Fischer, K. (1980).

A theory of cognitive develoDment: The control

A newborn baby is mostly helpless anc

unable to deal with much of the work
around him. Over the years the baby grow:
into a child, the child into an adult. Ex
plaining the psychological transformatior
that the individual undergoes in these 20
odd years is one of the most challenginl
tasks facing psychology.
The theory presented in this article, callec
skill theory, attempts to explain a large par
of this psychological transformation. I
focuses primarily on cognition and intelli
gence, and it deals with aspects of learnin)
and problem solving. Skill theory treat!
cognitive development as the constructior
of hierarchically ordered collections o
specific skills, which are defined formallj
by means of a set-theory description.
Of course, other psychologists have deal
 _I_

47 8 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT

with these same general issues before, and find that stimulation facilitates physical ously are subject to the functional laws very core of the study of cognitive devel
skill tkeory w 8R *heir *atf, k€t&Bg $ew€q”t iff pxmttwe idants. . outlined by environmeiifal€y oriented psy- ment: the issues ofsequence and synchrl
concepts from the work of Piaget (19361 Despite the general agreement on the chologists. The sets that describe the skill in development. Under what circumstan
1952, 1970; Piaget, Grize, Szeminska. & interaction of organism and environment, structures are always jointly determined by will skills show invariant developme
Vinh Bang, 1968; Piaget & Inhelder, 1966/ developmental psychologists have had dif- the actions of the organism and the environ- sequences, and under what circumstan
1969), Bruner (1971, 1973), Werner (1948, ficulty incorporating both organism and mental context that supports those actions: will specific skills develop with some
1957), and Skinner (1938. 1969), informa- environment into their theories. When The organism controls its actions in a par- gree of synchrony? In practice, a the
tion-processing psychology (Case, 1974; attempting to include both, they have ef- ticular environmental context. This resolu- of cognitive development must be able
Pascual-Leone, 1970, Note 1; Schaeffer, fectively emphasized one side or the other. tion of the organism-environment dilemma predict and explain developmental
1975), and the study of skill learning (Baron, For instance, Piaget is perhaps the de- allows some progress toward explaining quences and synchronies. This is, I belie
1973; Gagne, ‘1968, 1970; Reed, 1968). The velopmental psychologist best known for and predicting cognitive development, al- the most essential criterion for evalual
intent of skill theory is to integrate ideas his interactional approach (1936/1952, 1947/ though it also raises some problems of its any theory of cognitive development.
from these various approaches to produce 1950, 1975), yet his explanatory constructs own, which will be discussed later.
a tool for explaining and predicting the de- have focused primarily on the organism. One of the most immediate implica- The Theory
velopment of behavior and thought. It is the organism that changes from one tions of defining specific skills in terms of
Before describing skill theory in detail, stage to the next, with the environment both organism and environment is that rela- Skill theory provides an abstract re1
I will discuss several of the key issues that playing only a minimal role (see Beilin. tively minor alterations in the environ- sentation of the structures of skills 1
it attempts to deal with: the relation between 1971, and Flavell, 1971a). Piaget himself mental context of action will literally emerge in cognitive development, toget
organism and environment in cognitive de- has recognized this problem: Faced with a change the skill being used. That is, the with a set of transformation rules that re.
velopment and the issues of sequence and host of environmentally induced instances organism’s control of a skill depends on a these structures to each other. The str
synchrony. The theory will then be pre- of developmental unevenness in perform- particular environmental context. This im- tures and transformation rules compns
sented quasi-formally in terms of assump- ance (called horizontal decalage; Piaget. plication should be kept in mind because it tool for explaining and predicting devel
tions, definitions, notation rules, and de- 1941), he has said that he simply cannot has many important ramifications for the mental sequences and synchronies fr
scriptions of both the hierarchical levels of explain them (Piaget, 1971, p. 11). theory and its corollaries. birth to young adulthood. As I will dem
cognitive control and the transformation At the other extreme are the behaviorists, strate later, they may also allow for
rules for development from level to level. who, like Piaget, recognize the importance Sequence and Synchrony explanation and prediction of cognil
Several experiments testing the theory will of both organism and environment. Their Within the context of this proposed development in adulthood. The theory t
be described, corollaries of the theory will explanatory constructs, however, have ef- resolution of the organism-environment focuses on the organization of behavioi
be proposed, and general implications and fectively emphasized the environment and dilemma, skill theory attempts to provide is primarily a structural theory, althoug
limitations of the theory will be discussed. neglected the organism: Concepts such as a precise answer- or at least a framework is in no way incompatible with functio
reinforcement, punishment, practice, and that will allow the pursuit of a precise analyses (see Catania, 1973, 1978; Fisct
Both Organism and Environment imitation are used to explain behavior and answer-to five interrelated questions. On 1972; Piaget , 1968/1970).
development (Bandura & Walters, 1963; first reading, several of the five questions Here is a brief overview: Skills deve
Most psychologists agree that psycho- Reese & Lipsitt, 1970; Skinner, 1938, 1969). may seem similar, but as the theory is pre- step by step through a series of 10 h
logical theories, to be adequate, must Useful as these concepts are, they require sented, the distinctiveness of the questions archical levels divided into three tiers. 7
reckon with both organism and environ- important modifications to deal adequately should become clear. (a) What is the struc- tiers specify skills of vastly different tyF
ment (e.g., Aebli, 1978, Note 2; Endler & with organism and environment (Catania, ture of an individual’s cognitive skills at any sensory- motor skills, representatio
Magnuson, 1976; Greenfield, 1976). The 1973, 1978; Herrnstein, 1977; Premack, one point in development? (b) Which skills skills, and abstract skills. The levels spec
interaction of organism and environment 1965). develop into which new skills as the child skills of gradually increasing complexi
is even more obvious in development than To take advantage of the insights of such moves step by step from infancy to adult- with a skill at one level built directly
in most other areas of psychology. Even diverse positions as Piaget’s genetic epis- hood? (c) What is the process by which pres- skills from the preceding level. Each le
the maturation of the child results from a temology and Skinner’s behaviorism, one ent skills develop into new skills? (d) How do is characterized by a reasonably well
combination of organismic factors (in- must somehow put organism and environ- present skills relate to the skills that they fined type of structure that indicates
cluding genes) and environmental factors. ment together in the working constructs of have developed from? For example, are the kinds of behaviors that a person (child
For example, myelination of nerve fibers a theory. The present theory is based on previous skills included in the present skills, adult) can control at that level. The sk
in the cortex is controlled not only by genes the concept of skill, which itself connotes supplanted by the present skills, o r what? at each level are constructed by a per:
but also by environmental stimulation a transaction (Sameroff, 1975) of organism (e) Why is cognitive development so often acting on the environment. She perfor
(Fischer & Lazerson, in press; Peiper. and environment. The skills in the theory uneven in different domains? The attempts several actions induced by a specific
1963). G. Gottlieb (1976) reports that spe- are always defined jointly by organism and to answer these questions are anchored to vironmental circumstance, and the P
cific experiences are necessary for many environment. Consequently, the skills are specific cognitive skills investigated in the those actions occur in that circumstar
aspects of normal physical and behavioral characterized by structures that have prop- developmental research literature. provokes her to combine the actions: ’I
development even when the infant is still in erties like those described by organism- Underlying these five questions are two person thus combines and differentia
the womb, and Cornell and Gottfried (1976) oriented psychologists and that simultane- central issues that operationally form the skills from one level to form skills at
 “-

