In the public domain
Developmental Psychology
1992, Vol. 28, No. 4, 593-603
Functional Analysis of the Contents of Maternal Speech to Infants of 5 and
13 Months in Four Cultures: Argentina, France, Japan, and the United States
Marc H. Bornstein, Joseph Tal, and Charles Rahn
National Institute of Child Health and Human Development
Bethesda, Maryland
Marie-Germaine Pecheux
Celia Z. Galperin
Universidad de Belgrano, Buenos Aires, Argentina
Universite de Paris V and Centre National de la
Recherche Scientifique, Paris, France
Martine Lamour
Sueko Toda and Hiroshi Azuma
Fondations Rothschild, Paris, France
Shirayuri College, Tokyo, Japan
Misako Ogino
Catherine S. Tamis-LeMonda
Sophia University, Tokyo, Japan
New \brk University
Maternal speech to infants of 2 ages in 4 cultures was examined to probe how infant age and
cultural variation influence the contents of that speech. Argentine, French, Japanese, and U.S.
American mothers were individually videotaped in naturalistic free-play interactions at home with
their 5- and 13-month-old infants, maternal speech was transcribed, and the contents classified as
affect salient or information salient. Mothers in the 4 cultures use all speech categories to young
infants, speak to older infants more than to younger infants, but differ in the emphasis of their
speech. Similarities speak to the universality of maternal speech to infants, provoked perhaps by
infants' common psychological status; differences in the speech mothers choose to emphasize
speak perhaps to the expression of cultural preferences.
proximity, closeness, and warmth between mother and baby
(e.g., Kaye, 1982; Stern, 1985). Second, although prelingual infants may not benefit from lexical as much as prosodic features
of speech directed to them (Papousek, Papousek, & Bornstein,
1985), maternal speech is acknowledged to play a substantial
role in the child's early language learning (e.g., Bornstein &
Lamb, 1992; Garton, 1991; Harris, 1991; Rondal, 1985). Third,
maternal speech begins the process of informing infants about
themselves, about mother, and about the environment (e.g.,
Sachs, 1977). Last, but not least, to the extent that maternal
speech reflects the communicational code of a culture, it plays a
central role in socializing culturally appropriate communication style (e.g., Blount, 1990; Clancy, 1986; Givon, 1985; Ochs,
1986). These multiple functions elevate maternal speech to babies to the level of a central subject of psychological, linguistic,
and cultural study.
Developmental psycholinguists have approached the analysis of maternal speech to babies with different aims. One has
been to evaluate its functional aspects, as for example, to contrast categories of speech where the intent is social interaction
from speech that is directed toward describing or discovering
the environment. In this vein, Brown (1977, p. 6) divided the
universe of maternal speech addressed to infants into contrasting "features concerned with affection (affection-inspiring, tenderness-inspiring, intimacy-inspiring) and features concerned
with communicative competence (verbal production, verbal
Maternal speech to infants serves several significant functions in development. First, verbal interactions establish ties of
This article was presented at a symposium of the American Anthropological Association in November 1991 in honor of Harold Conklin.
Marc H. Bornstein was supported by Research Grants HD2O559
and HD20807 and by Research Career Development Award HD00521
from the National Institute of Child Health and Human Development
and by a Guggenheim Foundation Fellowship. Marie-Germaine
Pecheux was supported by Institut National de la Sant6 et de la Recherche Medicale Grant 859012. Martine Lamour was supported by a
fellowship from the Fondation pour la Recherche Medicale, Paris,
France. Sueko Toda was supported by a Fogarty International Visiting
Fellowship to the National Institute ofChild Health and Human Development. Misako Ogino was supported by the Sato Foundation, Tokyo,
Japan.
We thank the Universidad de Belgrano, the Laboratoire de Psychologie du Developpement et de l'Education de l'Enfant of the Sorbonne,
and the Faculty of Education of the University of Tokyo for gracious
hospitality in providing bases for data collection. We thank H. Bornstein, J. Galperin, S. D. Galperin, M. Gil-Montero, Y Hirabayashi, G.
Jaimsky, L. Pascual, J. Reul, K. Schulthess, K. Tribicono, B. Wright,
and M. I. Zarate for assistance.
Correspondence concerning this article should be addressed to
Marc H. Bornstein, Child and Family Research, National Institute of
Child Health and Human Development, Building 31—Room B2B15,
9000 Rockville Pike, Bethesda, Maryland 20892. Electronic mail may
be sent to BR2@NIHCU.
593
594
BORNSTEIN ET AL.
comprehension and cognitive competence)." This basic division
of maternal speech to infants into affect-salient and information-salient contents has been widely adopted (e.g., de Villiers &
de Villiers, 1978; Folger & Chapman, 1978; Penman, Cross,
Milgrom-Friedman, & Meares, 1983; Sherrod, Crawley, Petersen, & Bennett, 1978; Toda, Fogel, & Kawai, 1990). The present
study follows in this tradition.
In a related vein, investigators have long speculated that the
contents of maternal speech to infants should also be influenced by characteristics of the baby, including age and developmental level (e.g.. Bates, 1976; Bruner, 1974/1975; Cross,
1977; Phillips, 1973; Sherrod, Friedman, Crawley, Drake, &
Devieux, 1977). As infants grow over the 1st year, maternal
speech patterns and actions toward them shift from feeling-oriented and self-marking content to object-oriented and environment-marking content in ways consistent with infants' increasing interest in their surroundings (Adamson & Bakeman, 1984;
Bornstein & Tamis-LeMonda, 1990; Penman et al., 1983). In an
exemplary longitudinal analysis, Snow (1977) observed that
mothers changed emphasis from discussing their infants' internal states and feelings to discussing activities and events in the
immediate surround. For this reason, in the present study we
analyzed speech in mothers with babies of two ages, 5 months
and 13 months. By the middle of the infant's 1st year, conversations with mothers have taken on many mature characteristics,
such as turn taking (Bornstein & Tamis-LeMonda, 1990; Cohn
& Tronick, 1987; Kaye & Fogel, 1980; Stevenson, Ver Hoeve,
Roach, & Ixavitt, 1986), but the infant is still almost exclusively
preverbal. By the start of the 2nd year, infants comprehend
considerable amounts of speech they hear, and many are even
themselves beginning to talk (Bates, Bretherton, & Snyder.
