The Epigenetic System and the
Development of Cognitive
Functions
JEAN PIAGET
1
Preformation and Epigenesis
The problem that has always arisen before one could tackle ontogenesis has been
preformation or epigenesis. With the usual veering of fashion in the history of
ideas, the tendency of many writers today is to return to the more or less strict
preformation standpoint. Their grounds for this are that the chain or helical
structure of the DNA or deoxyribonucleic acid molecule is susceptible of a
combinatorial arrangement of its elements where "combinatorial" covers, by
definition, the set of all possibilities. But if it is difficult, from the phylogenetic
point of view, to conceive of man as preformed in bacterium or virus, it is every bit
as hard to make out how, from the ontogenetic point of view, the main stages of
"determination" or induction, and, most important, of the final functional
"reintegration" of differentiated organs, could already be present in the initial
stages of segmentation. Furthermore, Waddington has stated categorically that
the idea of an entirely predetermined system in the DNA, however fashionable it
may be at the moment, is just unacceptable in embryologya. At the symposium on
this subject at Geneva in 1964, in the course of discussion about the
regulations of development, he made a very profound comparison between
epigenetic construction and a progression of geometric theorems in which each is
rendered indispensable by the sum of those preceding it, though none is
directly derived from the axioms underlying the original one.
See Waddington (1975) — cited under Further reading. (Editor's note.)
"The Epige'netic System and the Development of Cognitive Functions" is an excerpt from:
Jean Piaget, Biology and Knowledge (Edinburgh University Press and University of Chicago
Press, 1971, section 2, pp. 14-23) and is reprinted by kind permission.
32
Jean Piaget
The comparison of epigenesis with a progressive mathematical construction
comes home to us all the more forcibly because the growth of elementary
logico-mathematical operations during the ontogenesis of intelligence in a child
raises the same problem of preformation or epigenetic construction as that which
forms the basis of discussion about causal embryology.
We shall, indeed, find ourselves compelled to trace the origin of
logico-mathematical operations back to an abstraction made from the general
coordination of actions. On the one hand, such operations cannot possibly be
based on the objects themselves, since abstraction from objects can give rise only
to non-necessitous statements (in the sense of deductive necessity) or, to put it
more precisely, to judgements which are merely probable, whereas it is characteristic of logico-mathematical operations that they have an internal necessity
attributable to their complete reversibility (and therefore not physical): for
example, if i= A/ — 1, then i x i = — 1. On the other hand, reunion, order, and
interchangeable schemata are to be found in the general coordination of action,
and these constitute the practical equivalent and even the motor equivalent of
future interiorized operations.
If these elementary logico-mathematical operations are based on the coordination
of actions, by means of reflective abstraction drawn from sensorimotor schemata,
do we have to conclude that the whole of mathematics is laid down in advance to
our nervous system? Not only is this unthinkable, but the facts prove that logic
itself, even in its most "natural" forms, is by no means innate in human beings in
the sense that it exists at any age. Even the transitivity of equals or of
cumulative differences (A = C if A = B and B = C, or A < C if A < B and B <
C) is by no means obvious to a child of four to six years when he has to make a
comparison between lengths and weights on first perceiving A and B
simultaneously, next B and C, but not A and C (A subsequently being hidden and
so presenting the problem).
The task of finding out about this transitivity raises all the main problems of
epigenesis. Is this transitivity inherent in the genotype of the human species? If so,
why does it not automatically come into play at about seven or eight years (and
about nine or ten for weights)? Because, it will be said, new conditions are
indispensable if the inherent virtual is later to become actual: for example, the
intervention of regulatory genes or the collaboration of a number of genes not so
far synergic (by reference to genetic or genie coadaptation, to use the currently
accepted term). However, as these differentiated regulations are not made at any
definite age in the particular case but may be accelerated or retarded according to
conditions of exercise or acquired experience, they certainly exercise factors
which are indirectly connected with environment.
Can it then be said that transitivity is utterly unconnected with the actions of
the genome and solely dependent on phenotypic actions of the organism in
relation to environment? In that case, how can it become "necessary" and
generalizable? Because those actions which exert an influence on environment
b
Piaget defines "reflective abstraction" as the process of "reconstruction with new
combinations, which allows for any operational structure at any previous stage or level
to be integrated into a richer structure at a higher level" (section 20, part 4, p. 320 of
Biology and Knowledge). (Editor's note.)
Epigenetic System and Cognitive Functions
33
are influenced in their turn by the more generalized forms of internal coordinations of action? If that were so, would then generalized coordinations depend, in
their turn, on the most common and deep-seated coordinations of the nervous
system, which brings us back to the genome?