480 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT

next higher level. The movement from one mental sequences has frequently found that posed to the social, emotional, or linguistic commonly classified as motor, percept1
level fo tfie next occurs in many micro- certain “IogicaT” sequences do not acfuaIIy worlds) or knowledge as measured by stand- or mentat. For example, an infant not o
developmental steps specified by a series of occur (e.g., Hooper, Sipple, Goldman, & ard Piagetian tasks. But there is confusion grasps a doll o r shakes a rattle or kick
transformation rules. Notice that the skills Swinton, 1979; Kofsky, 1966). and controversy about how the concept of blanket but also watches the doll, listen!
develop through levels, not stages: De- The reciprocal give and take between cognition should be used (Chandler, 1977; the rattle, and feels the blanket. Accord
velopment is relatively continuous and theory and data is, in my opinion, essential Flavell, 1977; Kessen, 1966). to skill theory, the higher-level cognition:
gradual, and the person is never at the same for theoretical progress in cognitive-develop- In skill theory, cognition refers to the pro- childhood and adulthood derive direc
level for all skills. The development of skills mental psychology (Feldman & Toulmin. cess by which the organism exercises from these sensory-motor actions: Rep
must be induced by the environment, and 1975; Furby, 1972; Hanson, 1961). The most operant control (Catania, 1978; Skinner, sentations are literally built from sensol
only the skills induced most consistently important test of the levels and of all other 1938, 1969) over sources of variation in its motor actions.
will typically be at the highest level that the predictions from skill theory is empirical. own behavior. More specifically, the person The definition of action in skill theory
individual is capable of. Unevenness in de- The theory must also be internally consistent, can modulate or govern sources of variation however, different from Piaget’s use of
velopment is therefore the rule, not the ex- but internal consistency will be for naught if in what he or she does or thinks. These term. First of all, within Piaget’s frar
ception. The level of skills that are strongly the theory cannot describe, predict, and sources of variation are denoted in the work, the sense in which cognitions beyc
induced by the environment is limited, how- explain the development of actual cognitive theory by sets: sensory-motor sets, repre- infancy are themselves actions (not mer
ever, by the highest level of which the per- skills. sentational sets, and abstract sets. As cog- derivative from actions) is not clear: w1-
son is capable. As the individual develops, In this article, I do not attempt to provide nitive development progresses, infants a child represents a leaf fluttering in ‘
this highest level increases, and so she can a comprehensive review of the large body first control variations in their own sensory- breeze, falling to the ground, having a grt
be induced to extend these skills to the new, of relevant research. Instead, the primary motor actions, then children control varia- color, and turning red in the fall, in w.
higher level. goal is to make the concepts of skill theory tions in their own representations, and sense is the child acting? According to SI
as clear as possible and to show how these finally adolescents o r adults control varia- theory, the child is controlling representatio
Relation Between the Theory and Its Data concepts can be tied to behavior. Concrete tions in their own abstractions. Representa- sets for leaves’ fluttering, falling to i
examples of specific skills are used to illus- tions subsume sensory- motor actions, ground, being green, and turning red. T
The formulation of Levels 1 to 7 is based trate most concepts. To demonstrate how and abstractions subsume representations. control of variations can be conceived as
on the large empirical literature on cognitive the concepts relate more broadly to the re- According to this conception, cognition action on the part of the child, in that i
development between birth and adoles- search literature, a few instances of research includes anything that involves the person’s child actively controls the variations C I
cence. Both the specific structures of the relating to each concept are cited. These controlling sources of variation, even when nitively. Also, all representational sets ;
levels and the numbers of levels were in- examples have been chosen to represent a these sources have conventionally been literally composed of sensory-motor I

ferred from these data. To the best of my wide variety of behaviors, including re- called emotions, social skills, language, or tions, as I will illustrate later.
judgment, a larger number of levels did not search from many different laboratories. I whatever. All these various domains share Second, an action involves a set (rat1
seem to be warranted by the data, and a also indicate which concepts or predictions the same processes of developing more and than merely a point) because it must alwz
smaller number did not seem sufficient to do not yet have good research documenta- more effective cognitive control. be applied to something, and in being ;
explain the data. The validity of this judg- tion. plied, it must always be adapted to tl
ment will, of course, be determined by Nature of Skills thing. Every time an infant grasps a rat
future research. or every time an adult recognizes a famil
The basis for prediction of developmental Assumptions and Definitions Skill theory assumes that cognitive skills face, the action is adapted to the speci
sequences, like the sequence of seven Skill theory is based on a number of can be described effectively and precisely thing acted upon. Thus, every time an acti
levels, has been at issue in the cognitive- specific assumptions and concepts. This in terms of elementary intuitive set theory is carried out, even on the same thing, if
developmental literature. A number of de- discussion of them is not exhaustive but (see Suppes, 1957). The general definition of done a little differently. Notice that ea
velopmental psychologists have argued that focuses on ideas that need to be especially a set is a collection of things. Why is it specific realization of an action always
developmental sequences can be predicted clear at the outset. The assumptions and necessary to talk about collections to ex- cludes both a subject and an object-
on a purely logical basis, where the term concepts divide roughly into three topics: plain cognitive development? When people organism and an environment. An action
logical seems to mean internally consistent the concept of cognitive control, the nature control sources of variation in what they do therefore a set of similar behaviors
(e.g., Brainerd, 1978; Kaplan, 1967; Kohl- of skills, and the characteristics of the levels or think, each such source is a collection or things, but not just any such set: In an actic
berg, 1969). According to this way of think- and transformation rules. set, since it is a class of variations. This the person can control the relevant var
ing, if a coherent, “logical” argument can quality of cognition can be made more con- tions in the behaviors on things. An infi
be made for a predicted developmental se- crete by discussion of how cognition is who can consistently grasp a rattle has a !
quence, that sequence must occur. Al- Concept of Cognitive Control based in action. for grasping that rattle. An adult who c
though the sequence of cognitive levels Cognition and action. All cognition repeatedly recognize a specific familiar fa
predicted by skill theory is internally con- Cognition is a complicated concept. In starts with action, in a very broad sense. has a set for recognizing that face. The thi
sistent, I do not believe that this consis- much of the developmental literature, the Piaget (1936/1952; Piaget & Inhelder, 1966/ is always included with the behavior in t
tency itself provides an adequate test of the term cognition is used to refer to skills of a 1969) has pointed out that cogni:ion is es- definition of a set. In many ways, this de
sequence (Fischer & Bullock, in press). Also, particular type of content-typically sentially what the organism, from its own nition of action is closer to the behavio
research that has explicitly tested for develop- knowledge of the physical world (as op- point of view, can do,whether the doing is concepts of operant and skill than to Piage
 --

