7.

Sex differences in parent­

child interaction
JEAN BERKO GLEASON

Since by now it is well documented that there are differences in the

:ways grown men and women speak, it seems reasonable at this point
to ask where those differences originate. There are, of course, a number
of plausible explanations of the origins of sex differences in language:
They can arise either from inborn differences or as a result of environ­
mental . forces, or perhaps as a result of an interaction between the two.
In this chapter, the emphasis will be on environmental forces, especially
the role mothers and fathers play in shaping the language of their daugh­
ters and sons. This is not an attempt, however, to say that there are no
inborn differences. The work of McKeever (Chapter 10 of this volum<D;
Witelson and Pallie (1973), and many others has shown that it is entirely
likely that the language areas of the brains of males and females are
not identical: Specialization for language appears to develop earlier in
the brains of females, and males appear to be more vulnerable to every
kind of insult that affects language development and retention at every:
-'
age from early childhood through advanced old age.
Even if there were no differences in the neuroanatomical bases of
language in males and females, there would be other obvious differences
that, though not themselves linguistic, could have a differential effect
on language development. Young males, for instance, are more phys­
ically aggressive than females in all cultures that have been studied
(Maccoby and Jacklin, 1974). It should not surprise us, therefore, to
find adults uttering more negative statements and more prohibitions to
boys than to girls, and that is exactly what Cherry and Lewis (1976)
found. Adults spent more time trying to control young boys, and ad1Ilt
language, of course, reflected those efforts. If children's language de-·
velopment is affected by the kinds of language they hear when inter­
acting with adults, girls and boys may develop different kinds of language,
189
 190 JEAN BERKO GLEASON
\because they are spoken to differently. Thus, males and females may
produce and elicit different kinds of language because of their different
neurological and behavioral dispositions; and these possibly inborn dif­
ferences may be amplified by society.
There are also powerful environmental forces that shape the war_
individuals speak and that lead to stylistic variation: Males and females
speak differently as a reflection of their gender roles. The use of certain
lexical items, syntactic forms, and intonation patterns cannot be rea­
sonably tied to either neurological or inborn behavioral differences,
since they are culturally constrained. The use of the adjective "darling"
in English, for instance, or a special set of pronouns in Japanese, may
be limited to women, but not for any intrinsic reason.
There is general agreement that, for whatever reason, men and women
speak differently as adults. Presumably, these differences began to
emerge at some point in childhood, and the most likely context of their
development lies in the arena of parent-child interaction.

