speech were significantly larger than in both
adult- [F(1,10) ϭ 7.63, P Ͻ 0.005] and pet-
directed speech [F(1,10) ϭ 10.98, P Ͻ 0.001]but did not differ between pet- and adult-directed
speech [F(1,10) ϭ 0.19, P Ͼ 0.05] (see supple-mentary text).
These results show that infant- and pet-di-
Denis Burnham,1* Christine Kitamura,1 Ute´ Vollmer-Conna2
rected speech are similar and distinctly differentfrom adult-directed speech in terms of height-ened pitch and affect. Interestingly, only infant-
When talking to babies, adults invariably use a
the words cannot be understood; and (iii) vowel
special speech register characterized by elevated
hyperarticulation, which is objectified by plotting
vowels. Thus, vowel hyperarticulation does not
fundamental frequency ( pitch), exaggerated into-
first and second formant (F and F ) values of the
accompany special registers simply because they
nation contours, and high affect (1, 2). It has also
“corner” vowels, /i/, /u/, and /a/, and comparing
differ from adult speech in pitch and affect.
been found that mothers hyperarticulate vowels
the resultant vowel triangles (3).
Rather, it seems to be a didactic device: Mothers
when addressing their infants but not when
Speech samples of the mothers, all monolin-
exaggerate their vowels for their infants but not
speaking to other adults (3). This phenomenon is
gual native speakers of Australian English, were
ubiquitous, occurring across various languages—
recorded on a portable Professional Walkman
Evidently, speakers are sensitive to their au-
English, Russian, Swedish, and Japanese (3, 4)—
(Sony) with lapel microphone left with the moth-
dience, both in regard to acoustic preferences
and is thought to facilitate infants’ linguistic de-
ers in their homes. To obtain the requisite corner
and emotional needs, and in terms of potential
velopment by amplifying the phonetic character-
vowel information, we asked mothers to play
linguistic ability. We might predict that for-
istics of native language vowels (3). However,
with and name three provided toys, a “sheep,” a
eigner-directed speech would have hyperar-
the very ubiquity of this speech style means that
“shoe,” and a “shark,” in naturalistic 10- to
ticulated vowels but little elevation of emo-
it is practically impossible to obtain direct evi-
15-min interactions with each recipient. Mothers
dence of its function as a language-teaching de-
made separate recordings in their own time talk-
differences within speech registers; there may
ticulation in parrot-directed speech.
propose that (i) specialspeech registers differin their mix of acous-tic, affective, and vow-el
perceive the emotionaland linguistic needs of
Fig. 1. (A) Pitch (fundamental frequency in hertz), (B) rated affect, and (C) vowel hyperarticulation (in F1-F2 vowel space) in
infant-, pet-, and adult-directed speech (IDS, PDS, and ADS, respectively).
vice; clearly, we cannot ask caregivers not to use
ing to their 6-month-old infant, to their pet cat or
their mix of speech components accordingly.
baby-talk with infants, as it appears to be elicited
dog, and to another adult (see supplementary text
automatically. So, as the nature of the speech
References and Notes
1. A. Fernald, P. K. Kuhl, Infant Behav. Dev. 10, 279 (1987).
input cannot be changed, we decided to approach
2. C. Kitamura, D. Burnham, in Advances in Infancy
this issue from another angle— by experimental-
pitch, affect, and vowel triangles to test differ-
Research, vol. 12, C. Rovee-Collier, Ed. (Ablex, Nor-
ly manipulating the nature of the recipients.
ences in infant-, pet-, and adult-directed speech
3. P. K. Kuhl et al., Science 277, 684 (1997).
The uncanny similarity of pet- to infant-di-
(no differences were found in speech to the five
4. J. E. Andruski, P. K. Kuhl, A. Hayashi, J. Acoust. Soc.
rected speech has been noted previously (5), al-
cats and seven dogs). For pitch (Fig. 1A), infant-
Am. 105, 1095 (1999).
though no objective comparison of either pitch or
and pet-directed speech was statistically equiva-
5. K. Hirsh-Pasek, R. Treiman, J. Child Lang. 9, 229 (1982).
6. We thank V. Averkiou, A. Olley, M. Nguyen, and C.
affective speech components has been attempted.
lent [F(1,11) ϭ 0.03, P Ͼ 0.05], but pitch in
Paterson for acoustic and ratings data collection; N.
Does this similarity between pet- and infant-
speech to both infants [F(1,11) ϭ 6.58, P Ͻ 0.05]
Yates-Goozee for phonetic analyses; T. Burnham and S.
directed speech imply that vowel hyperarticula-
and pets [F(1,11) ϭ 36.52, P Ͻ 0.001] was
Vollmer-Conna for hypothesis development; and D.
Hadzi-Pavlovic, K. Bird, and J. Wolfe for statistical advice.
tion also occurs when we talk to our pets? Are we
significantly higher than pitch in speech to adults.
( perhaps unconsciously) trying to teach our ani-
Ratings of low-pass–filtered speech (see Au-
Supporting Online Material
mals how to speak or at least understand our
dio S1, S2, and S3) on five scales were factor
language? Or maybe vowel hyperarticulation is
analyzed, and scores from the resultant affect
simply a by-product of the highly emotional
factor were derived. Affect was greater in in-
Audio clips (Audio 1, Audio 2, and Audio 3)
speech we use to both our infants and pets.
fant- than in pet-directed speech [F(1,11) ϭ
To resolve this issue, we made objective com-
10.76, P Ͻ 0.01], but affect in both infant-
parisons of 12 mothers’ speech to their infant,
[F(1,11) ϭ 94.34, P Ͻ 0.001] and pet-directed
1MARCS Auditory Laboratories, University of Western
their pet, and another adult in three domains: (i)
speech [F(1,11) ϭ 54.44, P Ͻ 0.001] was high-
Sydney, Post Office Box 1797, Sydney, 1797, Australia.
pitch, which is the psychological correlate of
er than in adult-directed speech (Fig. 1B).
2School of Psychiatry, Department of Human Behaviour,
University of New South Wales, UNSW Sydney 2052,
fundamental frequency; (ii) affect, which is mea-
Vowels for infant-, pet-, and adult-directed
sured by ratings of low-pass–filtered speech, in
speech are plotted in F -F space in Fig. 1C.
*To whom correspondence should be addressed. E-
which the intonation and rhythm can be heard but
Mothers’ vowel triangle areas in infant-directed
www.sciencemag.org SCIENCE VOL 296 24 MAY 2002
TRABAJOS CIENTÍFICOS DE APORTE PUBLICADOS COMO PRIMER AUTOR Y EN COLABORACIÓN 1. Caraballo R , Fontana E., Blasi Esposito S, Michelisa B, Sgro V e Dalla Bernardina B. CARBAMACEPINA E STATO DI PUNTA-ONDA CONTINUA NEL SONNO. Bolletino Lega Italiana Epilessia. 1989; 66/77 , 379-381. 2. Caraballo R , Cersósimo R , Pociecha J, Intruvini S, Fejerman N. SÍNDROME DE WEST EN PACIENTES C
Sjöfartsverkets författningssamling SJÖFS 2000:21 Sjöfartsverkets föreskrifter och allmänna råd om sjukvård och apotek på fartyg; Sjöfartsverket föreskriver1 följande med stöd av 5 kap. 21 §, 6 kap. 13 §och 7 kap. 4 och 5 §§ fartygssäkerhetsförordningen (1988:594) och beslutarallmänna råd enligt bilaga 1 . 1 KAP. INLEDANDE BESTÄMMELSER 1 § Dessa föreskr