482 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT

conception. Indeed, the term behavior in documenting new instances, especially both animals and people. The effects are so Horizontal
class might be superior to the term set, when the unevenness can be attributed to powerful that a number of analyses of
but class is commonly used in psychology environmental causes. Unevenness has human abilities have been developed that
to refer to a type of concept, and set has no been found so often and synchrony so deal primarily with the influences of task on
such surplus meaning. seldom that many developmental psycholo- performance (e.g., Fleishman, 1975; Horn,
Skills, schemes, and operants. Set and gists have begun to suggest that unevenness 1976). In addition, specific experience with a
-Segment
Vertical
action are clearly synonyms within the may well be the rule in development, and task has repeatedly been shown to be im-
theory. How do they relate to skills? A synchrony the exception (e.g., Carey, 1973: portant. These two factors, task differences
skill is a unit of behavior composed of one Cole & Bruner, 1971; Feldman & Toulmin, and experience, likewise account for many
or more sets. The characteristic structure 1975). Unevenness has been demonstrated instances of unevenness. For example, the
for each level is a type of skill, varying in repeatedly for every Piagetian period of type of task, the materials used in a task,
complexity from a single set at Level 1 to a development.* and simple changes in the format of a task
very large number of sets at the highest The concept of skill, in contrast to have all repeatedly produced unevenness Figure I . The spring-and-cord gadget.
levels. What makes a group of sets into a Piaget’s scheme, requires that unevenness (e.g., Barratt, 1975; Jackson et al., 1978:
skill is the person’s control over both each be pervasive in development, because skills Kopp, O’Conner, & Finger, 1975). Even
individual set and the relations between the are defined in terms of the environment as simple practice with a task affects stage of through the seven levels must occur withir
sets. For example, an infant who can shake a well as the organism. Changes in the en- performance (e.g., Jackson et al., 1978; skill domain, not across skill domains.
rattle in order to listen to it has a skill com- vironmental context of action produce Wohlwill & Lowe, 1962). other words, the development of cogniti
posed of two related sets, shaking the rattle changes in a skill. In this regard, skills The usefulness of the concept of operant skills occurs in much the way that beh2
and listening to the noise it makes. share important similarities with Skinnerian does not extend, however, to analysis of iorally oriented psychologists have st
The relation between the concept of skill operants (Skinner, 1938). The term operanr the organization of behavior. Although gested (Baron, 1973; GagnC, 1968, 19;
in the theory and the concepts of scheme refers to a behavior that is emitted by an reinforcement and punishment can be useful Schaeffer, 1975). The child masters speci
and operant from Piaget and Skinner may organism, not elicited by a stimulus. At the experimental operations for analyzing or- skills, builds other specific skills upon the]
help to clarify the meaning of skill. Piaget’s same time, the specific form of the behavior ganization, they are insufficient. What the and transfers skills from one domain to a
general word for cognitive structure is and the probability that the organism will concept of operant lacks in behaviorist other. This mastery process involves qua
scheme’- a structure for knowing, a pro- emit the behavior are affected by environ- theory is a system for analyzing the or- tative changes in skills, but the speci
cedure that the child actively applies to mental stimuli. The behavior is therefore ganization of operants and how that or- changes occur gradually, not abruptly.
things in order to understand them. In broad controlled by both the individual organism ganization changes with learning and de- Induction of a new skill. An example a
conception, there are many similarities be- and environmental stimuli. Hunt (1969) and velopment. Skill theory is designed to show how development is induced jointly
tween scheme and skill, as already indicated Aebli (1978, Note 2) have pointed out that provide such a system. both the person‘s skills and the envira
in the discussion of action, but there are also most of the behaviors studied by Piaget and In general, then, scheme and operant are mental circumstances in a particular situatic
major differences. One of the most im- his colleagues are in fact operants. synonyms for skill within the present The developmerit of conservation of leng
portant differences involves the organism- The phenomena of developmental un- theory, although of course they have dif- in the gadget shown in Figure 1 (adapt
environment problem: Piaget’s schemes evenness make good sense from a be- ferent psychological frameworks. The from Piaget et al., 1968, chap. 4) provide:
allot much less importance to the environ- haviorist perspective. Behavioral research levels of cognition are a hierarchy of skills, simple illustration of this joint induction 1
ment than the skills of the present theory has shown repeatedly that task factors have schemes, or operants in which each higher- organism and environment. In the gadget
do. Schemes are assumed to have a high potent effects on most kinds of behavior in level skill, scheme, or operant is actually cord is attached to a spring and drap
degree of generality, encapsulated in composed of lower-level skills, schemes, over a nail, so that the cord is divided in
Piaget’s concept of the structure d’ensemble or operants. The theory thus provides a tool two segments by the nail, a horizontal se
(Piaget, 1957, 1968/1970: Inhelder & Piaget, Many of the English translations of Piaget‘s works for analyzing skills, schemes, or operants ment and a vertical segment. Differi
1955/1958). This powerful generality of use the word schema instead of scheme to translate into units of widely varying complexity. weights attached to the cord will produ
the French sch?me. There is a problem with this The definition of sets has an important changes in the length of the horizontal se
schemes should produce a high degree of usage: In recent years Piaget has differentiatedschime
synchrony in development. Two tasks that from schema (Furth, 1969; Piaget & Inhelder, 19661 implication for the meaning of skill, scheme, ment of the cord and concomitant chang
according to Piagetian analysis require the 1971). SchCma refers to an internal image of some- and operant. Because an action always in- in the length of the vertical segment.
same scheme should develop at the same thing, which is very far from the meaning of schime. volves a particular object o r thing, a skill Consider a 5- or 6-year-old girl who :
time. Yet rather than synchrony, re- * Here are just a few of the relevant references: for must be specific to particular objects or ready has two skills (or schemes o r ope
the sensory-motor period, Butterworth, 1976: Jack- things. This implication is equivalent to
searchers typically find unevenness in de- son, Campos, & Fischer, 1978; Kopp, O’Connor,
ants) for the length of the cord: (a) SI
velopment (e.g., Flavell, 1971b; Jamison, & Finger, 1975; Uzgiris & Hunt, 1975; for the pre- saying that as children develop, they master understands approximately how the leng
1977; Liben, 1977; Toussaint, 1974). operational period, Gelman, 1978; Goldstein & specific cognitive skills; they do not develop of the vertical segment relates to the leng
The number of well documented in- Wicklund, 1973; Watson & Fischer, 1980; for the con- uniformly across the entire range of skills. of the horizontal segment; that is, she c;
crete operational period, Achenbach & Weisz, 1975; Similarly, since cognitive development roughly control the relation between tl
stances of unevenness has been increasing Hooper et al., 1971; Jackson, 1965; Smedslund, 1964:
astronomically in recent years; American and for the formal operational period, Martarano, proceeds by the coordination of specific vertical length and the horizontal lengt
psychologists seem to take special delight 1977: Neimark, 1975; Piaget, 1972; Wason, 1977. skills or schemes or operants, development using the vertical to predict the horizonG

