Chapter 2

The Role of Interaction Formats in

Language Acquisition

Jerome Bruner

Learning a native language is an accomplishment within the grasp of any

toddler, yet discovering how children do it has eluded generations of
philosophers and linguists. I would like to take this opportunity to ask anew
some puzzling questions about what it is, beyond a splendid nervous system,
that makes it possible for the young child to acquire language so swiftly and so
effortlessly. Perhaps they are no longer puzzling questions save to those of us
who have spent a great deal of time working and brooding over whether the
acquisition of knowledge about the social world and about the world generally
is in some sense constitutive of language.
The awkward dilemma that plagues questions about the original nature and
later growth of human faculties inheres in the unique nature of human
competence. For human competence is both biological in origin and cultural in
the means by which it finds expression. While the capacity for intelligent
action has deep biological roots and discernible evolutionary history, the
exercise of that capacity depends upon man appropriating to himself modes of
acting and thinking that exist not in his genes but in his culture.
I shall argue in this chapter that language acquisition "begins" before the
child utters hisl first lexicogrammatical speech. It begins when mother and
infant create a predictable format of interaction that can serve as a microcosm
for communicating and for constituting a shared reality. The transactions that
occur in such formats constitute the "input" from which the child then masters
grammar, how to refer and mean, and how to realize his intentions
communicatively.
The child, however, could not achieve these prodigies of language acquisi-
tion without, at the same time, possessing a unique and predisposing set of
language-learning capacities-something akin to what Noam Chomsky has

lMasculine pronouns are used throughout this chapter to refer to the child; feminine
pronouns refer to the mother.

J. P. Forgas (ed.), Language and Social Situations

© Springer-Verlag New York Inc. 1985
32 Jerome Bruner

called a Language Acquisition Device (LAD). But the infant's Language

Acquisition Device could not function without the aid given by an adult who
enters with him into a transactional format. That format, initially under the
control of the adult, provides a Language Acquisition Support System
(LASS). It frames or structures the input of language and interaction to the
child's Language Acquisition Device in a manner to "make the system
function." In a word, it is the interaction between LAD and LASS that makes
it possible for the infant to enter the linguistic community-and, at the same
time, the culture to which the language gives access. The remainder of this
chapter is an amplification of how this process works. .

Two Conflicting Views of Language Acquisition

Saint Augustine believed that language acquisition was quite simple. Allegedly
recollecting his own childhood, he said: "When they named any thing, and as
they spoke turned towards it, I saw and remembered that they called what one
would point out by the name they uttered .... And thus by constantly hearing
words, as they occurred in various sentences, I collected gradually for what
they stood." But a look at children as they actually acquire language shows
Saint Augustine to be far, far off target. Alas, he had a powerful effect both on
his followers and on those who set out to refute him.
Developmental linguistics is now going through rough times that can be
traced back to Saint Augustine as well as to the reactions against him. It is one
of the mysteries of Kuhnian scientific paradigms that this empiricist approach
to language acquisition persisted in psychology (if not in philosophy, where it
was overturned by Frege and Wittgenstein) from its first enunciation by Saint
Augustine to its most recent form in B. F. Skinner's Verbal Behavior. It would
be fair to say that the persistence of the mindless behavioristic version of
Augustianism finally led to a readiness, even a reckless readiness, to be rid of
it. For it was not only an inadequate account, but one that damped inquiry by
its domination of "common sense." It set the stage for the Chomskyan
revolution.
It was to Noam Chomsky's credit that he boldly proclaimed the old
enterprise bankrupt. In its place he offered a challenging, if counterintuitive,
hypothesis based on nativism. He proposed that the acquisition of the
structure of language depended upon a Language Acquisition Device that had
as its base a universal grammar or a "linguistic deep structure" that humans
know innately and without learning. LAD was programmed to recognize in
the surface structure of any natural language encountered its deep structure
by virtue of the kinship between innate grammar and the grammar of natural
languages. The universal grammatical categories that programmed the LAD
were in the innate structure of the mind. No prior nonlinguistic knowledge of
the world was necessary, and no privileged communication with another
speaker was required. The only constraints on rate of linguistic development
The Role of Interaction Formats in Language Acquisition 33

were psychological limitations on performance: the child's limited but growing

attention and memory span, for example. Linguistic competence was there
from the start, ready to express itself when performance constraints were
extended by the growth of requisite skills.
It was an extreme view, but in a stroke it freed a generation of
psycholinguists from the dogma of association-cum-imitation-reinforcement.
It turned attention to the problem of rule learning, even if it concentrated only
on syntactic rules. By declaring learning theory dead as an explanation of
language acquisition (one of the more premature obituaries of our times), it
opened the way for a new account.
George Miller put it well. We now had two theories of language acquisition:
One of them, empiricist associationism, was impossible; the other, nativism,
was miraculous. But the void between the impossible and the miraculous was
soon to be fIlled in, albeit untidily and partially.