1988; Tamis-LeMonda & Bornstein, 1990).
In this study, we also evaluated maternal speech to infants in
different cultural settings. Argentina, France, Japan, and the
United States comprise an attractive comparative set in which
to investigate universal as well as culture-specific aspects of
developmental change and continuity in maternal speech.
There are several reasons why. In these countries, we specifically chose to study maternal speech in Portenos, Parisians,
Tokyo people, and New Yorkers. These particular locales and
samples are similar in terms of modernity, level of industrialization, urbanity, per capita income, education, and standard of
living; in all four, infant caregiving is significant, the family is
nuclear in organization, and mother is normally the primary
caregiver in the family setting. So, locales and samples are
roughly comparable in terms of sociodemographics. However,
considerable differences exist among the four in terms of history, culture, beliefs, and values concerned with child rearing.
Thus, the particular comparison set we assembled directly contrasts cultural conditions of child rearing and disentangles
them (to the degree possible) from economic and educational,
urban-rural, modern-traditional, as well as ecological and climatic factors (see Jahoda, 1980; Munroe & Munroe, 1980;
Triandis, 1989). In comparing mothers from Argentina,
France, Japan, and the United States, we are also contrasting
four distinct language communities: Spanish, French, Japanese, and American English. Furthermore, this quartet of cultures contrasts three Western societies and one Eastern society
and so presents the possibility for creating a wider comparison
than is typical of cross-cultural research (see Bornstein, 1980,
1989, 1991; Brislin, 1983; Lewis & Ban, 1977; Piaget, 1966/
1974; Super, 1981; Whiting, 1981). To analyze maternal speech
comprehensively, it is both desirable and necessary to approach
the topic from cross-national and cross-language perspectives.
Our chief purposes in this study, then, were to analyze and
compare the language environments directed to younger and
older infants in terms of two prominent classes of speech—affect salient and information salient—and to assess the generality or specificity of infant-directed speech in Argentine, French,
Japanese, and American mothers. We hypothesized significant
main effects for age for both affect- and information-salient
speech; presumably, mothers communicate verbally more with
their 13-month-old infants than with their 5-month-olds. We
also hypothesized that significant differences would emerge for
culture; for example, previous work indicates that American
mothers emphasize information, and Japanese mothers, affect,
in their speech to babies. Finally; we expected Age x Culture
interactions; that is, cultural differences in functional emphases in maternal speech should more likely be manifest with
older than with younger babies (Azuma, 1986; Bornstein, 1989:
Caudill&Weinstein, 1969; Clancy. 1986; Morsbach, 1973: Toda
et al., 1990).
Method
Subjects
Primiparous mothers and their 5- and 13-month-old infants, recruited from patient populations of private obstetric and pediatric
groups in Buenos Aires, Paris, Tokyo, and New York City, were observed interacting at home. There were 24 and 24 Argentine, 18 and 18
French, 22 and 22 Japanese, and 22 and 25 U.S. American mother-infant dyads with infants of 5 and 13 months, respectively. All infants
were term at birth and healthy at the time of the study, with a mean
weight of 3.25 kg and an average length of 49.8 cm. The samples were
balanced for sex of baby and came from comparable middle-class to
upper middle-class households.' Babies in the Argentine, French, Japanese, and American samples were the same age at the time of the
5-month home visit (M = 164 days, range of country means = 161-165)
as well as the 13-month home visit (M = 407 days, range = 402-411).
Mothers of babies in each of the two age groups were themselves approximately the same age (A/ = 29.7 years, range = 27.2-31.6, for
mothers of 5-month-olds, and M = 30.4 years, range = 28.3-33.1, for
1
Factors external to child rearing per se—for example, differing
views on privacy or the actual physical layout of the home—may differentially influence mothers of different cultures. In several respects,
however, this Buenos Aires-Paris-Tokyo-New York City comparison
overcomes these potential shortcomings. These cities are among the
most cosmopolitan metropolitan areas in the world. Porteflos, Parisians, Tokyo people, and New Yorkers are all normally reticent to invite
strangers into the home; nonetheless, mothers in all four places volunteered participation, and all expressed interest in the study and cordiality toward the observers. Moreover, families in all four cities lived in
comparable apartments in multistory buildings. These factors tend to
diminish the possibility that differential sensitivity on the part of
mothers, or structural differences in their dwellings, systematically
affected maternal speech.
SPECIAL SECTION: CONTENTS OF MATERNAL SPEECH IN FOUR CULTURES
mothers of 13-month-olds) and had similar educational histories
(mean number of years of post-high-school education = 3.5, range =
2.8-4.5, for mothers of 5-month-olds, and M = 3.6, range = 2.1-6.3, for
mothers of 13-month-olds).
Procedure
Home observations were conducted identically in thefourcountries.
Mothers were asked to behave in their usual manner and to disregard
the observer's presence insofar as possible. Besides the observer, only
mother and baby were present in the home, and observations took
place at times of the day that were optimal in terms of individual
babies' being in awake and alert states (see Bornstein, 1985; Bornstein
& Tamis-LeMonda, 1990). Our goal in this study was to observe
mothers under the most natural and unobtrusive conditions possible
and not to alter the context of data collection beyond what would
naturally occur in the home. Naturalistic interactions of mothers and
infants were videotaped. A single female observer, always a native of
the country, visited the home to videotape.
Measurement of Maternal Speech
Transcription. From the 175 videotapes, 15-min segments of
mother-infant play with toys were first identified. Next, word-forword transcripts were made of maternal language during those periods
by native speakers. The transcripts were then compared with the videotapes by different native speakers, and necessary corrections were
made (see Woods, Fletcher, & Hughes, 1986).