~ ~~~
484 KURT W. FISCHER THEORY O F COGNITIVE DEVELOPMENT ‘
(b) She also understands approximately sential contribution of the environment to is clearly central to using the theory. The BefQrethe lev& tl%ems&es&an be
how the horizontaI Iength retates to the skiII deveIopment requires, of course, that central question for task analysis is: What scribed, a number of key concepts must
vertical length; that is, she can use the unevenness be the rule in development, but sources of variation must the person control introduced.
horizontal to roughly predict the vertical. it by no means excludes instances of syn- to perform a task? That is, what sets must Relations between skills and levc
But she does not yet understand that the chrony. Developmental synchrony in she or he control, and what relations be- Contrary to the use of stage or period in m
changes in the horizontal length compen- various degrees is predicted by the theory. tween sets? Guidelines for task analysis cognitive-developmental models, the lev
sate for the changes in the vertical length, Analysis of skill structures plus control of will be described later after the theory are used generally to characterize a chil
so that the total length of the cord does environmental factors such as practice and has been more fully elaborated. skills, not the child in general. A child 1
not change; she does not yet understand familiarity allow the prediction of special Closely related to the problem of specify- normally be at different levels for diffen
conservation of the length of the cord. instances of near-perfect synchrony, as well ing which sets and relations a person must skills. To characterize a specific child
To construct an understanding of this con- as predictions of various degrees of syn- control in a task is the problem of defining cognitive profile is required, indicating le
servation, she must coordinate her two chrony under differing circumstances. Such the boundaries of a set. Indeed, the most of performance on a wide range of sk
skills for predicting the length (vertical predictions will be illustrated later. useful form of set theory may prove to be (see, for example, Rest, 1976).
predicts horizontal, and horizontal predicts Because of the connotation of the word the theory offuzzy sets (Negoita & Ralescu, There is, however, one sense in which I
vertical). This combination will occur only skill, the phrase skill domain implies a fairly 1975), which does not require precise defini- levels are used to characterize the chi
if (a) the child has the two skills and (b) broad grouping of behaviors. However, tions of set boundaries. The problem of Each child has an optimal level, indicati
she plays with a gadget in which length in methods for determining the developing defining the boundaries of a set is virtually the best performance the child shows, whi
fact conserves. As she applies the two skills child’s groupings of behaviors into skill identical to the problem encountered by is presumably a reflection of both pract
repeatedly to the gadget, the task itself domains are crude at best (see Beilin, 1971; behaviorists in defining the boundary of an and the upper limit of his or her process1
induces the child to notice a connection be- Flavell, 1971b, 1972; Wohlwill, 1973). So as operant (Schick, 1971). The problem may be ability. Just as in information-processi
tween them, because the properties of the not to beg the question of which skills more serious theoretically than practically theories, this central processing limit
task make the two skills closely related. develop together in a single domain, I will (Catania, 1978), but it is still a problem. creases with development (Case, 1 9
Then the child explores the connection and distinguish task domain from skill domain. Skill theory at least points in the direction
Flavell & Wohlwill, 1969; Halford & $3
gradually constructs a new, higher-level A task domain is a set of behaviors that of a solution by specifying a universe of son, 1980; Pascual-Leone, 1970; Scandu
skill for conservation of the total length involve only minor variations in the same possible skill structures and thus providing a 1973). But skill theory does not require t
of the cord. task, in contrast to the broad grouping of tool for partially defining behavioral units. homogeneity of performance demanded
The importance of the contribution from behaviors across tasks in a skill domain. Development is analyzed into a hierarchy of many information-processing theories, sin
the gadget (the environment) should not be Within a task domain, there is virtually no operants-skill levels of increasingly com- the optimal level is merely an upper lim
underestimated. If the cord were not a cord problem in determining which behaviors plex cognitive control-plus various transi- not a characteristic of all cognitive behavi
but a rubber band, conservation of the belong to that domain. As will be shown tional forms specified by the transformation at a given point in development. Also, t
length would not obtain, because the differ- later, the theory can predict developmental rules. A particular behavior can be related tolimit is characterized by a skill structt.
ent weights that stretch the spring would sequences within a task domain. It may also one of the possible skill structures, and at (one of the cognitive levels) rather than
also stretch the rubber band. More gen- prove useful in determining the nature and the minimum, the theory will then imply simple whole number of items in worki
erally, the child’s possession of two skills scope of skill domains, but that usefulness particular kinds of changes in the cores and memory.
cannot by itself produce coordination of remains to be demonstrated. boundaries of sets across transformations The postulation of levels instead of cc
those skills. The child must be induced to Task analysis. Because task factors are and levels. tinuous monotonic increases in complexi
coordinate them by applying them to some- so important in skill theory, task analysis has implications for the form of the i
thing for which they do ~ o o r d i n a t e . ~ Concepts for Defining Levels and crease in optimal level with age: Associat
The joint action of organism and environ- Transformations with the levels, there will be spurts in t
ment in cognitive development is equally This analysis differs from that of Piaget et ai. Through the joint contributions of the speed of developmental change. That is, a:
(1968) in three major ways: (a) They do not grant the
important for all the skill levels in the same inductive role to the task. (b) They do not person and the environment, skills, schemes, child moves into a new level, she will shc
theory. A 1- or 2-month-old infant, for ex- ascribe to the 5-year-old the ability to relate vertical or operants develop through at least seven rapid change; but once the level has be,
ample, will typically not be able to control to horizontal and vice versa, although they do de- hierarchical levels. The skills at each level attained, she will show slower change.
the relation between shaking a mobile and scribe an ability to relate weight to length of the are characterized by a structure that this way, the speed of development will va
watching it jiggle. But when she has mas- spring (which is also consistent with skill theory). indicates the kinds of behaviors that the cyclically with the skill levels. Note that tk
These several abilities are both predicted by skill
tered the two individual skills, shaking the theory and supported by some research (Wilkering. person can control at that level. Also, at hypothesis does not mean that develo
mobile and watching it, she will be induced 1979). (c) They do not explain conservation as arising each level, the skills include all the lower mental change is abrupt or discontinuou
by the mobile to coordinate the shaking from the coordination of skills relating vertical and levels. For example, when a child is at Level The child moves into a new level gradual
and the watching. It is a property of mobiles horizontal. instead, they describe three stages: relating 5 for a specific skill. that skill subsumes over a long period, but the speed of chan,
weight to spring, then understanding conservation,
-and of many other things in the world- and finally understanding the proportional relation skills at Levels 4 through 1. Note, however, during this period is relatively rapid.
that shaking them produces interesting between weight and amount of displacement. Their that these lower-level skills become more Although I have defined the optimal lev
changes in their appearance. third stage develops much later than what is dis- differentiated at each higher level to which as a single upper limit, there is a possibili
Developmental synchronies. This es- cussed in the present example. the superordinate skill develops. suggested by ability research that at tl
 THEORY O F COGNITIVE DEVELOPMENT
486 KURT W FISCHER