The evidence thus proves that the problem of preformation or epigenesis has
nothing about it that appertains specially to organic embryogenesis, and it crops up
in its most acute form every time we discuss the ontogenesis of cognitive
functions. It may be objected that the problem is settled in advance, since the
various aspects of intellectual behavior are phenotypic reactions and a phenotype
is the result of interaction between the genotype and the environment. That is
indisputable, but one still needs to explain in detail how, in the field of knowledge as
in that of organic epigenesis, this collaboration between the genome and the
environment actually works - especially those details which concern
autoregula-tions or progressive equilibrations which admit of the exclusion of
both prefor-mism and the notion of a reaction caused entirely by environment.
2
The Sequential Character of Stages
In this attempt at elucidation, the first step forward should be an examination of
the sequential character of development. We call sequential a series of stages, each
one of which is a necessary part of the whole and a necessary result of all the
stages that precede it (except for the first one), as well as naturally leading on to
the next stage (except for the last one). This seems to be the case with the
embryogenesis of Metazoa, since the main stages constantly repeat themselves in
the same order. However, no experiments have yet been done to control the
impossibility of doing away with one stage, though these will doubtless be
performed some day if someone succeeds in isolating processes which entail
considerable speeding up or slowing down of the succession of stages. A further
argument in favor of the sequential character and generality of the stages is the
fact that, in mosaic-type embryos, namely at the initial level studied, those which
have shown incomplete regeneration when separated from a blastomere reach a stage
of partial control if the seed is split at the virgin egg stage (Ascidies de Dalcq).
Now this same problem about the sequential character of stages appears again
in psychology in connection with the development of the cognitive functions. It
is important to note that in this sphere the stages became increasingly clear and
sequential in relation to controls that are better differentiated and of wider
application.
Psychologists have relied too much on the notion of stage. Some speak as
though it were nothing but a series of actions, not always, though "generally," in
a constant order, and supposedly sharing a dominant characteristic, nothing
more - which opens the door to arbitrary thinking. This is what Freud means
by stages, for example, as far as the affective is concerned.
When it comes to intelligence, however, we use the term stage where the
following conditions are fulfilled: first, where the series of actions is constant,
independently of such speeding up or slowing down as may modify the middle
range of chronological age1 in terms of acquired experience and social environment (like individual aptitude); second, where each stage is determined not
34
Jfean Piaget
merely by a dominant property but by a whole structure which characterizes all
further actions that belong to this stage; third, where these structures offer a
process of integration such that each one is prepared by the preceding one and
integrated into the one that follows. For example, without going into great detail
about particular stages, three main periods can be seen in the case of operative
intelligence:
A. A sensorimotor period (from birth up to one and one-half to two years)
during which sensorimotor schemata ranging up to acts of practical intelligence
by means of immediate comprehension (using a stick or a piece of string, etc.)
are established as well as practical substructures of future notions (permanent
object schema, spatial deplacement "group," sensorimotor causality, etc.).
B. A period that begins when the semiotic function (language, game symbols,
picture making) manifests itself and goes through the preparatory phase of
preoperative representation (nonconservation, etc.). This ends not later than the
eighth or ninth year with the setting up of operations which are called "concrete"
because they still have a bearing on objects (classifying things, putting them in
series, noting connections, understanding numbers).
C. A period beginning at about the age of eleven or twelve which is charac
terized by propositional operations (implications, etc.) with their combinatorial
quality and their possible transformations made by relation to a quaternary group
— a combination of two elementary reversibility forms (inversion or negation and
reciprocity).
A stage system of this kind (stages which can actually be even further
differentiated into substages) makes up a sequential process: it is not possible to
arrive at "concrete" operations without undergoing some sensorimotor preparation
(which explains why, for example, blind people, having badly coordinated action
schemata, may be retarded). It is also impossible to progress to propositional
operations without support from previous concrete operations, etc. Thus, one is
confronted with an epigenetic system whose stages may be characterized by
fairly precise structures: coordination of sensorimotor schemata reaching certain
invariables and an approximate reversibility (though in successive actions);
"groupings" of concrete operations, that is, those elementary structures which
are common to classifications and serializations, etc.; and combinatorial with a
quaternary group at the third degree. 2
By contrast, in the field of primary perceptions (or field effects) no comparable
system of stages is to be found, and, as to behavior of medium complexity
(perceptive activity in exploration, etc., and mental images), an intermediary
situation is found halfway between an absence of stages and stages limited by
their progressive integrations. Thus, everything seems to happen as though the
more complex - in their organization and autoregulation systems - cognitive
systems are, the more their formation is dependent on a sequential process
comparable to a biological epigenesis.