Categorization. Maternal speech was coded in terms of the primary
function of each utterance. Several functional taxonomies of maternal
speech have been developed (Delia Corte, Benedict, & Klein, 1983;
Folger& Chapman, 1978; Morikawa, Shand, & Kosawa, 1988; Penman
et al, 1983; Rondal, 1985; Sherrod et al, 1978; Toda et al, 1990), and
our analysis followed them closely. We identified two principal classes
of utterances: (a) affect-salient speech—expressive, generally nonpropositional, idiomatic, or meaningless statements that included greetings,
recitations, onomatopoeia, endearments, and the like, and (b) information-salient speech—normally fully prepositional direct statements,
questions, and reports about the infant, mother, or environment. A
third class of other (including nonaffect, noninformation, and unintelligible) maternal speech was also identified. (Including this class in
analyzing the two classes of interest also eliminated statistical limitations of linear dependence.) Together these three classes are mutually
exclusive and exhaustive of maternal speech to infants. The Appendix
provides definitions of affect- and information-salient speech and examples in each language.
The coding unit consisted of each utterance for which a single functional category could be reliably identified, and the unit changed when
there was a change in coded utterance type or when an utterance terminated and a silence of at least 2 s ensued. Thus, the minimum unit size
could be a single word. On the basis of these parameters, we determined the frequency of each class of maternal speech.
Coders were all natives of their respective countries, and the Argentine, French, and Japanese coders were also English-language bilinguals. AH were trained in the functional classification system. During
practice sessions, examples were provided and discussions held among
coders to resolve infrequent disagreements. Maternal utterances were
classified on the basis of the written transcripts. Coders were first
trained on the transcripts of U.S. mothers; 20% of the U.S. transcripts
were initially coded by all coders. Intercoder reliability was evaluated
using correlations (r) for overall frequency of each category, the unit of
analysis. Average intercoder correlations were .88 for affect, .92 for
direct statements, .99 for questions, .99 for reports, and .60 for other.
During the coding process, 20% of the transcripts within each of the
595
four cultures were also coded by a second native-speaking coder. Average intercoder correlations were .90 for affect, .90 for direct statements,
.96 for questions, .83 for reports, and .72 for other. These are high
reliability values, and they give confidence in the resulting data. As can
be seen in the Appendix, it is fairly easy to tell when a mother is
making a direct statement to her baby, or questioning the baby, or
reporting about the baby, or using affect-salient speech.
Preliminary Analyses
Outliers. Prior to any statistical appraisal of the results, univariate
data were inspected in box plots (Tukey, 1977). Because of consistent
outliers, the data from 2 Argentine, 2 Japanese, and 2 American
mothers of 5-month-olds were omitted. Other outlying data were scattered and selectively eliminated from relevant analyses; therefore, ns in
analyses differ. In all analyses, the maximum available n was used.
Frequencies and proportions. Amounts of affect- and informationsalient maternal speech were analyzed in terms of frequency and were
also transformed into proportions. Analyses of frequencies and proportions address different questions. When the absolute amount of
speech to which a child is exposed is of interest, analysis of frequencies
is appropriate. Examining frequencies assumes that infants are sensitive to absolute amounts of stimulation. When relative amount of stimulation within a particular frame of reference (such as total stimulation) is of interest, analysis of proportions is appropriate. The use of
proportions eliminates individual and cultural differences in total
speech: All mothers display a total speech of 1.0, and so relative emphases within this total can be compared directly across cultures. Because there is no substantive rationale for preferring frequencies to
proportions, and because at one time or another a child might be sensitive to either absolute or relative amounts of different classes of maternal speech, we analyzed both frequencies and proportions. With one
exception, fundamentally similar patterns emerged from the two analytic approaches. Consequently, we report results for frequencies—the
more directly measured variable—and note the single difference that
emerged for proportions.
Infant gender and maternal sociodemographics. Mothers of babies
in each of the two age groups in each of the four countries were approximately the same age and had approximately the same number of years
of post-high-school education. Approximately equal numbers of
mothers ofboys and girls in each locale participated in the study. Analyses of variance (ANOVAs) incorporating infant sex, mother age, and
mother education were conducted for affect- and information-salient
speech. No significant main effects for these variables emerged. In
addition, none displayed significant interactions with either country
or age (of infant). For this reason, we pooled the data for maternal
speech to boys and girls.
Results
Multivariate analyses of variance (MANOYAs) were conducted separately for frequencies and proportions in the stepwise fashion. We first analyzed total affect and information as
dependent variables. With regard to information, a MANOYA
was used as a first step designed to determine whether it was
necessary to analyze subcategories. Total information itself was
not investigated further with respect to pairwise comparisons
among cultures. We then examined the three subcategories of
information—direct statements, questions, and reports—to
verify which might be "responsible" for culture and Age X Culture effects. The main effect of culture was explored via the
Newman-Keuls range test, controlling Type I error at a = .05.
Pairwise comparisons of cultures for significant Age X Culture
596
BORNSTEIN ET AL.
effects were conducted using two-way ANO\As. Because there
were four cultures, six pairwise ANOVAs were possible, only
three of which were orthogonal. To adjust for nonorthogonal
comparisons at a = .05, we used a modified Bonferroni test
(Keppel, 1982). This yielded the requirement that interaction
effects be considered significant only when p < .025.
Total Affect and Information
Table 1 displays mean frequencies for total affect and information classes of maternal speech across two ages and four
cultures. A MANOVA specifying total affect and information
as dependent variables yielded significant main effects for age
and culture and a significant Age X Culture interaction, F(2,
104) = 51.98, F(6, 210) = 16.93, and F(6, 210) = 7.35, respectively, ps < .001. As expected, univariate analyses revealed significant age effects for both affect- and information-salient
classes of maternal speech, Fs(l, 105) = 22.59 and 97.93, respectively, ps < .001. Table 1 shows that mothers in these four cultures spoke more frequently to their 13-month-old infants than
to their 5-month-old infants: with nearly one third more affect
and more than twice as much information. The analogous analysis for proportions yielded main effects of age for affect and
information, Fs(l,105)= 21.86 and 54.00, ps< .001. However,
the effect of age on proportions of affect-salient speech was in
the opposite direction of age's effect for frequency; the proportion of total speech that was affect-salient was, across the four
cultures, .58 at 5 months and .45 at 13 months. In other words,
the frequency of maternal affect-salient speech increased significantly between the ages of 5 and 13 months, but the proportion
of total maternal speech that was affect-salient decreased.