Table 1 clude much more complexity and detail than sets; phrases such as variations in length level. Uppercase letters designate sets, w
The Cycle of Four Levels That Repeats a mapping4 or the d x t o r r d e will be used as skort- diffefeitt t y w a e e s specifying the tier o f t
in Each Tier A third type of structure, called a system hand in place of longer descriptions such as set, as shown in Table 2.
of systems, is a relation between two sys- the child’s representation of variations in Superscripts and subscripts on a capi
Characteristic Set-theory the seen lengths of the cord or the child’s letter give additional information about a sf
Level structure description tems, as shown for Level IV in Table 1. The
psychological interpretation of a system of representation of variations in what she can Lines and arrows indicate relations betwe
I Single set Wl or [XI systems is that people can relate two sys- mahe a doctor doll d o in examining a patient sets, and letters above or next to a line
tems in a single skill, which allows them to doll. arrow indicate a particular relation. Brac
I1 Mapping [W -XI Transformation rules. The five major
form a new kind of set: the most elementary ets designate a skill, and certain math
111 System W.4.B Xml set M at the next higher tier. In this new transformation rules specify how a skill can matical symbols and abbreviations speci
1v set, each system is one element, so that the be transformed in development. Several the application of transformation rules.
simplest set has just two elements. rules deal with the ways that skills can be
Note that in all these structures, a set is a combined to produce more complex skills Recurring Cycle of Four Levels
source of variations that the person can and how they change as a result of the com-
binations. The other rules indicate altera- The progression of skills through t
control- variations in actions, representa- hierarchical levels shows a repetitive cyc
highest levels, a person may have a few dif- tions, or abstractions. In each case, the tions in skills that are less drastic but that
nevertheless produce clear-cut develop- diagramed in Tables 1 and 3. This kind
ferent optimal levels in different broad do- variations involve behaviors-on-things, but repetition of structure has been discussed I
mains. For example, an adult’s optimal level the level of complexity of the organization mental orderings of skills. Although the
rules specify qualitative changes in skills, both Piaget (1937/1954, 1967/1971) a
in spatial skills may be different from his or of those behaviors increases markedly at the Werner (1948), although neither of them h
her optimal level in verbal skills (see higher levels. Consequently, I will at times these changes occurgradually, not abruptly.
The transformation rules are central to described the exact nature of the propos
Horn, 1976). use simplified descriptions of higher-level parallels. The structures of Levels 1 to
Mappings and systems. The concepts the theory, for they allow much more de-
tailed predictions of sequence and syn- are parallel to the structures of Levels 4
of mapping and system define the possible 7 and 7 to 10, but at each cycle the stru
relations between sets within a skill, and chrony than the cognitive levels alone. The
The concepts of mapping and system are both
levels produce only macrodevelopmental tures are composed of a different type
both of these concepts can be described in derived in part from Piaget’s and Wemer’s work. set, as illustrated in Table 3.
set-theory terms. A mapping is a structure Piaget and his colleagues (Piaget et al., 1968; see also predictions (across levels), but the trans-
Flavell, 1977) have analyzed several behaviors of the formation rules also provide microdevelop- Each cycle of four levels is a tier and
relating two sets: a collection of ordered named for its type of set. For the first tie
preschool child in terms of what they call a function, mental predictions (within a level). By the
pairs in which the first member in each pair which is similar to a mapping. But they seem to restrict Levels 1 to 4, the sets are sensory-motc
is from one set (W) and the second member microdevelopmental transformations, more
their analysis to only a limited group of behaviors they are actions and perceptions of the chi
is from another set (X). The first set is said and analyze those behaviors in terms of the degree to complex skills can be constructed than the
ones shown in Tables 1 and 3, which are the on things or events in the world. Within tf
to be mapped onto the second: [W - XI. which the behaviors match the characteristics of a
tier, the combinations of sensory-mot
A system is composed of a relation be- mathematical function. Their conclusion is therefore simplest possible at each level. Adequate
that preschool children can sometimes show quasi formal definitions of the transformation sets grow more and more complex as tl
tween two subdivided sets. Each set is functions (called constituent functions) but not real child develops through the first four level
divided into two subsets, which are related functions (called constituted functions). The skill- rules depend on the formal descriptions of
the levels, and so the rules will be defined until at Level 4 the combinations create se
to the two subsets in the other set. The two theory concept of mapping may be viewed as a re-
definition, generalization, and extension of Piaget‘s precisely later. of a new kind, representational sets.
subdivided sets are said to form a system, These representational sets designa
concept of function. The concept of system is not
with the subsets noted by subscripts: directly present in Piaget’s work, but it derives in part concrete characteristics of specific object
[wA,B XA,B]. The double-headed arrow Notation
from Piaget’s concept of concrete operations (Inhelder events, or people (including the child he
indicates that the structure is a system even & Piaget, 195911964; Piaget, 1942, 1949). One of the The introduction of a notation system self). Note that the new sets subsume se
when the subsets are not expressly listed in most central aspects of a concrete operation is that it
is reversible, but research does not support Piaget‘s will allow semiformal description of both sory-motor sets, as shown in Table 3; tl
the formula: [W ++ XI. argument that reversibility is absent in the pre-opera- the characteristic structures for the levels sets from the earlier tier do not disappea
The psychological interpretation of map- tional period and first emerges in the concrete opera- and the transformation rules. It will thus For Levels 4 to 7, the representation
pings and systems is straightforward. In a tional period (Moore & Hams, 1978; Schmidt & facilitate use of the theory as a tool for tier, the new sets are again combined
mapping, a person can relate two sets in a Paris, 1978; Fischer & Roberts, Note 3). The con- analyzing development. The notation sys-
cept of system in skill theory is intended to explain the more and more complex ways, producing
single skill-two sensory-motor actions, behaviors that have been documented by Piaget‘s tem and the structural descriptions are not cycle parallel to that for Levels 1 to 4.
two representations, or two abstractions. In research on concrete operations but without re- rigorously formal; they are only as elaborate At Level 7, the combinations of repr
a system, a person can relate two subsets of quiring that reversibility be absent from mappings. as is necessary to convey the intended sentational sets create new sets of anothj
each of two sets in a single skill-two Both mapping and system also incorporate explicitly meanings. kind: abstract sets, which are genera
components of two actions, representa- Werner’s hypothesis that development proceeds by
simultaneous differentiation and hierarchical integra- The notation rules are described in Table intangible attributes of broad categories (
tions, or abstractions. The ability to deal tion (coordination). The meanings of mapping and 2. Numbers and plain capital letters J, K , objects, events, or people. These new se
with two subsets in each set means that the system are sufficiently different from Piaget’s usage and L designate skill levels. Lowercase subsume the representational and sensory
person can control two sources of variation that attempts to plug Piaget’s usage into skill theory italic letters indicate skills of unspecified motor sets from earlier tiers, as shown I
in each set. As a result, a system can in- will lead to serious errors.
 _ ~ -_ ~

',
,
\

488 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT ,\

**
Table 3. What happens after Level 7 is
,trecausetmsretras
been so little research on cognitive develop-
When a mathematician says that one set is
mapped onto anofher, he or she means,
roughly speaking, that variations in the first
Table 2 (continued)
Type of symbol Examples Meant.
ment in adulthood. Yet the predictions of
the theory are clear and direct: The abstract
set produce predictable variations in the
second one. In an analogous way, at Level
Subscript to the left or right Information of intei
the sets; used to ~ c r i r
_.
sets should produce an abstract tier-an- I1 a child can understand situations where inate related sets; two
other progression through the cycle of four he or she can relate one set of variations to letter subscripts indicate
the set is composed of
levels. When the combinations of abstract a second set of variations. two subsets
sets reach Level 10, they should produce Level 111 is characterized by a system
still another new kind of set. Specification -arelation between two sets each of which Brackets IA *Fl Sets and relations inside
brackets constitute a sing
of the nature of the new sets at Level 10 is divided into two subsets, indicated by the skill
must await future research on cognitive two-headed horizontal arrow between sets
development in adults. in Tables 1 and 3. The child is no longer Long line connecting two letters Mapping: Relation betweer
two sets
To distinguish the general cycle of levels limited to the two simple sets in the map-
from the specific levels, the Roman nu- pings of the previous level but can control Horizontal two-directional arrow System: Relation between
merals I to IV will be used to refer to the relations between two subsets for each set. two sets, each composed
of two subsets
levels of the cycle, and the Arabic numerals That is, the child can understand situations
1 to 10 will be used to refer to the actual where he o r she can systematically relate Vertical two-directional arrow A*F System of systems: Relatio
behaviorally defined cognitive levels. two components of one set of variations to between two systems
[ R E A
As shown in Table 1, Level I is character- two components of a second set of varia-
ized by single sets-single sources of vari- tions. In this way the child can deal with b m
ation that the child can control by them- one subset while still keeping the other in Lowercase letters above or next [A -F ] . [ A - F1 A particular relation
to line or arrow
selves but not in relation to each other. That mind and as a result can control much finer
is, the child cannot yet coordinate sets into a covariations in the two sets than at Level 11. Multiplication sign Intercoordination of two
skills
higher-level skill. The characteristic structure for Level IV
The characteristic structure for Level I1 is a system of systems- a relation between Addition sign [A * F ] + [F *R ] Compounding of two skills
is a mapping- a relation between two sets, two systems, indicated by the two-headed Sign for greater than [ A * F ] > [f * R ] Change in focus from the fir
indicated by the long line in Tables 1 and 3. vertical arrow in Tables 1 and 3. At this skill to the second
Equals sign [A * F ] + IF ++ RJ= [A F ++ R] The skill on the right is the
result of the transformatic
indicated on the left
Table 2 Foc A change in focus between
Notation Rules the two skills on the left
produces the skill on the
Type of symbol Examples Meaning right
Roman numerals I, 11, 111, IV Cycle of levels that repeats Sub Substitution of a set in the
in each tier skill on the left produces
the skill on the right
Arabic numerals I , 2, .. . 5, 10 The ten hierarchical levels
Diff A = AM,A, Differentiation of the set or
Plain capital letters J, K, and L L,L+l Undesignated level skill on the left into the set:
Diff [A ++ F ] = [AG,,,++ Fx,x,y]
Lowercase italic letters b, s, e Skills of unspecified level . or skills on the right