3
Chreods
If a detailed study is to be made, that is, if the evolution of broad concepts or of
particular operative structures is to be studied separately, then each one may
Epigenetic System and Cognitive Functions
35
give rise to its own respective stages in the midst of which is to be found the
same sequential process. But the interesting thing about this point is that it
presents us with differentiated channels, each one of which is nevertheless
relatively even and follows its own course while still giving proof of varied
interactions with the rest.
Waddington has suggested the name "chreods" (necessary routes) to describe
developments particular to an organ or a part of an embryo, and he applies the
term epigenetic system (or, epigenetic "scene") to the sum of the chreods, taken as
being - to a greater or a lesser degree - channeled0. But the main interest of this
idea is not just in the names he gives things (or in the symbolic patterns thereby
presented to us, of channels, some wide, some narrow, that the processes must
follow). It is, rather, in a new concept of equilibrium as something which is, as it
were, kinematic and which, in determining such processes, is nevertheless quite
distinct from homeostasis: there is a kind of "homeorhesis" when the formatory
process, deviating from its course under outside influence, is brought back on
course by the interplay of coercive compensations. In Waddington's opinion, such
a mechanism is dependent upon a network of interactions rather than upon the
action of individual genes; each group of genes is not even homeorhetic, and its
return to a moral course or chreod presupposes, in this way, a complex interplay
of regulations. It is true that some influence systematically exerted by the
environment may eventually lead to lasting deviations in the chreod and to the
consolidation of a new homeorhesis, but this is not the moment to raise such a
problem. On the contrary, we would do better to emphasize the fact that the
chreod and its homeorhesis do have a space-time aspect, not merely a space one.
Differentiation in chreods is regulated in both time and space. The various
channelings as well as the autocorrections which assure their homeorhetical
equilibrium are under the control of a "time tally," which might well be described
as a speed control for the processes of assimilation and organization. It is, then,
only at the completion of development or at the completion of each structural
achievement that homeorhesis gives place to homeostasis or functional equilibrium.
In the latter case, the question naturally arises of determining the relationship
between the two.
It is impossible to take note of such a picture without immediately thinking of
the far-reaching analogies it has with the development of schemata or ideas in
the intelligence, and with that of operational structures.
To put the matter in a familiar way, let us begin by noting that these analogies
are very far from being universally accepted; very rarely have I been able, in
America, to expound any aspect of my stage theory without being asked, "How
can you speed up this development?" And that excellent psychologist, J. Bruner,
has gone so far as to state that you can teach anything to any child at any age if
only you set about it the right way. My answer to this is in the form of two
questions: first, would it ever be possible to make the theory of relativity or even
the simple handling of propositional or hypothetico-deductive operations comprehensible to a four-year-old? And, second, why does a human baby not discover
the continued presence of something that he sees you hide beneath a screen until
he reaches the age of nine months and upward, whereas kittens (in a study made
c
See Editors note a.
Jean
36
by H. Gruber when he discovered the same preliminary stages in them as in us) do
so at three months, even though they make no further progress in coordinating
successive positions?
The truth, it seems to me, is that every notional or operational construction
implies some optimum length of time, the expression of the most favorable
transformation or assimilation speeds. This is because such a construction
contains a certain number of necessary stages whose itinerary is the equivalent of
a "chreod." In the sphere of the mind, where social influences are added to factors
of physical experience (material environment), deviations easily occur, and short
circuits too. Thus, the natural way for the mind to attain the concept of whole
numbers consists of syntheses of inclusion of classes and the sequence of transitive
asymmetrical relationships, in spite of the fact that the latter two systems develop
along partly independent lines. Now the natural structure of the number concept can
be modified in various ways. First of all, as is done by many parents, it can be taught
the child verbally - ten to twenty, etc. But this only modifies the child's
comprehension very slightly; we are constantly coming across subjects of four to
five years old who will deny the equality of two piles of objects, even though
they have counted what is in each pile as being perhaps seven or ten, because the
way the objects were arranged in space or subdivided into small groups was
changed each time. In such cases, outside influences, such as counting out loud,
only produce a slight deviation leading back to the "chreod" at the four- to
five-year-old level, for lack of any means of assimilation at higher levels. In other
cases, a genuine acceleration can be set up, but only at one point (for example, in
experiments where transfers are made one at a time in succession, thus facilitating,
by repetition of the same actions, the synthesis of inclusions and the serial
order).3 This local synthesis is not necessarily followed by comprehension, nor
will it guarantee retention of the number in transfer experiments between groups of
objects arranged differently on different planes.