Input Language

We may well ask the questions, When do little boys and girls first begin
to sound like males and females? and What role do parents have in the
development of whatever differences there are? But questions of this
sort are very recent indeed. The ontogenesis of sex differences in lan­
guage has hardly been explored, and only very recently have we had
any information at all about the possible differences in the speech di­
rected to girls and boys by their mothers and fathers.
One major reason for the dearth of information on what is obviously
an interesting and important topic lies in the nature of the theories that
have dominated the study of children's language development in the
years since Chomsky first published Syntactic Structures (1957). The
models of child language acquisition that dominated the field in the
1950s and 1960s were child-centered and did not consider the role of
adults, except insofar as they were thought to provide a rather degen­
erate sample of language that the young language-learning child could
feed into her or his Language Acquisition Device. It was generally
assumed that differences in the language the child heard (and this lan­
guage was called Input Language) did not matter, since the child's Lan­
guage Acquisition Device was equipped with suitable filters for
processing out those elements that were not of use at a given time. The
search was for universals, with an emphasis on the acquisition of syntax.
The burden of acquisition lay on the child, and the role of adults in the
child's environment was minimized; it was assumed that all the child
 Sex differences in parent-child interaction 191
needed to set the Language Acquisition Device in motion was to
overhear a sufficiently large sample of the target language.
In the late 1960s this picture began to change for a variety of reasons.
Among other things, a number of researchers (for instance, Gleason,
1973; Remick, 1971; Snow, 1972) began to wonder if it was really true
that young children had to learn the rules of language from listening to
a complex and degenerate corpus of adult speech. This led investigators
to study the language of mothers of young language-learning children.
The results of those studies are well known: Mothers' speech to young
children is much less complex than their speech to other adults and
appears to contain design features that may make the learning of lan­
guage easier. The language of mothers to their 2-year-olds is slow,
redundant, simple, and, above all, grammatical. This special speech may
or may not make the acquisition of syntax simpler: There is still a raging
controversy in the field on this subject, with some (e.g., Moerk, 1975)
claiming that mothers' speech contains all of the elements necessary to
teach children grammatical language, and others (e.g. Gleitman, New­
port, and Gleitman, 1984) claiming that what appears to be simple in
mothers' speech is not and that mothers' speech at best can have only
a superficial influence on children's acquisition of language.
No one contests, however, that mothers' speech has the particular
form that has been described by so many researchers, and it is that very
special kind of speech that is now referred to as input language. (Those
who think it is unimportant and uninteresting tend to call it "Moth­
erese," but for a number of reasons that will soon become apparent,
this is a misnomer. ) Another name for input language is Child Directed
Speech (CDS), which is a bit more accurate, since the special features
of this speech are necessitated by the child who is being addressed rather
than by the person, who may or may not be a mother, who produces
it.
Once it became clear that mothers have a special way of speaking to
young children, a number of questions arose in addition to those that
center on the acquisition of syntax. These questions have to do with
stylistic (or registral) variation: Input language, or CDS, is clearly a
separate style, or register. It appears in the speech of women who are
not mothers, in the speech of fathers, and, indeed, in the speech of all
speakers, child and adult, who are addressing young children. Shatz and
Gelman (1973) showed that even 4-year-olds make some modifications
in their usual speech when they speak to 2-year-olds. Other researchers
(Giattino and Hogan, 1975; Golinkoff and Ames, 1979) showed that
fathers' speech also contained the simplifying and clarifying modifica­
tions that had been noted in mothers' speech. Thus, input language
(CDS) containing some special features is produced by all speakers
192 JEAN BERKO GLEASON
addressing young children. Bohannon and Marquis (1977) suggest that
it is children themselves who cause these modifications, because it can
be demonstrated that speakers adjust the complexity of their utterances
in accordance with the signals of comprehension or noncom prehension
produced by their addressees. While this can be shown experimentally,
it is also true that speakers have preconceived notions of how to talk
to young children: Adults simplify and clarify their speech when they
only think they are talking on the phone to 2-year-olds (Snow, 1972),
and young children in preschool produce typical "baby talk" when play­
ing with dolls (Sachs and Devin, 1976; Andersen, 1977). Some of the
features of CDS are undoubtedly tied to communication pressure (for
instance, clear enunciation), but others are part of a conventionalized
speech register (for instance, calling a rabbit a "bunny"). Young children
acquire this register as part of their developing communicative com­
petence, and adults use this register in speaking to young children.
Input language is not a unitary phenomenon, however. It changes
over time and becomes more complex as children's ability to compre­
hend it changes. By the time children are 4 or 5, adults speak to them
in a "language of socialization" that emphasizes not so much syntactic
clarity, or the rules of language, as the rules of society. Speech to a 2­
year-old contains many phrases like "See the bunny. It's a nice bunny.
Pat the bunny," while speech to a 5-year-old contains many phrases like
"Look both ways before you cross the street," "Say thank you to Mrs.
Williams," and "Sit up at the table. "
CDS thus occurs in different forms, depending on the age of the child
being addressed. There may be some argument about the relation be­
tween the syntax used by adults in their CDS and the acquisition of
syntax by children, but there is general agreement that adults explicitly
teach children social conventions and that the adult language is the
medium of that education.
What remains to be determined is whether the sex of the child as well
as the age of the child has an effect on the CDS, and, additionally,

whether CDS varies according to the sex of the speaker. Unless girls
and boys are exposed to different adult models or are spoken to dif­
ferently, we are hard pressed to provide an environmental explanation
for how sex differences in their own language might possibly originate.
In the rest of this chapter, a number of studies originating in our own
laboratory will be discussed. The questions to be considered involve:
(1) differences between mothers' and fathers' speech to children, re­
gardless of the child's sex, (2) differences in parents' speech to boys and
to girls, and (3)' emerging sex-associated differences in the speech of
children.
 Sex differences in parent-child interaction 193

Research setti ngs

The research was carried out in both naturalistic and laboratory settings.
Initially, we obtained a small sample of families whom we visited in
their homes, making audiorecordings of family interaction. At the same
time, for comparison, we made recordings of male and female teachers
in a day-care setting (Gleason, 1975). We then obtained funding to
conduct a laboratory and home study of a much larger sample of families.
Twenty-four families participated in this study. Each family had a child
between the ages of 2 and 5; the mothers were the primary caretakers,
and the fathers worked outside the home in professional occupations.
Twelve of the child subjects were girls and twelve were boys, about
evenly matched for age.