Univariate analyses revealed a significant main effect for culture and a significant Culture X Age interaction for affect-sa-
lient speech, Fs(3,105) = 8.83 and 3.98, respectively, ps < .001.
(See Figure 1, Panel A) A range test showed that Japanese
mothers displayed significantly higher frequencies of affect-salient speech than mothers in the other three cultures. In addition, Argentine mothers displayed significantly higher frequencies of affect-salient speech than French mothers. Significant
pairwise interactions were found between U.S. American and
French mothers, F(l, 61) = 7.21, and between French and Japanese mothers, F(l, 58) = 6.59, ps = .01. As can be seen in Panel
A of Figure 1, only French mothers displayed similar frequencies of affect-salient speech to babies of the two ages; mothers in
the other three cultures displayed an increase in this frequency
between 5 and 13 months.
Univariate analyses revealed a significant main effect for culture and a significant Age X Culture interaction for information-salient speech, Fs(3,105) = 12.87 and 6.48, respectively, ps
< .001. (See Figure 1, Panel B.) To explore these effects further,
we examined the three subcategories of information separately:
direct statements, questions, and reports.
Subcategories of Information
Considered together. Table 2 displays mean frequencies of
maternal speech for the three subcategories of information-salient speech. A MANOVA specifying the three categories as
dependent variables revealed significant main effects for age
and culture and a significant Age X Culture interaction, F(3,
126) = 39.14, F(9, 374) = 13.66, and F(9, 374) = 6.00, respectively, ps < .001.
Direct statements. Univariate analysis revealed significant
main effects for age and culture, F{\, 128) = 79.42 and F(3,
128) = 31.86, respectively, ps < .001. In addition, a significant
Age X Culture interaction emerged, F(3,128) = 8.08, p < .001.
Table 1
Mean Frequencies for Affect-Salient and Information-Salient Maternal Speech
Across Two Ages and Four Cultures
Age
Argentina
France
Japan
United States
All cultures
Affect-salient speech
5 months
M
SD
13 months
M
SD
Both ages
M
SD
77.2
26.7
64.3
20.6
82.7
46.8
56.2
30.4
70.4
30.0
92.2
21.7
59.9
19.6
129.9
48.6
86.6
26.1
93.4
39.3
84.5
25.2
61.9
19.8
107.0
52.8
72.3
26.1
Information-salient speech
5 months
M
SD
13 months
M
SD
Both ages
M
SD
58.8
25.5
55.2
29.2
38.8
26.2
50.4
29.4
50.4
28.1
145.9
31.8
68.6
37.6
90.8
38.9
129.2
34.4
115.0
45.1
109.6
52.3
60.9
33.0
65.5
42.1
92.8
50.9
SPECIAL SECTION: CONTENTS OF MATERNAL SPEECH IN FOUR CULTURES
597
AFFECT-SALIENT
160
13 MONTHS
5 MONTHS
INFANT AGE
INFORMATION-SALIENT
13 MONTHS
5 MONTHS
INFANT AGE
Figure 1. Maternal affect-salient speech (top) and maternal information-salient speech
(bottom) to infants of two ages in four cultures.
Examination of the age effect showed that mothers make
direct statements to their 13-month-olds more than 2.5 times
as much as mothers do to their 5-month-olds, and the Newman-Keuls range test showed Argentine mothers display significantly higher frequencies of direct statements than their
counterparts in France, Japan, and the United States. No other
pairwise comparison for the main effect of culture was significant. Pairwise examination of the Age X Culture interaction
showed significant effects for the three comparisons involving
Argentine mothers: Argentina-France, F(l, 68) = 12.24; Argentina-Japan, F(l, 71) = 18.71; and Argentina-United States,
F(l, 75) = 8.11; ps < .01. Figure 2, Panel A shows these interactions; they must be interpreted with caution, however, because they do not cross over (Argentine mothers at 5 months
showed the highest frequency of the four cultures, and it is
possible that their greater slope is due to scaling). All three
interactions show the same pattern—the increase from 5 to 13
months is greater for Argentine mothers as compared with
mothers in the three other cultures—as expected, because the
French, Japanese, and U.S. American mothers displayed statistically equivalent patterns.
Questions. Univariate analyses revealed significant age and
culture main effects, F(l, 128) = 33.73 and F(3,128) = 7.08,
respectively, ps < .001. Mothers of 13-month-olds asked their
infants almost twice as many questions as mothers of 5-montholds asked their infants. The Newman-Keuls range test showed
that U.S. mothers more frequently asked their infants questions
than did French or Japanese mothers and that Argentinean
mothers asked more questions than Japanese mothers. (See Figure 2, Panel B.)
Reports. Univariate analysis revealed significant main effects for age, F(l, 128) = 72.20, p < .001, and culture, F(3,128) =
3.42, p < .05. In addition, a significant Age X Culture interaction emerged, F(3,128) = 9.65, p < .001. Mothers of 13-montholds reported to their infants about the infants, about themselves, or about the environment almost 2.5 times as much as
598
BORNSTEIN ET AL.
Table 2
Mean Frequencies for Subcategories of Maternal Information-Salient Speech
Across Two Ages and Four Cultures
Age
Argentina
France
Japan
United States
All cultures
Direct statements
5 months
M
SD
13 months
M
SD
Both ages
M
SD
19.8
11.6
13.3
10.8
6.4
4.4
9.3
6.7
12.3
10.1
52.1
16.8
24.6
13.3
15.9
7.6
26.1
11.9
31.2
19.2
37.0
21.7
18.6
13.1
11.2
7.8
17.9
12.8
Questions
5 months
M
SD
13 months
M
SD.