Boldface capital letters M,G, P Sensory-motor sets

Italic capital letters A, F. R Representational sets
Script capital letters 9, x Abstract sets level, a person can control the relation be- cess by which Level IV of one tier bc
Plain capital letters w, x,c Sets of undesignated tier tween two systems, keeping in mind one comes Level I of the next tier. Level I ca
G.4 '
p system while dealing with the other. This co- be thought of as a simple building bloc1
Superscript to the right Sets designated by the main
letter are components of ordination of two systems produces a new Level I1 is then a combination of thos
the set designated by the kind of set, the most elementary set M at the building blocks in one dimension to fori
superscript, which is at the next tier, as shown in Table 1. lines. At Level 111, lines are combined t
next higher tier The metaphor drawn in Figure 2 illus- make two-dimensional objects, such as th
Superscript to the left LA,zP Level of set trates the cycle of four levels and the pro- square in the figure. Finally, at Level I\
 490 KURT W. FISCHER THEORY O F COGNITIVE DEVELOPMENT

representational, and abstract tiers will help and what they make happen. They do not gadget, one set, IS, involves the infar
LEVEL
to ctarify the general picture o€ cogdive Lul€kr&dthat objects, events, a d people krokkg at tke gadget tk.keR it CFOSSeS
development presented by the theory. The have their own characteristics independent field of vision and maintaining it in her sig
I 0
child s potential skills with the spring-and- of what the infants themselves do; that Another set, ‘G, involves her grasping
cord gadget in Figure 1 will be traced ability awaits the development of repre- spring when it touches her hand and ms
through the levels as a continuing example. sentational sets at Level 4. Consequently, taining her grasp on it. Most of Piagr
a child does not realize, for example, that (1936/1952) primary circular reactions se
Sensory-Motor Tier: Levels 1 to 4 her favorite rattle has properties like hard- to be Level 1 single sensory-motor actio
ness and the capacity to make noise that The infant can control many such single s
The first four levels constitute the sensory- are independent of her own actions on it. at Level 1, but she cannot control the rt
motor tier, as shown in Table 3. In this tier, Nor does she understand that people and tions between sets.
all skills are composed of sensory-motor many other things can act by themselves Single sensory-motor sets are not limi
sets-actions (including perceptions) on independently of her actions. To emphasize to adult-defined modalities in percept
objects, events, or people in the world. the domination of this world by action and and action. The young infant does not knc
Skills at this tier have most of the charac- to avoid confusion from terms like object for example, that seeing is different fr
teristics that have been called “sensory- or person, I will refer to objects, people, listening. When she is attempting to look
motor” by a long and distinguished line of and other things in the infant’s experience the doll swinging in front of her, any soul
psychologists (e.g., Baldwin, 1925; Dewey, as tab1eaus.j For example, an infant grasps it makes can be incorporated into her
1896; Hobhouse, 1915; Lashley, 1950: a tableau, not an object, and listens to a ID. So long as she does not have to relate
Piaget, 1936/1952; Werner, 1948): Both tableau, not an object. sights and sounds independently, sight i
Figure 2. A metaphor for the cycle of four levels. sensory and motor components are integral Several independent investigators have sound can be mixed together in the sa
parts of the skills and for most purposes recently reported data that generally sup- Level 1 skill. This lack of differentiation
planes are combined to form three dimen- cannot be genuinely separated. Because the port the pattern of developmental changes Level 1 contrasts with Piaget’s (1936/19
sional objects, such as a cube-a new type infants can control only sensory-motor predicted by the four sensory-motor levels argument that young infants have diff
of building block. In this way the cycle actions, their skills are purely practical: (Emde, Gaensbauer, & Harmon, 1976; entiated schemes for seeing, hearing, gra
begins over again, with Level IV of one tier They understand how to act on specific Kagan, 1979; McCall, Eichorn, & Hogarty, ing, and so forth, which must be coordina
serving as Level I of the next tier. things in the world but cannot think about 1977; Uzgiris, 1976). McCall’s analyses are together. The undifferentiated and unf
An elaboration of how this cycle of Levels those things independently of acting on especially relevant: In examining patterns of ordinated status of Level 1 skills in
I to IV applies in the sensory-motor, them. They understand what they can do correlations among items in infant tests, present theory fits Wemer’s (1948,19
he found changes in correlation patterns characterization of a developmentally pn
Table 3 that suggested four successive periods of itive state. Many of the studies of classi
S e n s o y - M o t o r and Representational Levels of Skills change and consolidation in the first two and operant conditioning in young infa
years of life-times of instability in cor- may involve such undifferentiated mu
Sensory-motor Representa- Abstract relations followed by times of stability. If modality sets (see Papousek, 1967; Sam
Level Name of structure setsa tional sets setsb infants are in fact developing through Levels off, 1971).
1 Single sensory-motor set [‘A] or [‘B] 1 to 4 in an age-related progression, one Although infants cannot control any re
would expect periods of change and con- tions between sets at Level 1, they
2 Sensory-motor mapping [*A -‘B] solidation in correlations exactly like those readily drift from one set to another, usua
3 Sensory-motor system *
[IAG.H 3 B ~ , d that McCall found. Further research to test led by some tableau. Consequently, e%
* the relation between McCall’s findings and though they cannot yet coordinate two se
4 System of sensory-motor systems, which is a
single representational set [ ‘AR
5:
‘CR *
‘BE

4I)R
] E IW the levels clearly needs to be done.
The characteristic structure of Level 1 is
they do not become stuck on one set
long. Indeed, their drifting from set to
the single sensory-motor set (shown in eventually leads them to explore the relati
5 Representational mapping [‘R -‘TI
the top row of Table 3), a set, ‘D, of acting between two sets and so to intercoordin:
6 Representational system WJ,S* ‘TJ.KI on tableaus, such as looking at a doll. A them into a Level 2 skill, simultaneou:
7 System of representational systems, which is a 12-week-old infant may look for long differentiating them from each other.
single abstract set periods at the tableau produced by a doll The characteristic structure of Level 2
hanging on a string in front of her. Even
when the doll swings back and forth in a
~~

a Sensory-motor sets continue after Level 4, but the formulas become so complex that they have been 5 For the same reasons, Piaget occasionally u:
omitted. To fill them in, simply replace each representational set with the sensory-motor formula for Level 4. wide arc, she can keep her gaze on it. This is
this term in his works on infancy (Raget, 1936/19
Development through the abstract tier shows the same cycle as development through the sensory-motor and a single sensory-motor set o r action, the 1937/1954). He did not, however, use it consistentl)
representational tiers. Abstractions are built from representational and sensory-motor sets in the same way set ‘D of adaptations of looking at the doll these works, and he has not used it in subsequ
that representations are built from sensory-motor sets. tableaus. Similarly with the spring-and-cord works.
 492 KURT W. FiSCHER THEORY OF COGNITIVE DEVELOPMENT L