Methods

In the laboratory portion of the study, each child was seen and video­
taped twice, once with the father and once with the mother, in a coun­
terbalanced design. Recording sessions lasted a half hour, which was
divided among three activities: "reading" a picture book that had no
words (Mercer-Mayer's The Great Cat Chase); taking apart (and at­
tempting to put back together) a toy Playskool car; and playing store
with a number of grocery items, paper bags, and a toy cash register.
Toward the end of the session a research assistant entered the laboratory
playroom with a gift for the child. This assistant followed a script, de­
signed to maximize the likelihood that the parent and child would say,
"Hi," "Thanks," and "Goodbye" (see our article of that name: Greif
and Gleason, 1980). This was accomplished by, for instance, holding
out the gift; saying, "Here's a little gift just for you"; and then waiting
expectantly. Obviously, the pressure on a parent under the circumstan­
ces is to tell the child to say "Thanks" or personally to say "Thanks. "
In this way we were able to look at sex differences in politeness behavior
in fathers, mothers, girls, and boys.
The laboratory videotapes were transcribed and analyzed in all of the
standard ways (e.g., for mean length of utterance and sentence type),
as well as for features thought to be differentially represented in the
speech of females and males. We looked for tag questions, for instance,
as in "It's hot in here, isn't it?," a construction often claimed to be used
more by women than by men.
The home and day-care studies relied only on audiotapes, since we
194 JEAN BERKO GLEASON
felt that taking a videocamera into subjects' dining rooms was too in­
trusive; the same held true in the day-care center. Since these studies
have been reported in detail elsewhere, the major findings, along with
their implications for the study of sex differences in language, will be
reported here rather than the means and standard deviations associated
with their statistical analyses. The interested reader is referred to Bel­
linger and Gleason, 1982; Gleason, 1973, 1975, 1980; Gleason and Greif,
1983; Gleason and Weintraub, 1976; and Masur and Gleason, 1980.
These report on both the laboratory and home studies. It should be
added here that the twenty-four families who participated in the labo­
ratory study were also seen at home: A recording of a family dinner
where both parents and the child were present was made in each family.
Our current work centers on these dinner transcripts (Gleason, Peri­
mann, and Greif, 1984).

Differences in the speech of mothers and fathers to girls and boys

Home studies

Our first study was of several families at home. Like other researchers,
we found that there were very few substantive differences in the speech
of the mothers and fathers; but there were some notable exceptions. It
should be noted here that in this first home study a male research
assistant participated and remained with the family while the recording
was made. This may have led to some exaggerated "macho" behavior
m the part of the fathers. In our later home study, where we recorded

1 he dinner interactions of our twenty-four families, we learned to leave

small cassette player with them, with the instruction to turn it on when
they were about to have dinner. This much less intrusive method resulted
in what seemed to be a more natural interaction.
In the first study (with the male assistant in attendance) we found
that the syntactic measures of males and females were roughly equiv­
alent. The only real syntactic difference was that the mean length of
utterance of the fathers was less closely related to the child they were
addressing than the mothers' MLU was to the child they were address­
ing. This seemed to reflect two things. The first is that the mothers
appeared to be more "in tune" with their child than the fathers. Other
evidence for this lies in the fact that in our study and others (see Stein,
1976) mothers were better able to understand what their child was saying
than fathers were, and misunderstood less often than fathers did. The
second thing that the disparity between father and mother MLU reflects
is the fact that in the home sample fathers, especially when they talked
 Sex differences in parent-child interaction 195
to their sons, used many more imperatives than mothers did. Since an
imperative lacks a subject (e. g. , "Stop that"), the MLU is quite short.
The fathers used many more direct imperatives than the mothers did,
especially when talking to their sons, and so had disproportionately short
MLUs when talking to boys. In one family, which had two children, the
father had a longer MLU when talking with his 3-year-old daughter than
with his 5-year-old son. Needless to say, he used many imperatives with
his son.
In these home samples, the fathers produced approximately twice as
many direct imperatives as the mothers - in fact, 38 percent of the
fathers' utterances were in this form. Mothers were more likely to couch
their imperative intentions in conventionalized polite forms (e. g. ,
"Would you take your plate off the table, sweetie?"). It should be noted
that this tendency to give orders is somewhat mitigated in the laboratory,
where the differences in imperatives were slight, since, apparently, pub­
lic behavior is a good deal more polite than private behavior at home.
The other major differences we found, which appear to be robust,
were in choice of lexical items: Fathers used rarer vocabulary than
mothers. Again, this finding has been replicated elsewhere, including
in our own laboratory, where a father talking to his very young son
referred, for instance, to a "construction site." The fathers at home also
used a number of rather disparaging terms of address with their sons:
One father called his son "dingaling"; another referred to his pre­
schooler as "nutcake" and "Magoo." Again, we observed similar no­
menclature in the laboratory, where one father called his son "wise
guy." The fathers at home also had a tendency to threaten their sons.
e. g. , "Don't go in there again or I'll break your head. " We did not see
threats in the laboratory. Thus, rare lexemes, direct imperatives, threats,
and rather pejorative names marked the fathers' speech at home. All
of these features except the rare vocabulary were more likely to be
found in speech to boys than to girls.
Unfortunately, these results have been neither replicated nor discon­
firmed by others, since home studies are rare. Each of the features
mentioned is a good candidate for emerging differences in the speech
of boys and girls; but, with the exception of imperatives, which will be
discussed shortly, these features have not been found (or sought) in the
language of young children. If fathers serve as models for their sons,
we would expect young boys, when compared with girls, to use more
threats, more imperatives, more "funny" names. The research remains
to be done, however, and if early sex differences of this type are found,
they cannot all be attributed to the inlluence of parents, since children
also find models among their peers and in society's stereotypical rep­
resentation of the sexes.
196 JEAN BERKO GLEASON