Initial Cognitive Endowment

If we are to consider the transition from prelinguistic communication to
language, particularly with a concern for possible continuities, we had better
begin by taking as close a look as we can at the so-called "original
endowment" of human beings. Might that endowment affect the acquisition
and early use of language?
Let me begin with some conclusions about perception, skill, and problem-
solving in the prelinguistic infant, and consider how they might conceivably
predispose the child to acquire "culture" through language.
The first of these conclusions is that much of the cognitive processing going
on in infancy appears to operate in support ofgoal-directed activity. From the
start, the human infant is active in seeking out regularities in the world about
him. The child is active in a uniquely human way, converting experience into
species-typical means-end structures.
To say that infants are also "social" is to be banal. They are geared to
respond to the human voice, to the human face, to human action and gesture.
Their means-end readiness is easily and quickly brought into coordination
with the actions of their caretakers. The pioneering work of Daniel Stern and
Berry Brazelton and their colleagues underlines how early and readily
activated infants are by the adults with whom they interact and how quickly
their means-end structuring encompasses the actions of another. The infant's
principal "tool" for achieving his or her ends is another familiar human
being.
Infants are, in a word, tuned to enter the world of human action. Obvious
though the point may seem, we shall see that it has enormous consequences
for the matter at hand. This leads directly to the second conclusion, which
concerns infant "endowment."
An enormous amount of the child's activity during the first year and a half
34 Jerome Bruner

of life is extraordinarily social and communicative. Social interaction appears

to be both self-propelled and self-rewarding. Many students of infant
behavior, like Tom Bower, have found that a social response to the infant is
the most powerful reinforcer one can use in ordinary learning experiments.
Conversely, withholding social response to the infant's initiatives is one of the
most disruptive things one can do-an unresponding face, for example, will
soon produce tears.
While the infant's attachment to the mother (or caretaker) is initially
assured by a variety of innate response patterns, there very quickly develops a
reciprocity that the infant comes to anticipate and count on. For example, if
during play the mother assumes a sober immobile face, the infant shows fewer
smiles and turns his head away from the mother more frequently than when
the mother responds socially, as Edward Tronick and his colleagues have
shown. The existence of such reciprocity-buttressed by the mother's
increasing capacity to differentiate as infant's "reasons" for crying as well as
by the infant's capacity to anticipate these consistencies-soon creates a form
of mutual attention, a harmony or "intersubjectivity," whose importance we
shall take up later.
In any case, a pattern of inborn initial social responses in the infant, elicited
by a wide variety of effective signs from the mother, is soon converted into a
very complex joint anticipatory system that converts initial biological
attachment between mother and child into something more subtle and more
sensitive to individual idiosyncracies and to forms of cultural practice. The
third conclusion is that much ofearly infant action takes place in constrained,
familiar situations and shows a surprisingly high degree of order and
"systematicity." Children spend most of their time doing a very limited
number of things. Long periods are spent in reaching and taking, banging and
looking, etc. Within anyone of these restricted domains, there is striking
"systematicity." Object play provides an example. A single act (like banging)
is applied successively to a wide range of objects. Everything on which the
child can get his hands is banged. Or the child tries out on a single object all
the motor routines of which he or she is capable-grasping the object,
banging it, throwing it to the floor, putting it in the mouth, putting it on top of
the head, running it through the entire repertory. There may be differences of
opinion concerning the "rules" that govern this orderly behavior, but there can
be no quarrel about its systematicity.
It is not in the least surprising, in light of this conclusion, that infants enter
the world of language and of culture with a readiness to find or invent
systematic ways of dealing with social requirements and linguistic forms. The
child reacts "culturally" with characteristic hypotheses about what is required
and enters language with a readiness for order.
There are two important implications that follow from this. The first is
obvious, though I do not recall ever having encountered the point. It is that
from the start, the child becomes readily attuned to "making a lot out of a
The Role of Interaction Formats in Language Acquisition 35