Both ages
M
SD
24.4
11.8
18.9
9.9
17.8
13.0
26.5
20.0
22.3
18.5
40.6
15.0
33.6
20.7
26.6
12.9
49.2
17.8
38.3
18.4
32.9
15.7
26.0
17.4
22.5
13.5
38.6
21.5
Reports
5 months
M
SD
13 months
M
SD
Both ages
M
SD
19.2
12.2
27.6
16.7
17.3
13.1
19.3
12.5
20.7
14.0
55.3
19.4
30.6
21.4
46.5
23.4
56.0
20.9
48.4
23.0
39.4
24.5
29.1
19.0
32.6
24.1
39.7
25.4
mothers of 5-month-olds reported to their infants. Despite an
overall significant F for culture, no significant pairwise differences emerged: Pairwise differences that might exist, therefore,
are small. Pairwise examinations of the Age X Culture interaction for reports were significant for the three comparisons involving French mothers: France-Argentina, F(l, 68) = 22.93;
France-Japan, F(l, 63) = 13.05; and France-United States, F(\,
67) = 17.10; ps <. .001. As with direct speech, all three significant interactions show the same pattern (Figure 2, Panel C):
Relative to mothers of other cultures, French mothers showed a
smaller increase in reports between 5 and 13 months.
Discussion
This study compared the functional contents of Argentine,
French, Japanese, and US. American mothers' speech to their
5- and 13-month-old infants in the natural setting of the home.
We first review and comment on the results of the study in
terms of cultural universals and specifics in maternal speech
and then discuss more general issues about maternal speech to
infants in its cultural context.
Mothers in these three Western cultures and one Eastern
culture showed general similarity in certain aspects of speech to
infants. These findings submit to cross-cultural evaluation the
universality of developmental processes related to maternal
speech. First, the consistency and face validity of findings resulting from the cross-cultural application of the present categorization scheme of maternal speech give evidence of its generalizability to culturally contrasting settings. Apparently, maternal
language to infants nearly universally contains affect-salient as
well as information-salient topics. Results of the present study
also show that a functional perspective on maternal speech is
applicable to infants of two ages in four different language communities across three Western cultures and one Eastern culture.
It is likely that mothers everywhere intend to share feelings and
contribute to emotional exchanges via their affect-salient
speech to babies, just as they wish to impart or confirm cognitive information referential of infant perceptual experiences.
Perhaps these are universal aspects of parenting (Papousek &
Papousek, 1987; Vigilant, Stoneking, Harpending, Hawkes, &
Wilson, 1991). Furthermore, no systematic sex effects or effects
ascribable to maternal age or education emerged for maternal
speech in these samples; but we limited variation in these dimensions, and these factors could be expected to influence
maternal speech in other samples.
Mothers in the four cultures showed some similarity in the
SPECIAL SECTION: CONTENTS OF MATERNAL SPEECH IN FOUR CULTURES
DIRECT STATEMENTS
4CY
50 40
UI
D 30
United State*
O
111
£ 20
10
0
5 MONTHS
13 MONTHS
INFANT AGE
QUESTIONS
S MONTHS
13 MONTHS
INFANT AGE
REPORTS
5 MONTHS
13 MONTHS
INFANT AGE
Figure 2. Maternal information-salient speech to infants of two ages in four cultures for (top panel) direct
statements, (middle panel) questions, and (bottom panel) reports.
599
600
BORNSTEIN ET AL.
types and frequency of functional classes of speech they addressed to infants. Notably, predicted age effects emerged for
affect- and information-salient speech. Mothers as a group increase the frequency with which they use both categories of
speech as their infants approach their first birthdays. Provocatively, the age effect that emerged for affect was reversed when
we analyzed proportions—between 5 and 13 months, the frequency of affect-salient speech climbs, but affect becomes a
smaller proportion of total maternal speech. Penman et al.
(1983) also found relatively greater increases in information-oriented versus affect-oriented classes of Australian mothers'
speech to 3- and 6-month-old infants. Thus, independent of
culture, mothers speak more to older infants, but "relatively
speaking," information-salient versus affect-salient speech increases in emphasis with age in all cultures.
Argentine, French, Japanese, and U.S. American mothers
also favored affect over information in speaking to their 5month-old infants and information over affect in speaking to
their 13-month-old infants. Examination of information subcategories showed that the mean frequency of the three approximately doubled with age. Apparently, over the second half of the
1st year of life, mothers tend to expect that their infants need to
be directed more, expect that their infants know more or will
perhaps better comprehend their questions, and expect that
they can and should give their infants more information about
the infants themselves, their mothers, and the environment.
During this period, infants themselves are also of course beginning to talk, a fact that may encourage mothers to speak to
them more in an increasingly adult-oriented conversational
manner.
There were also exceptions to these generalities and variations in the speech of mothers in these four cultures that appear
to reflect general cultural beliefs and values. Variation in affectoriented speech appeared across the four cultures: Japanese
mothers were highest in the use of affect-salient speech, meaning that they used the most grammatically incomplete utterances and that they played with sounds in speaking to their
babies, using nonsense, onomatopoeia, song, and the like,
more than did mothers of the other cultures. This result replicates and expands on the work of Toda et al. (1990), who found
in the laboratory that Japanese mothers used nonsense, onomatopoeic utterances, and calling by name to their 3-montholds more than did American mothers. The Japanese mother's
goal in early child rearing is to empathize with her infant's
needs—to meet her infant at the infant's level—as is well recognized, rather than to show authority as mother (Befu, 1986;
Bornstein, Azuma, Tamis-LeMonda, & Ogino, 1990; Doi,
1973; Kojima, 1986a, 1986b). Reciprocally, the evidence that,
relatively speaking, mothers from the three Western cultures
favor information-salient speech and more often use grammatically complete utterances in speaking to babies—that is,
speech more characteristic of adult-adult conversation—is consistent with the view that these mothers are more interested in
supporting individual expression and imparting information to
their children from an early age.
Variations in categories of information-oriented speech also
emerged across these cultures. First, Argentine mothers displayed higher frequencies of direct statements than mothers in
the three other cultures. In Argentina, the traditional child-
rearing orientation is authoritative; that is, to direct a child's
action is to behave positively toward the child and to express
care and love in preparing the child for development (Aguinis,
1988; Diaz Rossello, 1988; Fillol, 1961). Blount (1990) noted
that Spanish-speaking mothers rely comparatively more than
English-speaking mothers on attentionals (utterances designed
to attract the child's attention) and high volume in interactions
with their 9- to 22-month-olds. Even in schooling, Argentines
favor direct transmission of knowledge as, for example, by demonstration (Pascual, 1991; Petty, 1986). Perhaps the use of these
features of speech stems from a view of infants and children as
in need of instruction. The interaction effect showed that Argentine mothers more often directed their 13- than their 5month-olds relative to mothers in the three other cultures, implying that as culture takes hold of language directed at the
child, its effects may increase.