Laboratory studies

A number of our own studies in the laboratory revealed sex differences

in parents' speech, some of them also confirming earlier home findings.
In addition to those already discussed, several might be mentioned here.

Lexical differences. In this study (Masur and Gleason, 1980) we looked

at the speech of mothers and fathers to their children in the laboratory
situation where they played with an automobile that could be disassem­
bled. The auto, of course, is not a neutral toy; most observers would
agree that it is male-oriented. In their conversations with their children
in this situation there were several differences between mothers and
fathers but few in the way boys and girls were addressed. Basically, the
fathers were more likely than the mothers to provide the actual names
of the car parts and to ask their children to produce them. The fathers
were more cognitively demanding, expecting their children to display
their knowledge by naming the car parts and tools and explaining their
function. Thus, in this laboratory situation, fathers modeled different
behavior from mothers while treating boys and girls in roughly equiv­
alent fashion.
The mothers, rather than name the car parts or associated tools, said
things like "That's the turn thing" rather than "wrench," and they
frequently referred to nuts as bolts and vice versa. If children follow
the models of their same-sex parent, we might expect to find among the
emefging differences in children's language a greater specificity and
demandingness among boys, at least when dealing with topics perceived
as in their domain, such as sports and tools, and more lexical vagueness
among girls, perhaps reflected in the use of more general vocabulary ­
more use of words like "thing" and "whatsis. "

Politeness. In this-study, we looked at politeness behavior in the labo­

ratory situation described above, where the child was given a gift (Greif
and Gleason, 1980). We did not find differences in the ways boys and
girls were treated: Both sexes were encouraged (actually urged) to say
thank you when given a gift and, to a lesser degree, to say "Hi" and
"Good-bye" as well. There were large differences in the mothers' and
fathers' own politeness behavior, however: The mothers were much
more likely to say thank you to the assistant when their child was given
a gift than the fathers were. There were also some differences in the
children's own spontaneous behavior: Boys were more likely to say "Hi"
than girls were. If parents provide models for their same-sex children,
we would, therefore, expect to see in the emerging language of the little
 Sex differences in parent-child interaction 197

girls more conventional expressions of politeness than in the speech of

boys. The greater percentage of greetings in the speech of boys is quite
interesting, since it may reflect either less shyness on the part of boys,
which itself may be related to sex-role expectations, or it may be more
directly related to the fact that there is more emphasis on greeting
behavior among males than among females in our society - adult males
are obliged, for instance, to rise and offer their hands, whereas females
have a great deal more latitude in what is permissible. It is impossible
to discuss differential modeling of greeting behavior by parents, since
virtually all of the parents said hello when the assistant said hello to
them.