little" by combination. He typically works on varying a small set of elements

to create a larger range of possibilities.
The second implication is more social. The acquisition of prelinguistic and
linguistic communication takes place, in the main, in the highly constrained
settings to which we are referring. The child and his caretaker readily combine
elements in these situations to extract meanings, assign interpretations, and
infer intentions. A decade ago there was considerable debate among
developmental linguists about whether in writing "grammars" of child speech
one should use a method of "rich interpretation" -taking into account not
only the child's actual speech but also the ongoing actions and other elements
of the context in which speech was occurring. Today we take it for granted
that one must do so. For it is precisely the combining of all elements in
constrained situations (speech and nonspeech) that provides the road to
communicative effectiveness. It is for this reason that I shall place such heavy
emphasis on the role of "formats" in the child's entry into language.
A fourth conclusion about the nature ofinfant cognitive endowment is that
its systematic character is surprisingly abstract. Infants during their first year
appear to have rules for dealing with space, time, and even causation. A
moving object that is transformed in appearance while it is moving behind a
screen produces surprise when it reappears in a new guise.
Objects explored by touch alone are later recognized by vision alone. The
infant's perceptual world, far from being a blooming, buzzing confusion, is
rather orderly and organized by what seem like highly abstract rules.
It is not the case that language, when it is encountered and then used, is the
first instance of abstract rule following. It is not, for example, in language
alone that the child makes such distinctions as those between specific and
nonspecific, between states and processes, between "punctual" acts and
recurrent ones, between causative and noncausative actions. These abstract
distinctions, picked up with amazing speed in language acquisition, have
analogues in the child's way of ordering his world of experience. Language will
serve to specify, amplify, and expand distinctions that the child already has
about the world.
These four cognitive "endowments"-means-end readiness, transactional-
ity, systematicity, and abstractness-provide foundation processes that aid
the child's language acquisition. None of them "generates" language, for
language involves a set of phonological, syntactic, semantic, and illocutionary
rules and maxims that constitute a problem space of their own. But linguistic
or communicative hypotheses depend upon these capacities as enabling condi-
tions.
Such sensitivity grows in the process of fulfIlling certain general, non-
linguistic functions-predicting the environment, interacting transactionally,
getting to goals with the aid of another, and the like. These functions are first
fulfilled primitively if abstractly by prelinguistic communicative means. Such
primitive procedures, I will argue, must reach requisite levels of functioning
36 Jerome Bruner

before any Language Acquisition Device (whether innate or acquired) can

begin to generate "linguistic hypotheses."

Support for Language Acquisition

We can say, I think, that the last decade of research strongly supports the view
that language acquisition is aided by the acquirer gaining world knowledge
concurrently with or in advance of language, and is aided also by maturation
and by a privileged social relationship between the child and an adult who is
moderately well tuned to the child's linguistic level. If there is a Language
Acquisition Device, the input to it is not a shower of spoken language but a
highly interactive affair shaped, as we have already noted, by some sort of an
adult Language Acquisition Support System.
The view that acquisition depends upon interaction as the clue-giving source
for language acquisition has several variants. The most recent grows out of
speech-act theory. Its central argument is that prelinguistic infants already
know, say, how to delare and demand (e.g., Clark & Clark, 1977) by means
other than language-by gesture and intonation, for example. Mastering the
more conventional linguistic forms for carrying out these acts is a matter of a
substituting new linguistic procedures for old nonlinguistic ones with the aid
and/or modeling of an adult who already knows the language and its social
conventions. Much of the literature on Motherese deals with how this is
presumably brought about (e.g., Snow & Ferguson, 1977). Two of my own
studies are typical of this approach (Ninio & Bruner, 1978; Bruner, Ratner, &
Roy, 1982). The question that such studies pose, to out it baldly, is whether
the prelinguistic communicative functions that the child can fulfill before the
development of language proper are either constitutive of the language that he
is about to learn, or whether they even provide any clues to the aspiring
learner about the formal structure of language.
To this point we have been presupposing that the child is operating pretty
much of his own initiative, even in social interaction and certainly in use of
previously acquired world knowledge as a guide. Insofar as the adult partner
has come into the picture, it is rather as a model from whom the child can get
an input of the language in order to make his or her own inductions, piercing
discoveries, or intuitive recognitions, depending upon what view you may take
of the process. But may not the adult herself arrange the environment and her
encounters with the child in ways that scaffold language input and interaction
to make it better fit the child's "natural" way of proceeding? She, after all,
knows the language that the child is trying to master and she probably has an
implicit and practical theory about how to help the child learn it. There may, in
a word, be a Language Acquisition Support System that readies or that
formats the input of language to the child in a way that makes its rules more
transparent to the child's Language Acquisition Device. LASS, so to speak,
helps LAD. We shall want to reexamine the fine-tuning hypothesis later with
this in mind.
The Role of Interaction Formats in Language Acquisition 37