Briefly, intellectual growth contains its own rhythm and its "chreods" just as
physical growth does. This is not, of course, to say that the best teaching
methods, by which we mean the most "active" ones, cannot, to a certain extent,
speed up the critical ages dealt with so far, but this speeding up cannot be
indefinitely continued.
4
Maturation and Environment
The epigenesis of the cognitive functions, like any other, does, in fact, presuppose
an increasingly close collaboration between the factors of environment and the
genome, the former increasing in importance the larger the subject grows.
The factors relative to the genome are certainly not to be left out of account, in
spite of what some scholars, empirically oriented, have said about all knowledge
being drawn from outside experience. At this stage of our knowledge, these
factors certainly cannot be tested in detail, but the best indication that they do
intervene is the fact that the maturation of the nervous system is continuous
right up to the age of fifteen or sixteen years. This, of course, in no way implies
that ready-made knowledge is written into the nervous system from the outset
in the way that "innate ideas" are, and, even if this idea proves acceptable in
Epigenetic System and Cognitive Functions
37
the case of certain instincts, there does not seem to be any similar phenomenon
where human knowledge is concerned. On the contrary, heredity and maturation
open up new possibilities in the human child, possibilities quite unknown to
lower types of animal but which still have to be actualized by collaboration with
the environment. These possibilities, for all they are opened up in stages, are
nonetheless essentially functional (having no preformed structures) in that they
represent a progressive power of coordination; but this very power is what makes
possible the general coordinations of action on which logico-mathematical operations are based, which is why the continuous maturation of the nervous system
that goes on until fifteen or sixteen years is a factor by no means to be ignored.
Such maturation does not, moreover, depend solely on the genome. But it does
depend on that among other things (with the intervention of exercise factors,
etc.), and, in general terms, it is admitted today that every phenotypic growth
(including, therefore, cognitive functions in general) is the product of close
interreactions between the genome and the environment.
The analysis of this collaboration remains, it is true, very complex and has
scarcely been touched on so far. At this point we might begin by referring to an
idea for which we are indebted, once more, to Waddington. This time it dates back
to the work he did in 1932 on the phenomena of induction in the embryos of hens
and ducks, to the idea of "competence," or the physiological state of a tissue,
which permits it to react in a specific way to given stimuli. Competence is
naturally subject to time conditions such as we talked about earlier, and a tissue
may be competent at one particular phase without having been so previously or
even remaining so afterward.
Surely no one can fail to see the analogy between this notion in relation to the
embryonic mechanism and the facts brought out by experiments in the field of
learning in logico-mathematical operations. The work of such people as Inhelder,
Sinclair, and Bovet opened this up. When mechanisms favorable to the acquisition of
knowledge are thus presented (for example, retaining the idea that there is the same
amount of liquid when changing it from one vessel to another of a different
shape), the results are utterly different according to the stage of the child's
development, and the particular presentation which causes one subject to learn
more quickly about a constant quantity will leave another utterly unmoved. The
explanation of this again lies in the fact that sensitivity to stimuli (not only
perceptual stimuli but in some cases those which set up a reasoning process) is a
function of such assimilation schemata as are available to the subject. In this case,
then, "competence" is a particular instance of what we call cognitive "assimilation,"
but assimilation schemata are built up by the interplay among the subject's powers of
coordination and by the data of experience and environment.
To put it briefly, the epigenetic process which is the basis of intellectual
operations is rather closely comparable to embryological epigenesis and the
organic formation of phenotypes. Of course, the part played by environment is
much larger, since the essential function of knowledge is to make contact with
environment. To the effects of physical environment we must add those of social
environment (for the individual genome is always the reflection of multiple
crossbreedings and of a fairly broad range of "population"). But the essential
question does not concern the quantitative sum of the respective influences
exerted by endogenitive and external factors; rather, it has to do with qualitative
Jean Piaget
38
analogies, and from that point of view it seems obvious that internal coordinations of
the necessary and constant type, which make possible the integration of exterior
cognitive aliment, give rise to the same biological problem of collaboration
between the genome and the environment as do all the other forms of organization
which occur in the course of development.
NOTES
1 In psychology the distinction is always made between chronological and mental age.
2 This sequential character of the stages of intelligence certainly seems to prove the
necessity of an endogenic factor in nervous maturation, but by no means excludes
either the intervention of the environment (experience) or, more particularly, the
interaction of environment and maturation at the center of a process of equilibration
or progressive autoregulation.
3 In this case, it was the putting of beads, simultaneously, one in each hand, into transparent
bottles. See Inhelder and Piaget, La formation des raisonnements recurrentiels, Etudes
d'epistemologie genetique, 17 (Presses Universitaires de France, 1963), chapter 2.