Second, U.S. American mothers tended to question their infants the most. It is unlikely that American mothers believe that
their 5- or even 13-month-olds will answer their questions, or
even need to, because many questions are asked about ongoing
activities or about things in plain view. Rather, question asking
of infants may be a way for American mothers to emphasize the
information components of speech. After all, asking "What
does that toy do?" goes beyond simply posing a question; it
conveys information via labeling, for example, and at the same
time does so in a most attention-getting way, as through the use
of interrogative prosody (Bornstein & Lamb, 1992). Questions
also emphasize the child's participation in his or her own language development in contrast to mothers' use of descriptions;
this is a "distancing strategy" thought to promote cognitive development (Sigel, 1982).
Third, French mothers speak to their younger infants like
other mothers, but they increased the frequency of their reports
to older infants at a lower rate than mothers in the other three
cultures. Reputedly, French mothers place less emphasis on
achievement stimulation and more emphasis on emotional support and use language to establish closeness (Dion, 1989), an
opinion consonant with their lower level of reporting to babies
(found here) and their lower level ofdidactic activity and responsiveness overall (Bornstein et al., in press; Bornstein, Tamis-LeMonda, Pecheux, & Rahn, 1991). It might also be that French
mothers approach speech to their babies with a more egalitarian, adult-adult orientation.
It is important to recognize that mothers in only four cultures
were sampled in this study, and that the samples of mothers in
these four locales were comparatively restricted in terms of
sociodemographic characteristics, including urban location,
economic status, and educational history. These restrictions
actually aid the comparisons we undertook, because the four
samples were similar on these variables. However, they also
have implications for the generalizability of the findings, because different results could emerge in other samples or in samples from less metropolitan environments, from other regions
of the socioeconomic status scale, or from less educated families in these four countries. Moreover, we concentrated these
analyses on functional aspects of maternal speech occurring in
free-play interactions at home; analyses of other aspects of
speech in mothers or of the same or different aspects of speech
in other situations may result in similar or different patterns of
SPECIAL SECTION: CONTENTS OF MATERNAL SPEECH IN FOUR CULTURES
findings. Finally, mothers in the four countries spoke different
languages, and in this study we did not attempt to disentangle
language from culture. Perhaps language and culture are most
meaningfully considered together. However, "natural experiments" that separate language and culture (to the degree possible) could be invoked to identify cultural versus linguistic influences on maternal speech, as for example by comparing different cultures in which the same language is spoken.
One of the ways infants become socialized into culture is
through integrating maternal behavior and speech into their
development, and one of the ways they do this is via mother-infant communication. The general association between culture
and development has been conceived in terms of several fundamentally different models. One traditional and influential paradigm suggests that cultural norms, convictions, images, and
rules influence the development of parental beliefs about children, their abilities, character, needs, temperament, and so
forth, and that in turn these parental beliefs (somehow) translate themselves into the verbal, enactive, and responsive practices that parents use to achieve their child-rearing goals (e.g.,
Bateson & Mead, 1942; Benedict, 1938; Caudill, 1973; Tapp,
1980; Whiting & Child, 1953). The present findings suggest
that this model, at least in part, validly describes cultural forces
that shape maternal speech to infants. At the same time, however, universal aspects of infancy, such as the dramatic growth
of competence over the 1st year, appear also to exert control
over the contents of maternal speech.
References
Adamson, L. B., & Bakeman, R. (1984). Mothers' communicative acts:
Changes during infancy. Infant Behavior and Development, 7, 467478.
Aguinis, M. (1988). Un pais de novela. Buenos Aires, Argentina: Editorial Planeta.
Azuma, H. (1986). Why study child development in Japan? In H. W
Stevenson, H. Azuma, & K. Hakuta (Eds.), Child development and
education in Japan (pp. 3-12). New York: W. H. Freeman.
Bates, E. (1976). Language and context: The acquisition ofpragmatics.
San Diego, CA: Academic Press.
Bates, E., Bretherton, I, & Snyder, L. (1988). Fromfirstwords to grammar. New York: Cambridge University Press.
Bateson, G., & Mead, M. (1942). Balinese character: A photographic
analysis. New York: New York Academy of Sciences.
Befu, H. (1986). Social and cultural background for child development
in Japan and the United States. In H. W Stevenson, H. Azuma, & K.
Hakuta (Eds.), Child development and education in Japan (pp. 1327). New York: W H. Freeman.
Benedict, R. (1938). Continuities and discontinuities in cultural conditioning. Psychiatry, 1,161-167.
Blount, B. G. (1990). Parental speech and language acquisition: An
anthropological perspective. Pre- and Peri-Natal Psychology, 4, 319337.
Bornstein, M. H. (1980). Cross-cultural developmental psychology. In
M. H. Bornstein (Ed.), Comparative methods in psychology (pp. 231—
281). Hillsdale, NJ: Erlbaum.
Bornstein, M. H. (1985). How infant and mother jointly contribute to
developing cognitive competence in the child. Proceedings of the
National Academy of Sciences, 82, 7470-7473.
Bornstein, M. H. (1989). Cross-cultural developmental comparisons:
The case of Japanese-American infant and mother activities and
601
interactions. What we know, what we need to know, and why we need
to know. Developmental Review, 9,171-204.
Bornstein, M. H. (1991). Approaches to parenting in culture. In M. H.
Bornstein (Ed.), Cultural approaches to parenting (pp. 3-19). Hillsdale, NJ: Erlbaum.
Bornstein, M., Azuma, H., Tamis-LeMonda, C, & Ogino, M. (1990).
Infant and mother activity and interaction in Japan and in the
United States: I. A comparative macroanalysis of naturalistic exchanges. International Journal of Behavioral Development, 13, 267287.
Bornstein, M. H., & Lamb, M. E. (1992). Development in infancy: An
introduction. New York: McGraw-Hill.