One final point before turning to a more detailed consideration of those

already introduced. Beyond early world knowledge and social interaction as
aids to the acquisition of language, there is one other possibility, one
particularly well developed in a recent paper by Shatz (1982). Might it be the
case that, though language is a problem space of its own, there are certain
general cognitive processes developed in the child's early years that can also
serve him or her in cracking the linguistic code? This is not the same as
Piaget's argument that cognitive development produces language as one of its
spinoffs or symptoms. Rather, it asserts that common processes are involved
in acquiring world knowledge, skilled social interaction, and language.
Examples are not hard to find: Learning to decompose tasks into constituent
routines and then to recombine the constituents into new procedures is, for
example, a well-observed feature of the infant's sensorimotor learning (Bruner,
1973). In a formal sense, it is the same kind of process involved in learning to
decompose and recompose the flow of language into its constituents. Or,
indeed, certain rules of perceptual attention can operate as effectively in
spotting diacritica in speech as they can for spotting distinctive sensory
features in sensory learning. Segmenting action into goal-completion cycles
may, as noted, predispose to spotting aspectual completives. Slobin's (1973)
account of how the child learns to pick up stressed, initial, and terminal
elements of words and phrases is based upon this assumption.
There is a diachronic side to this issue as well that, alas, must remain
untestable. Language systems in their earliest development probably had to be
based upon cognitive skills that were widely distributed and readily accessible
to all members of the human group, in order to assure universal participation.
If that were the case (the alternative being that the possession of language
initially marked off a human elite that were then selected for their language
ability), then the forms of language that came into being would be of a kind
that would somehow be easy for human beings to learn. This would either
mean that they were "natural' cognitive skills, or that they were peculiarly
well-matched to the mode of social interaction into which human beings
naturally entered. On a priori grounds, based on Julian Huxley's dictum of
biological redundancy in all communication systems, I find it most plausible to
accept both these propositions. There is one other good reason to give a
central place to the operation of general cognitive capabilities in the
acquisition oflanguage. So much of human language operates deictically, by
dint of using nonlinguistic context for making meaning clear, that it is difficult
to imagine how a special gift for language could have emerged independently
of other ways of processing information about the environment.

Shared Formats and Language Acquisition

I want to begin with the role of pragmatics in language generally and in
language acquisition particularly. I think of pragmatics as entailing quite
38 Jerome Bruner

different processes from those involved in mastering a set of syntactic or

semantic codes. Semantics and syntax are formulated to deal almost
exclusively with the communication of information, and that, I suppose, is
why one can refer to each as emboyding a code of elements which "stand for"
some knowledge in the "real world." Pragmatics is not restricted in that way.
It is the study of how speech is used to accomplish such social ends as
promising, humiliating, assuaging, warning, declaring, requesting. Its elements
do not "stand for" anything: They are something. Even silence, though it
cannot be specified syntactically or semantically, may speak volumes in the
context in which it occurs. It is cenainly not just like a grammatical deletion
rule where patterned absence implies presence. In this perspective, language is
a vehicle for doing things with and to others, many of which could not be
conceived but for language. Pragmatics deals, then, with the extension of
social interaction by the use of speech. It is a commitment to social interaction
by the use of speech.
In this view, pragmatics necessarily relates to discourse and, at the same
time, is always context dependent, that is, dependent upon a shared context.
Discourse presupposes a reciprocal commitment between speakers. It is a
complex commitment that includes at least three elements: (1) a shared set of
conventions for establishing speaker intent and listener uptake, including
procedural conventions like those proposed by Grice (1975) in his celebrated
discussion of Conversational Principles; (2) a shared basis for exploiting the
deictic possibilities of spatial, temporal, and interpersonal context, subject to
"shifting" in lakobson's classic sense (1971-79); and (3) a conventional
means for jointly establishing and retrieving presuppositions. These three
elements-announcement of intention, regulation of deixis, and control of
presupposition-give discourse its future, present, and past orientations.
A great many acts of discourse will be found to be ways of "tuning" these
forms of reciprocal commitment. Indeed, some linguistic theorists have even
proposed that the grammatical categories of language exist, inter alia, to
assure such tuning and calibration as well as to assure reference and meaning.
Benveniste (1971) raised the question of the function served by personal
pronouns, a universal feature of all known languages. Why are they needed,
he asked, when in fact we could accomplish the same semantic ends more
reliably by using nominals to specify people or objects rather than having to
employ tricky, shifter pronominals. His answer, of course, was that shifters
like "I" and "you" serve as economical ways of sharing and calibrating the
perspectives of two speakers through reciprocal role shift.
I think it will be apparent from the foregoing that pragmatics of discourse
cannot be based upon ordinary grammatical categories alone. For grammer is
traditionally based upon the concept of the sentence and on "sentence parts."
Yet the performance, deictic, and presuppositional rules of discourse depend
for their power upon the privileges of occurrence of expressions in discourse,
not just in individual sentences. You will perhaps recall that the object of the
The Role of Interaction Formats in Language Acquisition 39