Interruptions. This study (Greif, 1980) examined the frequency of in­

terruptions in the speech of parents and children across the three lab­
oratory situations (playing store, reading a book, taking apart the toy
auto). Greif looked at both simultaneous speech, where the speakers
overlap, and outright interruptions, where one speaker wrests the floor
from another. Here she found differences in the amount of interruption
experienced by girls and boys, as well as differences in the number of
interruptions produced by mothers and fathers. Fathers interrupted their
children more than mothers did, and both fathers and mothers inter­
rupted little girls more than they interrupted little boys. Children in­
terrupted their parents less frequently than parents interrupted their
children, despite a cultural belief that it is children who often interrupt.
Given the large number of interruptions produced by parents, it seems
likely that the cultural strictures against interruptions are more related
to status considerations involving who may interrupt whom than to a
belief that interrupting per se is unacceptable behavior. Interruptions
are therefore another area where we might look for emerging sex dif­
ferences in the language of girls and boys, but they might be expected
to be found only in certain permissible contexts, as, for instance, among
peers.

Directives. A competent speaker can produce many different surface

structures in order to express a directive intent. We looked at the pro­
duction of directives in the situation where parent and child were playing
with the toy car (Bellinger and Gleason, 1982). Three forms of directives
in particular were examined: direct imperatives, conventionalized polite
imperatives that occur in question form, and implied indirect impera­
tives. Examples are:
Direct imperative: "Turn the bolt with the wrench."
Conventionalized polite imperative: "Could you turn the bolt with the
wrench?"
Implied indirect imperative: "The wheel is going to fall off."
198 JEAN BERKO GLEASON
In this study we found that mothers were more likely to use the
question forms and that fathers' speech had a higher proportion than
mothers' speech of both the direct imperatives and the implied indirect
forms. If children follow the model of their same-sex parent, we would
expect little girls to use the polite question forms more and little boys
to begin to use the direct imperatives and the implied forms more than
the girls. As they get older, we would expect, for instance, that females
would say, "Would you move your car, please?" to a person who has
double-parked next to them, whereas males would be more likely to
say, "Your car is blocking mine. " Looking at the speech of the young
children in our laboratory situation (for purposes of this study ten fam­
ilies were used, five with boys and five with girls}, we found that by the
age of 4 these young children were indeed producing the same forms
of directives as their same-sex parent: Boys produced more direct im­
peratives and more implied forms than girls, and girls produced more
polite question forms than boys.

Sex differences in the language of children

As this paper has tried to indicate, very little work has been done on
the emergence of sex differences in the language of children. Yet we
know that since men and women speak differently, those differences ·

must begin to emerge at some point in time. The research cited here
has described differences in parents' speech to children in the use of
jocular names, threats, directives, complex vocabulary, politeness, and
interruptions. For directives, we were able 'to show that preschoolers
were already stylistically similar to their same-sex parent. Some re­
searchers (e. g. , Lakoff, 1973a) have suggested that all children speak
"women's language" until the age of 5 or 6, but this is probably a
reflection less of the facts than of our lack of sophisticated methods of
analysis and gob d hypotheses about what to look for when seeking
differences: Difef rences can be found in the speech of even very young
children when we have precise features to investigate. This observation
is provocative: Developmental psychologists representing various the­
oretical schools (Freudian, cognitive, social learning) have suggested
that children do not have a firm sex-role identity until about the age of
5. Others (e.g., Money and Erhardt, 1972) have pointed out, however,
that after the age of 18 months it is very difficult to reannounce the sex
of a child if there has been an initial misidentification; that is, a child
who was thought to be of one sex cannot after this point easily make
the transition to the other sex, even when the chromosomes say it must .
be so. Perhaps one of the reasons for this is that the child has already
 Sex differences in parent-child interaction 199
begun to absorb sex-role related behaviors at some level. These may
include specific linguistic features.
This chapter has attempted to suggest some areas for further research.
More careful research on the language of parents and other adults is
certainly in order, but it is also time to turn our attention to the emerging
language of children in order to find the earliest evidences of linguistic
sexual dimorphism.
In doing this research it will be important to examine children's lan­
guage in a variety of contexts that allow us to separate out age, sex­
role, and status considerations. Since, for instance, status factors militate
against children's using imperatives with their parents and with older
people, one area to look for these differences is in peer language. The
speech children use when talking with one another is, of course, a
separate register itself, and one that has hardly been studied. Some of
the features of peer language must surely be learned from other children
or other models: Male parents rarely make noises like dive-bombers or
machine guns; yet these sound effects are common in the speech of
young boys and not in that of young girls. By the same token, male and
female teenagers undoubtedly use different features in their speech.
Since communicative competence requires appropriate linguistic use
even before adulthood, all of these populations are worth studying in
order to understand the nature of sex differences in language. Transitory
phenomena should be noted, along with those enduring features that
ultimately mark as distinct the language of grown women and men.