Bornstein, M. H., & Tamis-LeMonda, C. S. (1990). Activities and interactions of mothers and their firstborn infants in thefirstsix months
of life: Covariation, stability, continuity, correspondence, and prediction. Child Development, 61,1206-1217.
Bornstein, M. H., Tamis-LeMonda, C. S., Pecheux, M.-G., & Rahn,
C. W (1991). Mother and infant activity and interaction in France
and in the United States: A comparative study. International Journal
of Behavioral Development, 14, 21-43.
Bornstein, M. H., Tamis-LeMonda, C. S., Tal, J., Ludemann, P., Toda,
S., Rahn, C. W, Pecheux, M.-G., Azuma, H., & Vardi, D. (in press).
Maternal responsiveness to infants in three societies: The United
States, France, and Japan. Child Development.
Brislin, R. W (1983). Cross-cultural research in psychology. Annual
Review of Psychology, 34, 363-400.
Brown, R. (1977). Introduction. In C. Snow& C. Ferguson (Eds.), Talking to children (pp. 1-27). New York: Cambridge University Press.
Bruner, J. S. (1974/1975). From communication to language: A psychological perspective. Cognition, 3, 255-287.
Caudill, W (1973). The influence of social structure and culture on
human behavior in modern Japan. Journal of Nervous and Mental
Disease, 157, 240-257.
Caudill, W, & Weinstein, H. (1969). Maternal care and infant behavior
in Japan and America. Psychiatry, 32,12-43.
Clancy, P. (1986). The acquisition of communicative style in Japanese.
In B. Schieffelin & B. Ochs (Eds.), Language socialization across
cultures (pp. 213-250). Cambridge, England: Cambridge University
Press.
Cohn, J. E, & Tronick, E. Z. (1987). Mother-infant face-to-face interaction: The sequence ofdyad ic states at 3,6 and 9 months. Developmental Psychology, 23, 68-77.
Cross, T. G. (1977). Mothers' speech adjustments: The contribution of
selected child listener variables. In C. Snow & C. Ferguson (Eds.),
Talking to children (pp. 151-188). Cambridge, England: Cambridge
University Press.
Delia Corte, M., Benedict, H., & Klein, D. (1983). The relationship of
pragmatic dimensions of mothers' speech to the referential-expressive distinction. Journal ofChild Language, 10, 35-43.
de Villiers, J. G., & de Villiers, P. A. (1978). Semantics and syntax in the
first two years: The output of form and function and the form and
function of the input. In F. D. Minifie & L. L. Lloyd (Eds.), Communicative and cognitive abilities: Early behavioral assessment (pp. 309348). Baltimore: University Park Press.
Diaz Rossello, J. L. (1988). Larelacionmadre-hijo en el periodo inicial.
In M. Cusminsky, E. Moreno, & E. Suarez Ojeda (Eds.), Crecimiento
y desarrollo (pp. 149-163). (Publicacion Cientifica No. 510). Buenos
Aires, Argentina: Organizacion Panamericana de la Salud.
Dion, F. (1989). Les representations parentales du developpement cognitifde leur b6be et de lew role dans ce developpement. Unpublished
manuscript, Laboratoire de Psychologie du Developpement et de
l'Education de l'Enfant, Universite Paris V
Doi, T. (1973). [The anatomy ofdependence] (J. Bester, Trans.). Tokyo:
Kodansha International.
602
BORNSTEIN ET AL.
Fillol, T. R. (1961). Social factors in economic development: The Argen- Mothers' speech to prelingual infants: A pragmatic analysis. Journal
tine case. Westport, CT: Greenwood Press.
ofChild Language, 10,17-34.
Folger, J., & Chapman, R. (1978). A pragmatic analysis of spontaneous
Petty, M. (1986). Autoritarismo y participacion. Buenos Aires, Argenimitations. Journal ofChild Language, 5, 25-38.
tina: Congreso Pedagogico.
Garton, A. F. (1991). Social interaction and the development oflanguage Phillips, J. (1973). Syntax and vocabulary of mothers' speech to young
children: Age and sex comparisons. Child Development, 44, 182and cognition. Hillsdale, NJ: Erlbaum.
185.
Givon, T. (1985). Function, structure, and language acquisition. In D. I.
Slobin (Ed.), The cross-linguistic study of language acquisition (Vol. Piaget, J. (1974). [Need and significance of cross-cultural studies in
genetic psychology] (C. Dasen, Trans.). In J. W Berry & P. R. Dasen
2, pp. 1005-1027). Hillsdale, NJ: Erlbaum.
(Eds.), Culture and cognition (pp. 299-309). London: Methuen. (OrigHarris, M. (1991). Language experience and early language developinal work published 1966)
ment: From input to uptake. Hillsdale, NJ: Erlbaum.
Rondal, J. (1985). Adult-child interaction and the process of language
Jahoda, G. (1980). Cross-cultural comparisons. In M. H. Bornstein
acquisition. New York: Praeger.
(Ed.), Comparative methods in psychology (pp. 105-148). Hillsdale,
Sachs, J. (1977). The adaptive significance of linguistic input to prelinNJ: Erlbaum.
guistic infants. In C. Snow & C. Ferguson (Eds.), Talking to children.
Kaye, K. (1982). The mental and social life ofbabies. Chicago: UniverNew York: Cambridge University Press.
sity of Chicago Press.
Sherrod, K. B., Crawley, S., Petersen, G., & Bennett, P. (1978). Maternal
Kaye, K., & Fogel, A. (1980). The temporal structure of face-to-face
language to prelinguistic infants: Semantic aspects. Infant Behavior
communication between mothers and infants. Developmental Psyand Development, 1, 335-345.
chology, 16, 454-464.
Keppel, G. (1982). Design and analysis: A researcher's handbook. Eng- Sherrod, K. B., Friedman, S., Crawley, S., Drake, D, & Devieux, J.
(1977). Maternal language to prelinguistic infants: Syntactic aspects.
lewood Cliffs, NJ: Prentice-Hall.
Child Development, 48,1662-1665.