Prague School was to derive sentence grammar constitutively from discourse

such that, for example, topic and subject were said to be the "given" in
discourse and comment or predicate the "new."
There is another sense in which interaction motivates grammatical rules.
One of the major tasks in interacting with another is the regulation of joint
attention. Fillmore (1977) proposes that the function of sentence grammars is
to establish a perspective on a scene that the sentence depicts or represents.
Perspective setting demands forefronting and backgrounding for the direction
of attention, and there are many grammatical devices for accomplishing these
ends, such as subject placement, passivization, clefting, etc.
We will see in a moment that early interaction abounds in procedures for
regulating attentional perspective on scenes in the form of vocatives,
demonstratives, pointing gestures, intonation contours, etc., employed by both
adult and child. It also abounds in role shifting and in the other forms of
discourse tuning to which I have referred. This brings me to the central
issue.
All of this leads to the hypothesis that in order for the young child to be
clued into the language, he must first enter into social relationships of a kind
that function in the manner consonant with the uses of language in
discourse-relating to intention sharing, to deictic specification, and to the
establishment of presupposition. Such a social relationship I shall call a
format. The format is a rule-bound microcosm in which the adult and child do
things to and with each other. In its most general sense, it is the instrument of
patterned human interaction. Since formats pattern communicative inter-
action between infant and caretaker before lexico-grammatical speechs
begins, they are crucial vehicles in the passage from communication to
language. Let us consider their nature in more detail.
A format entails formally a contingent interaction between at least two
acting parties, contingent in the sense that the responses of each member can
be shown to be dependent upon a prior response of the other. Each member of
the minimal pair has a goal and a set of means for its attainment such that two
conditions are met: first, that a participant's successive responses are
instrumental to that goal, and second, that there is a discernible stop order in
the sequence indicating that the terminal goal has been reached. The goals of
the two participants need not be the same; all that is required is that the
conditions of intraindividual and interindividual response contingency be
fulfilled. Formats, defined formally in this sense, represent a simple instance of
a "plot" or "scenario."
Formats, however, grow and can become as varied and complex as
necessary. Their growth is effected in several ways. They may in time
incorporate new means or strategies for the attainment of goals, including
symbolic or linguistic ones. They may move toward coordination of the goals
of the two partners not only to agreement, but also to a division of labor and a
division of initiative. And they may become conventionalized or canonical in a
40 Jerome Bruner

fashion that permits others within a symbolic community (e.g., a "speech

community") to enter the format in a provisional way in order to learn its
special rules.
Formats are also modular in the sense of being amenable as subroutines for
incorporation in larger-scale, longer-term routines. A greeting format, for
example, can be incorporated in a larger-scale routine involving other forms of
joint action. In this sense, any given format may have a hierarchical structure,
the parts being interpretable in terms of their placement in a larger structure.
The creation of higher-order formats by incorporation of subroutine formats
is one of the principal sources of presupposition. What is incorporated
becomes implicit or presupposed.
Formats, save when highly conventionalized, cannot be identified inde-
pendently of the perceptions of the participants. In this sense, they have the
property of contexts generally in being the result of joint definition by the
participants. The communal definition of formats is one of the major ways in
which a community controls the interaction of its members. Once a format is
conventionalized and "socialized," it comes to be seen as having objective
status. Eventually, formats provide the basis for speech acts and can be
reconstituted as needed by linguistic means alone.
One special property of formats involving an infant and an adult (though it
may be a property of formats in general) is that they are asymmetrical with
respect to the "consciousness" of the members, with one "knowing what's up,"
and the other knowing not or knowing less. Consciousness in this sense is not
intended to imply psychological heavy weather. I hope I can make that clearer
later. I intend it in the sense used by Vygotsky (1962) when he discussed how
the adult helps the child achieve realization of the Zone of Proximal
Development. The adult serves as model, scaffold, and monitor until the child
can take over on his own. A good illustration of this is provided in a study by
Kaye and Charney (1980) in which the adult takes over the function of
keeping turns in discourse alternation until the child develops the procedures
necessary to do so on his own.
Let us now return to the three rubrics with which I introduced the idea of
discourse: intentions, deixis, and presupposition. With respect to the goal-
oriented aspect of formats, early formats usually involve joint, overt activity
with a clear-cut, ritualized, successive structure (e.g., games like Hide-and-
Seek, Give-and-Take, Peek-a-Boo, etc.) As my colleagues and I (Bruner &
Sherwood, 1976; Ratner & Bruner, 1978) have tried to show, signaling marks
the successive steps toward the final goals of these games, with such aspectual
completives as "All-Gone" and "Dere" among the first on the scene, much like
the young Brazilians reported by Campos (1979). Once children learn to
respond to these action formats, they soon learn to call them up and to expect
uptake. The signaling becomes increasingly conventionalized and consensual
(with the mother imitating the child more often than in the reverse direction)
and the child increasingly takes over initiative. (Bruner, 1978). As the
The Role of Interaction Formats in Language Acquisition 41

