r/s were investigated in languages like American English,
effect declined for 90 dB /bae/ pairings, in contrast to analogous Yoruba, French, Southern Swedish, Hausa, and Edo, using several
pairings in other stimulus sets, presumably because of greater speakers of each language. The results do not show any single
simultaneousmasking of F1 on F2 at higher intensity. Psycho- acoustic parameter underlying the phonological class of [+rhotic].
acoustic fusion [J. E. Cutting, Psychol. Rev. 83, 114-140 (1976); The third formant does not lower for all types of/r/s. There is often
B. H. Repp, J. Acoust. Soc. Am. Suppl. 1 58, S76 (1975)] occurred a decreased intensity associated with the rhotic, but this is by no
equally often regardlessof stimulusset and regardlessof dichotic means consistent. The natural class of [+rhotic] sounds is a phono-
or monaural presentation. [Work supported by NSF.] logically convenient class, but it is not controlled by any single
articulatory or acoustic correlate. Instead, this phonologicalclass is
associated with complex combinations of both articulatory and
3:36
acoustic parameters. [Work supported by NSF.]
K9. Cross-seriesadaptation with vocalic stimuli. John J. Godfrey
(The University of Texas at Dallas, Callier Center for Communi-
4:12
cation Disorders, 1966 Inwood Road, Dallas, TX 75235)

ß
The identificationboundaryshift after adaptationis now generally K12. Native language effects on the perception of liquids. K. S.
MacKain (Haskins Laboratories, New Haven, CT 06511 and Cornell
attributed to auditory contrast, rather than the operation of feature
detectors. That this contrast must be in a prephonetic form was University Medical School, Ithaca, NY), C. T. Best (Haskins
emphasizedby the demonstration that mutual adaptation may take Laboratories, New Haven, CT 06511), and W. Strange (University
place even between vocalic and consonantal segments [J. J. God-
of Minnesota, Minneapolis, MN and Haskins Laboratories, New
Haven, CT 06511)
frey, J. Acoust. Soc. Am. Suppl. 1 66, Sl14(A) (1979)]. Further
investigation shows that cross-seriesadaptation may be obtained In a study of the effect of language experience on speech per-
with brief vocalic stimuli differing in the amount of formant fre- ception, Miyawaki et al. (Percept. Psychophys. 1975) found classic
quency transitions from steady-state [e] to [et]. Cross-series categoricalperception of the liquids [r] and [1]in native speakersof
adaptationwas attempted with three adaptors: [3], [3co],and [at]. American English, but not in native Japanese,whose languagelacks
In the first two, F1 was identical to one of the test series stimuli, a liquid contrast. They tested oddity discrimination, using a syn-
but F2 was different; in [at], both F1 and F2 were different, but thetic [r]-[1] continuum containing systematic variations in the F3
the directional pattern of formant changesresembled [et]. Only the transition. Since the oddity task is believed to be very difficult, the
[at] adaptor caused a boundary shift, suggesting a fairly abstract current research investigated effects of the task requirements on
level of auditory contrast. [Work supported by NIH.] discrimination via comparison ofA x B and oddity discrimination
tests with American and Japaneseadults. Because spoken liquids
3:48
show F1 and F2 differences in addition to the F3 difference, a
, new ten-item synthetic continuum was used, which involved con-
K10. Palatalization of/s/ in American English: when is a /g/ not a current changes between [r] and [l] in all three formants, The
/g/? Victor W. Zue (Research Laboratory of Electronics, Massa- American results replicated the earlier findings. Peaks and troughs

$27 J. Acoust. Soc. Am. Suppl. 1, Vol. 67, Spring 1980 99th Meeting: Acoustical Society of America $27