Kojima, H. (1986a). Child rearing concepts as a belief-value system of
Sigel, I. E. (1982). The relationship between parental distancing stratethe society and the individual. In H. W Stevenson, H. Azuma, & K.
gies and the child's cognitive behavior. In L. M. Laosa & I. E. Sigel
Hakuta (Eds.), Child development and education in Japan (pp. 39(Eds.), Families as learning environments for children (pp. 47-86).
54). New York: W H. Freeman.
New York: Plenum Press.
Kojima, H. (1986b). Japanese concepts of child development from the
Snow, C. (1977). The development of conversation between mothers
mid-17th to mid-19th century. International Journal of Behavioral
and babies. Journal of Child Language, 4,1-22.
Development, 9, 315-329.
Stern, D. N. (1985). The interpersonal world of the infant. New York:
Lewis, M., & Ban, P. (1977). Variance and invariance in the mother-inBasic Books.
fant interaction: A cross-cultural study. In P. H. Leiderman, S. R.
Stevenson, M., Ver Hoeve, J., Roach, M., & Leavitt, L. (1986). The
Tulkin, & A. Rosenfeld (Eds.), Culture and infancy: Variations in the beginning of conversation: Early patterns of mother-infant vocal
human experience (pp. 329-355). San Diego, CA: Academic Press.
responsiveness. Infant Behavior and Development, 9, 423-440.
Morikawa, H., Shand, N, & Kosawa, Y (1988). Maternal speech to
Super, C. M. (1981). Behavioral development in infancy. In R. H.
prelingual infants in Japan and the United States: Relationships
Munroe, R. L. Munroe, & B. B. Whiting (Eds.), Handbook ofcrossamong functions, forms, and referents. Journal of Child Language,
cultural human development (pp. 181-270). New York: Garland
15, 237-256.
STPM Press.
Morsbach, H. (1973). Aspects of nonverbal communication in Japan.
Tamis-LeMonda, C. S., & Bornstein, M. H. (1990). Language, play, and
Journal ofNervous and Mental Disease, 157, 262-277.
attention at one year. Infant Behavior and Development, 13, 85-98.
Tapp, J. L. (1980). Studying personality development. In H. C. Triandis
Munroe, R. L., & Munroe, R. H. (1980). Perspectives suggested by
& A. Heron (Eds.), Handbook ofcross-cultural psychology: Developanthropological data. In H. C. Triandis & W W Lambert (Eds.),
Handbook ofcross-cultural psychology: Perspectives (Vol. 1, pp. 253- mental psychology (Vol. 4, pp. 343-423). Boston: Allyn & Bacon.
Toda, S., Fogel, A., & Kawai, M. (1990). Maternal speech to three317). Boston: Allyn & Bacon.
month-old infants in the United States and Japan. Journal of Child
Ochs, E. (1986). Introduction. In B. B. Schieffelin & E. Ochs (Eds.),
Language, 17, 279-294.
Language socialization across cultures (pp. 1-13). Cambridge, EnTriandis, H. C. (1989). The self and social behavior in differing cultural
gland: Cambridge University Press.
contexts. Psychological Review, 96, 506-520.
Papousek, H., & Papousek, M. (1987). Intuitive parenting: A dialectic
Tukey, J. W (1977). Exploratory data analysis. Menlo Park, CA: Adcounterpart to the infant's integrative competence. In J. D. Osofsky
dison-Wesley.
(Ed.), Handbook of infant development (2nd ed., pp. 669-720). New
Vigilant, L., Stoneking, M., Harpending, H., Hawkes, K., & Wilson,
York: Wiley.
A. C. (1991). African populations and the evolution of human mitoPapousek, M, Papousek, H., & Bornstein, M. (1985). The naturalistic
chondrial DNA. Science, 253, 1503-1507.
vocal environment of young infants: On the significance of homogeWhiting, J. W (1981). Environmental constraints on infant care pracneity and variability in parental speech. In T. Field & N. Fox (Eds.),
tices. In R. H. Munroe, R. L. Munroe, & B. B. Whiting (Eds.), HandSocial perception in infants (pp. 269-297). Norwood, NJ: Ablex.
book ofcross-cultural human development (pp. 155-179). New York:
Pascual, L. (1991). Democracy and educational reforms in Argentina: Garland STPM Press.
Possibilities and limitations. Paper presented at the Annual Confer- Whiting, J. W M., & Child, I. L. (1953). Child training and personality:
ence of the Comparative and International Education Society, PittsA cross-cultural study. New Haven, CT: Yale University Press.
burgh, PA.
Woods, A., Fletcher, P., & Hughes, A. (1986). Statistics in language
Penman, R., Cross, T, Milgrom-Friedman, J., & Meares, R. (1983).
studies. Cambridge, England: Cambridge University Press.
SPECIAL SECTION: CONTENTS OF MATERNAL SPEECH IN FOUR CULTURES
603
Appendix
Maternal Speech
Affect-salient speech consisted of expressive, generally nonpropositional, idiomatic, or meaningless statements and included, among
other categories, the following:
1. Greetings: Vocatives.
English: Hello there.
French: Bonjour.
Japanese: Konnichiwa.
Spanish: Hola!
2. Recitations: Nursery rhymes and the like.
English: Peek-a-boo.
French: La petite bete qui monte, qui monte.
Japanese: Inai inai baa!
Spanish: Cucu.
3. Onomatopoeia: Formation of a word by imitation of the sound of
the thing named.
English: Meow, meow (cat).
French: Tic, tac (clock)
Japanese: Moo, moo (cow).
Spanish: Guau, guau, guau (dog).
4. Endearments.
English: Sweetie.
French: Ma puce.
Japanese: Bokuchan.
Spanish: Mi bichito.
Information-salient speech consisted of fully prepositional sentences
about infant, mother, or the environment and included the following:
1. Direct statements.
English: Try one more time.
French: Tu vas le chercher.
Japanese: Ocha chodai.
Spanish: Mini, mini el oso.
2. Questions.
English: What does that toy do?
French: Tu veux jouer?
Japanese: Nani shiteruno?
Spanish: iQueres gatear bebe?
3. Reports.
English: You sure do like your blocks, don't you?
French: Tu es contente de la voir.
Japanese: Kowakatta none.
Spanish: Ese solo te gusta ahora.
Received January 3, 1992
Accepted January 5, 1992 •