signaling becomes more adept, it begins to pace the game rather than being
merely an accompaniment to it (Bruner, 1978).
Before the second year is far advanced, the child-mother pair are well
launched not only into games but also into procedures for realizing basic
linguistic functions like indicating and requesting. Consider the case of
requesting reported by Bruner, Ratner, and Roy (1982). There is a long
preliminary period in which Richard, one of the children they observed, first
mastered ways of requesting nearby visible objects by pointing and then by
intonation on an appropriate nominal; then invisible objects by indicating
direction or canonical locus of objects; then on requesting assistance in
carrying out actions: "invitations" and offers. This is particularly difficult, for
it requires his analyzing the structure of a task and signaling that he wants
assistance, on what object, but also what kind of assistance it is. He has
learned nominals, verbs, vocatives, demonstratives-all in aid of his requests.
At 20 months old, Richard adopted a "successive guidance" strategy for
managing complex requests for assistance in action. He starts the round of
exchanges with a requestive vocative or with an intonationally marked
nominal or verb. When his mother signals uptake but incomprehension of
what kind of help he is requesting, he follows by introducing a second element,
usually a locative, to indicate the place where the objects is or the locus of the
desired action. This may be followed by a verb indicating the action requested.
And so it continues step by step until he succeeds in getting the message across
in successive steps.
By the time Richard is 22 months, however, his mother will no longer
tolerate being dealt with in this robotic way and insists that he fulflll one of the
felicity conditions on requestiong-full disclosure of intention in advance.
"No, Richard, tell me what you want first," she demands. Richard responds
with one of his first three-word sentences, strung together with slight pauses
but including "Mummy" as Agent, the required Action, and the sought-after
Object, the whole marked with what is to be the requestive intonation contour
of such utterances on later occasions. All of which is not to say that there was
anything in the prior interaction per se that could have given Richard any
clues about how to linearize such a sentence. There is no "natural order" in
action that tells you the order of corresponding elements in a sentence-
though I was once tempted to believe something like that about Subject-Verb-
Object orders (Bruner, 1975). It is the familiarity and structure of the request
formats that frees Richard and guides him in fmding the linguistic procedures
required. The adult helps hold the child's goal invariant against distraction,
reduces the degrees of freedom in the choices he has to make lexically and
grammatically by coordinating her own utterances with established action
segments, and generally serves as linguistic scaffold. Above all, she helps him
link his intentions to linguistic means for their attainment.
In the limited space available, I would like to touch very lightly on the
second feature of formats: their role in providing a base for context sensitivity
42 Jerome Bruner

and deixis. It was Grace DeLaguna (1927) who noted in her remarkable book
of more than half a century ago that you could not know what a child meant
without knowing what he or she was doing while speaking. The key to going
beyond this primitive deictic indexicality depends upon waht C. S. Peirce
(1931-59) called the transformation of "sign vehicles." (It is to Michael
Silverstein (1981) that I am particularly grateful for pointing out the relevance
of Peirce's proposal for developmental linguistics.)
Peirce proposed that initial language is indexical and necessarily deictic,
dependent upon a contiguity or "pointing" relation between sign and
significate. With the development of a sign system, a second feature is added:
Language can then operate intralinguistically in the sense that signs can point
to or be related to other signs. The context to which reference is made may still
be concrete and specific, but it is linguistic. As the child gains further insight
into the language as a codified system of representation, he comes to operate
not on concrete events, whether directly in experience or represented in words,
but upon possible combinations derived from operations on the language
itself. This last accomplishment Peirce refers to as the meta pragmatic level,
and at this point the child is able to turn around on his language, correct it as
needed, quote it, amplify what was meant, even define it. A good example of
the transition from intralinguistic to meta pragmatic speech is provided in
Maya Hickman's (1982) paper on children reporting what they had seen and
heard in an animated cartoon. This last stage of the child's development,
however, takes place later than the ages I want to consider and need not
concern us.
Let me illustrate the manner in which, thanks to the presupposition-
conserving structure of a continuing and growing format, the child and
mother switch from indexical to intralinguistic procedures. Consider how the
mother and child come to signal "given" and "new" in their interaction when
Richard is between 18 and 22 months. In the growth of labeling (cf. Ninio &
Bruner, 1978), Richard's mother sets up a routine for book reading in which
she employs four invariant discourse markers: an initial attentional vocative in
the form "Oh look, Richard"; followed, when his attention is gained, by the
query "What's that, Richard?" with stress and rising intonation peaking at the
second word; followed, if Richard should reply by any vocalization even in the
form of a babble string initially, by "Yes, that's an X"; and terminated by a
reinforcing remark like "That's very good."
At this point when Richard can reliably produce the correct label or some
phonologically constant form that his mother can imitate herself, her
intonation contour changes. For items of this "known-to-be-known" class, she
still uses her second discourse marker ("What's that, Richard?"), but now
with a falling intonation on the second word. It is as if she is signaling that she
knows that he knows, and the shift often produces "knowing smiles" between
the two. Then, shortly after, she introduces an extended routine where, after
the presuppositionally marked request for a label, she asks a second question
calling for an answer in the form of a predicate of action or of state related to
The Role of Interaction Formats in Language Acquisition 43

the child's just-provided label-for example, "What's the X doing?"-with

stress and rising intonation on the terminal word. The same sorting of given
and new can be observed in the development of request, when Richard's
mother responds to his wave toward a canonical locus, where he thinks an
object may be, with "Something in the ice box? What do you want in the ice
box?" At each opportunity the mother is cannily adding nominals or even
anaphoric pronominals to the indexical procedures the child is using, and then
using these an intralinguistically presupposed in later discourse.
In short, formats in discourse provide the necessary microcosms in which
the child can signal intentions, operate indexically and then intralinguistically,
and develop presuppositions, all within the interactions that have properties
that are easily mapped onto the functions and forms oflanguage. At the start,
formating is under the control of the adult. Increasingly formats become
symmetrical and the child can initiate them as readily as the adult. All cultures
do not, of course, format early discourse in the same way-as we know form
the pioneering work of Schieffelin (1979) on the Kaluli of Papuan New Guinea
who, unlike us, "show" more than they interact. But the hypothesis I am
putting forth is that all cultures have ways of formating interaction and
discourse so as to highlight those features of the world and of social interaction
that map most readily onto linguistic categories and grammatical rules. It is
this feature of early interaction that I have referred to as the Language
Acquisition Support System, or LASS, without which an acquisition device,
LAD, could not make much progress.

Some Conclusions
Let me return to Marilyn Shatz's (1982) discussion of the ways in which social
interaction might aid the child in developing insights into syntax-and it is
important to note that it is syntax with which she is exclusively and (I think)
properly concerned. She presents four views that can be briefly characterized
as follows: (1) Syntax is derived directly from prior social knowledge; (2)
syntax is derived from prior semantic representations that achieve deep
structure by being transformed by social interactions; (3) syntax is not derived
from social interaction but merely faciliated by the routinization of social
interaction which frees necessary processing capacity; and, as already
mentioned, (4) syntax and the complexities of rule-bound social interaction
depend upon the same types of cognitive processes at any given stage of
development, and consequently children about the learning of the two
different systems in a common way.
Shatz's analysis is useful; I can even sail happily under the flag of her last
two rubrics. Yet I find her classification system constricting for its failure to
give a full enough role to the adult and for its incomplete analysis of the nature
of the formats in which the child's interactive learning and syntactic
acquisition occur. In effect, she treats the child as if he were flying solo, in the
44 Jerome Bruner

best tradition of both theories of learning and of information processing. But

the solo model of social learning, however useful it may be in goading us to
look for internal structures and processes, is just not good enough.
What I mean by this assertion requires me to revert briefly to Vygotsky's
(1962) conception of the Zone of Proximal Development, mentioned earlier in
passing. Vygotsky comments on the child's progress from arithmetic to
algebra in mathematical learning in a way that is relevant to our discussion. It
is not possible, he notes, for the child to move to the "higher ground" of
algebra unless he has grasped enough of concrete arithmetic operations to
appreciate hints that relate to the more categorical status of these opera-
tions-that any number can be treated as an unknown, x; that while a blind
Venetian and a Venetian blind cannot be substituted for each other, three twos
can substitute for two threes; and so on. In teaching language, unlike in
teaching algebra, the tutor is by the nature of the medium bound to be implicit
or tacit in the lessons given. The progress that results is much more like that
described by Braine in the movement from rote nonproductive utterances, in
formulaic pivot grammar, to a more productive use of the same constituents
once the child has had an opportunity to master and then to extend the forms.
By means of scaffolded use, the child learns what a form can do.
Vygotsky offers the hypothesis that mastery of a lower form has as its
terminal state an increase in consciousness (or if you prefer, metacognition).
He rather picturesquely characterizes this step across the Zone of Proximal
Development as a "loan of consciousness" by the adult to the child until the
time when the child can manage on his own. It is done not only by arranging
the world suitably, but by providing "hints" and "props." Now, to the degree
to which adult and child can stay within an informative but undemanding
format, the hinting and the propping will be assimilable. And it is for this
reason that I have made so much of the role of formats as essential aids to
assisted learning.
Finally, to revert to the argument of writers like Peirce (1931-59),
Benveniste (1971), and Jakobson (1971-79) about the 'intersubjectivity" of
linguistic forms. Peirce commented particularly on the duality of symbolic
forms in natural language. They serve both to represent concepts and to
communicate them, he says. This creates complexities, since one's own
perspective differs from the perspective of an interlocutor. It was Benveniste
who noted the resort to shifters as universal means of dealing with problems of
perspective. The credit goes to Jakobson, finally, for exploring the inter-
connections of pronominal shifters in such contrastive deictic pairs as this and
that, here and there, to and from, and even verb forms like come and go.
The message that I read into the writings of these towering linguists is that it
would be impossible to learn a language without knowing in advance or
learning concurrently the perspectival complexities involved in using the same
set of symbols for representation and communication. That is why I am so
reluctant to consider language acquisition to be either the virtuoso cracking of
The Role of Interaction Formats in Language Acquisition 45

a linguistic code, or the spinoff of ordinary cognitive development, or the

gradual takeover of adult speech by the child through some impossible
inductive tour de force. It is, rather, a subtle process by which adults
artificially arrange the world so that the child can succeed culturally by doing
what comes naturally, and with others similarly inclined.

Acknowledgment. Permission to reproduce some sections from my book "Child's talk:

Learning to use language" by the publisher, W. W. Norton & Co., is gratefully
acknowledged. Copyright © 1983 by Jerome Bruner.

References

Benveniste, E. (1971). Problems in general linguistics. Coral Gables, FL: University of

Miami Press.
Bruner, J. S. (1973). Organization of early skilled action. Child Development, 44, 1-1.
Bruner, J. S. (1975). The ontogenesis of speech acts. Journal of Child Language, 2,1-
19.
Bruner, J. S. (1978). Learning how to do things with words. In J. S. Bruner & A.
Garton (Eds.), Human growth and development: Wolfson lectures, 1976. Oxford:
Oxford University Press.
Bruner, J. S., & Sherwood, V. (1976). Early rule structure: The case of "peekaboo." In
R. Harre (Ed.), Lifo sentences: Aspects of the social role of language. New York &
London: Wiley.
Bruner, J. S., Ratner, N., & Roy, C. (1982). The beginnings of request. In K. Nelson
(Ed.), Children's language (Vol. 3).
Campos, F. (1979). The emergence of causal relations and the linguistic development
of Brazilian children. Unpublished doctoral thesis, University of Campinas,
Brazil.
Clark, H. H., & Clark, E. V. (1977). Psychology and language: An introduction to
psycholinguistics. New York: Harcourt Brace Jovanovich.
DeLaguna, G. (1927). Speech: 1tsfunction and development. New Haven, CT: Yale
University Press.
Fillmore, C. J. (1977). The case for case reopened. In P. Cole & J. M. Sadock (Eds.),
Syntax and semantics: Vol. 8. Grammatical relations. New York: Academic
Press.
Grice, H. P. (1975). Logic and conversation. In. P. Cole & J. L. Morgan (Eds.), Syntax
and semantics: Vol. 3. Speech acts. New York: Academic Press.
Hickman, M. E. (1982). The implication of discourse skills in Vygotsky's develop-
mental theory. In J. Wertsch (Ed.), Culture, communication, and cognition: Vygot-
skian perspectives. New York: Academic Press.
Jakobson, R. (1971-79). Selected writings. The Hauge: Mouton.
Kaye, K., & Charney, R. (1980). How mothers maintain "dialogue" with two-year-
olds. In D. Olson (Ed.), The social foundations oflanguage and thought. New York:
W. W. Norton.
Ninio, A., & Bruner, J. S. (1978). The achievement and antecedents of labelling.
Journal of Child Language, 5, 1-15.
Peirce, C. S. (1931-59). Collected papers. Cambridge, MA: Harvard University
Press.
Ratner, N., & Bruner, J. S. (1978). Games, social exchange, and the acquisition of
language. Journal of Child Language, 5,391-401.
46 Jerome Bruner

Schieffelin, B. (1979). How Kaluli children learn what to say, what to do, and how to
feel. Unpublished doctoral dissertation, Columbia University, New York.
Shatz, M. (1982). Relations between cognition and language acquisition. In W.
Deutsch (Ed.), The child's construction of language. New York & London:
Academic Press.
Silverstein, M. (1981). Aspects offunction in culture and language. Paper presented at
SSRC Seminar on Culture and Mind, New York.
Slobin, D. (1973). Cognitive prerequisites for the development of grammar. In C. A.
Furguson & D. Slobin (Eds.), Studies of child language development. New York:
Holt, Rinehart & Winston.
Snow, C., & Ferguson, C. A. (Eds.). (1977). Talking to children: Language input and
acquisition. New York & London: Academic Press.
Vygotsky, L. (1962). Thought and language. Cambridge, MA: MIT Press.