the sensory-motor mapping, in which one considered a t every level and especially at of his Q W ~actions. In the w8, every vertka€OT kori-zmA segfftefft. Tasks 6-
sensery-zS&x set, 2A, i s mapped a the eadiest k& within a tier, where dif- object is in fact the focus of a number of designed that will help him to separate su
second sensory-motor set, 2B, as shown in ferentiation is always poor. different sensory-motor systems; that is, factors in one situation, but when the factc
Table 3. One type of sensory-motor map- Level 3 is characterized by the sensory- every object can be made to participate in covary in a task, the child will treat thr
ping is a means-end mapping, in which a motor system, in which two components of or produce many different types of actions. in a single representation.
child can use one action in order to bring one sensory-motor set, 3AG,H,are related to The ability to understand objects in this way Many different types of representatiol
about a second action. For example, a 7- two components of a second sensory- (as independent agents of action) first de- sets should develop at Level 4, according
month-old infant looks at a tableau of a doll motor set, 3BG,H,as shown in Table 3. The velops at Level 4 (Watson & Fischer, 1977). the theory; and Piaget (194611951, Obseri
and uses what she sees to guide her attempts most investigated type of sensory-motor tion 64) described a behavior that demc
to grasp the tableau (Field, 1976; Lasky, system is the means-end system. Unlike the Representational Tier: Levels 4 to 7 strates a second type, a set of objects
1977; Ruff, 1976). She has combined two means-end mapping of Level 2, the means- events that all share a single action
simple actions, looking at the doll and grasp- end system allows the infant to control Level 4 is the culmination of the sensory- characteristic. His daughter Jacqueline U S
ing it, into one means-end mapping in complex variations in means and ends motor tier, and so it produces a new type the word bimbam to mean swaying or flutt
which looking is used as a means to bring (Fischer, Note 4). For example, Piaget’s of set aad begins a new tier, the representa- ing. She combined her sensory-mol
about grasping. That is, she has mapped the 10.5-month-old son Laurent drops a piece of tional tier. In terms of the repetitive cycle of system for rocking back and forth on a pie
sensory-motor set ZDof looking at the doll bread, watches it fall, breaks off a crumb levels, the characteristic structure for Level of wood with her system for making a It
onto the sensory-motor set ZHof grasping and drops it, watches it fall, and so forth 4 is the system of sensory-motor systems flutter and used “bimbam” to refer to bo
it, as shown in Table 3. She may also have (Piaget, 1936/1952, Observation 141). He (sensory-motor Level IV), which is the Then she gradually extended this rep
a separate, complementary mapping, in constantly varies the means (the way in same as the single representational set sentational set to a wide range of object
which she maps grasping the doll onto look- which he drops the bread) and watches (representational Level I). This type of skill Other examples of the construction of sin)
ing at it. For example, she grasps the tableau closely the variations in the end (seeing the is a relation between two sensory-motor representational sets from sensory-mol
of the doll and brings it before her eyes so bread fall). systems, as shown in Table 3. The combina- systems have been described by Bertentl
that she can look at it. Similarly, with the At Level 2, he was unable to perform tion of these systems generates the single and Fischer (1980), Watson and Fisck
spring-and-cord gadget, an infant pulls on such a complicated experiment in action; he representational set in which children can (1977), Fischer and Corrigan (in pres
the spring, *G, so that she can watch it could learn little more than that dropping represent simple properties of objects, Fischer and Roberts (Note 3), and Fisck
stretch, 2S. Many of Piaget’s (1936/1952) produced falling. The reason for this limita- events, and people independently of their and Jennings (in press).
secondary circular reactions are means- tion was that he could relate only one aspect own immediate actions. A word of caution may be helpful at ti
end actions of this sort, although a number of dropping the bread to one aspect of seeing With the spring-and-cord gadget, the child point about the meaning of representatic
of the behaviors that he classifies in this the bread fall. can combine Level 3 systems for the gadget The term is often used as a virtual synonj
category seem to be complex forms of Level At Level 3, he can relate two aspects of into a single Level 4 representation. One for recall memory o r for symbol use. But
1 actions. each action, and therefore he can build skills such skill involves the child’s understanding skill theory, representation is different frc
Sensory-motor mappings should include that coordinate and differentiate types of that the spring itself stretches. For example, both of these meanings. It refers to the c
many types of skills besides means-end variations in dropping with types of varia- the following two systems can be co- ordination of two or more sensory-mot
mappings, especially skills involving two tions in falling. Similar kinds of skills can be ordinated at Level 4: When he pulls the systems to form a single representatior
components within the same modality. built with the spring-and-cord gadget-for spring, it stretches; when he sees someone set, not to recall memory or symbolizati,
Bertenthal, Campos, and Haith (in press) example, learning not only that pulling the else pull the spring, it stretches. Therefore, per se. Skills involving both recall memo
describe one such skill: By 7 months, infants string makes it stretch but that pulling it in a characteristic of the spring is that it and symbol use can develop before Level
can apparently relate several visual com- different ways makes it stretch differently. stretches; the child controls a representa- and in addition skills can be constructed
ponents such as angles to form a line (see Examples of such means-end systems tional set 4L for the spring’s stretching. In Level 4 that do not centrally involve eith
Level I1 in Figure 2). Presumably many abound in the research literature (e.g., the same way: he constructs a set *W repre- recall or symbol (Fischer & Corrigan,
more such mapping skills develop within Bryant, 1974, p. 162 ff.; Koslowski & senting that the weight itself can “pull,” press). A single representation is defined 1
modalities such as looking, grasping, and the Bruner, 1973; Fischer & Roberts, Note 3). independently of his feeling it; and he con- its structure, not by its function as reca
like in infants. As with earlier levels, researchers have structs a set *C representing that the cord symbol, o r any other such psychologic
Just as with Level 1, however, Level 2 neglected other types of Level 3 skills, such can be big, independently of his making category.
skills cannot be subdivided according to as those within a modality (see Fischer & it move. The characteristic structure for Level 5
adult conceptions of modalities. If stimuli Comgan, in press). With these single representations, the the representational mapping, in which o
from two different adult-defined sensory Despite all the sensory-motor sophistica- child shows a lack of differentiation anal- representational set, jR, is mapped onto
modalities, for instance, co-occur in such a tion of Level 3, the skius are still definitely ogous to that with Level 1 single sensory- second representational set, ”T, as shoh
way that the infant can treat them as one limited: The infant is only able to control motor actions. In the gadget, he will confuse
source of variation, then at Level 2 the one sensory-motor system at a time, and the pressure exerted by the weights with
The rules of compounding, substitution, c
infant can treat them as a single set that she therefore he cannot yet deal with many of their size, mixing them both together as ferentiation, and intercoordination nicely account
can relate to a second set. The same kind of the complexities of acting on objects, nor “big.” Similarly, he will confuse the total the sequence by which this skill developed, as (
concern about the definition of sets must be can he understand objects independently length of the cord with the length of the scribed by Piaget (194611951).
 494 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT

in Table 3. With this kind of skill, the child combines the vertical and horizontal lengths terizes broad categories of objects, events, lengthandwidth “ p a s & s h s h e t
can r e b e variaiions in one representation to o f t k cord when m e weight is used with the or peopte. (Note that, as with representa- The coordination of these two conc
variations in a second repre~entation.~same length when another weight is used, and tion, abstraction has many different mean- conservation skills produces the abst
Consider a 4- or 5-year-old who is given the thus he knows how the lengths vary together ings in psychology; see Pikas, 1965. These concept of Conservation, which can the]
spring and several weights of different sizes and compensate for each other (Piaget et al., various meanings should not be confused generalized to other tasks. Other instar
from the gadget in Figure 1. If he has had 1968; Verge & Bogartz, 1978). With the with the specific meaning used here.) of Level 7 skills include most analo
sufficient experience with the task, he can gadget, he can also construct several other In a Level 7 skill, the person can control (Lunzer, 1965), political concepts like
roughly use the size of the weight (one set) to Level 6 systems, each involving the relation the relation between two representational and society (Adelson, 1972). and a fec
control the length of the spring (the other of two concrete variables to each other. systems, as indicated by the Level 7 struc- Piaget‘s simpler formal operational t:
set), thus understanding in an approximate Other representational systems that have ture in Table 3. Consider a 15-year-old boy (Inhelder & Piaget, 1955/1958).
way that large weights will make the spring been studied in the research literature in- who can control a system of systems for the Following the recurring cycle, abst
stretch farther than small weights. clude most of Piaget’s concrete opera- sets in the spring-and-cord gadget. He can tions should develop through Levels 7 to
Notice in Table 3 that this structure (like tional tasks (e.g., Inhelder & Piaget, 19591 integrate several of the systems from the For example, with the spring-and-c
all representational structures) can be de- 1964) and a number of other tasks (e.g., previous level into a single Level 7 system gadget, the individual will start with sir
scribed either in terms of representational Watson & Fischer, 1980; Winner, Rosen- that controls the relations among the weight, abstractions such as conservation, t
sets without visible reference to their sen- stiel, & Gardner, 1976). the vertical length of the cord, the hori- relate two such abstractions in a mapp
sory-motor origins or in terms of the Despite all this sophistication, however, zontal length of the cord, and the length and so forth. Because so little resea
sensory-motor sets on which the repre- the skills of Level 6 are still definitely of the spring. When he is thinking, for has been done on cognitive developrr
sentational sets are based. For example, limited. The child can only deal with one example, about how the changes in weight beyond adolescence, however, no data
the child’s understanding of the relation Level 6 system at a time. He cannot relate produce changes in the length of the spring, available to provide a strong test of s
between weight and spring ties directly to various systems to one another. Even if he he can simultaneously consider how those predictions. To illustrate the kind of
his overt actions of manipulating and seeing understands every one of the possible Level changes relate to the changes in the vertical velopmental progression that is predic
the weight and spring, because the repre- 6 systems in the gadget, for example, he and horizontal lengths of the cord, H e can and to emphasize the applicability of
sentational sets are actually composed of cannot integrate them into a single higher- thus understand how all the changes covary. theory to things other than cold cogniti
sensory-motor systems specifying what the level skill. More generally, he cannot yet This skill not only allows him to control I will present a hypothesized sequencc
child can do with the gadget. understand objects independently of their the gadget effectively, but it also gives him the development of a person’s iden
Besides the representations for weight overt characteristics, because he is limited an abstract set for the general state of the (Erikson, 1963)-one’s sense of the k
and length of spring, the child could also to dealing with one Level 6 system at a time. gadget. of person one is.
construct representations for the length of That is, he cannot think of objects in the Many different kinds of abstractions can At Level 7, single abstract sets, a per
the vertical segment of the cord and the abstract. be constructed at Level 7. For example, a can for the first time construct absti
length of the horizontal segment (or depend- person can for the first time understand the identity skills (see Erikson, 1974). Th
ing on the nature of the specific gadget, a Abstract Tier: Levels 7 io I O abstract concept of conservation-varia- identity concepts result from the coordi
representation for the total length of the tions in two related quantities compensate tion of two representational systems ab
Level 7 is the culmination of representa- for each other so as to produce no change the self. For instance, a certain 9-year-
cord). A child who is very familiar with the
tional development, generating a new kind in some superordinate quantity. With only may have a Level 6 system for identificat
gadget could conceivably possess at least 12
of set and starting a new tier, the abstract Level 6 skills, the person can understand with his father’s career as a psycholog
different mappings, all possible pairings of
tier. In the recurring cycle of four levels, most of the individual kinds of conserva- He relates his representation of himsell
the four sets-weight, spring length, ver-
the characteristic structure for Level 7 is tion that Piaget and his colleagues have his representation of his father as a psycf
tical and horizontal lengths of cord (see
the system of representational systems documented (Piaget & Inhelder, 194111974; ogist (Kagan, 1958). Likewise, he has
Wilkering, 1979). Despite all this knowl-
(representational Level IV), which is the Piaget & Szeminska, 1941/1952), but cannot other system relating his representation
edge, the child’s understanding of the gadget
same as the single abstract set (abstract integrate those separate conservations into himself as both skilled with other people i
would be peculiarly disjointed because of
Level I). In an abstract set, the person an abstract concept of conservation. good at science to his representation
his inability to consider two aspects of each
abstracts an intangible attribute that charac- For instance, the person combines the what psychologists do: They are peog
set simultaneously. That is why, for ex-
ample, he has difficulty treating the vertical skill for conservation of the length of the oriented scientists. Most 9-year-olds are
and horizontal lengths as segments of a ’ Piaget does not postulate the existence of a major cord in the gadget with the ski11 for con- yet capable of coordinating two such Lev1
single cord of constant length. cognitive-developmental change in the middle pre- servation of amount of clay (where the same systems into a Level 7 skill.
school years, but some of his own research suggesls
Level 6 is characterized by a representa- that there might be such a change (Piaget et al., 1%8), piece of clay is squeezed into different A few years later, when the child (
tional system, in which the child relates two and many studies over the last two decades have shapes, such as from a ball to a sausage). coordinate the two systems, he can tl
subsets of one representation, ‘Rj,K, to two documented that preschool children’s abilities are far In the conservation-of-length task, the two construct his first abstract set for his car,
subsets of a second representation, 6TJ,K, greater than prior research had indicated (Gelman, lengths are equal because the vertical and identity. With the addition of a few otl
1978). Also, two developmental theories that are not as horizontal lengths compensate for each representational systems to the Level 7 si
as shown in Table 3. For example, he can well known as Piaget’s posit a major developmental
understand conservation of length of the shift at about age 4 (Bickhard, 1978; Isaac & O‘Connor, other. In the conservation-of-clay task the via microdevelopmental transformatio
cord in the gadget, as described earlier. He 1975). two amounts are equal because changes in he can build a complex abstract set relat
 496 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT L

various of his own characteristics to various progression from pre-operational to formal ing a computer (Fischer & Lazerson, in sets when one of the other four transforr
aspects a f t h e career mat he is considefing- operatianal thuglit, accadhg t~ sk;iA press, chap. 13). tiofff wcws, b& it can &e be & se
At Level 8, abstract mappings, the person theory, because both the Piagetian periods rately to predict microdevelopmental ste
can relate one abstract identity concept with and the scientific progressions involve Transformation Rules The microdevelopmental transformatic
another. For example, he can coarsely development within a tier from Levels I of differentiation, substitution, focusi
relate his own career identity with his con- to IV.8 Oddly, Piaget too (1970; Piaget in Now that the structures of the levels have and compounding eventually produce
ception of his potential spouse’s career Beth & Piaget, 1961/1%6) has suggested been described, the operation of the five macrodevelopmental transformation
identity: Perhaps he sees his own career that there may be general parallels between transformation rules can be illustrated with intercoordination. These five transforn
identity as requiring that his spouse be in a the development of scientific theories and some precision. The five rules specify how tion rules are probably not exhausti
closely related career or perhaps as re- the development of cognition in the child. a skill is transformed into a new, more ad- future research will indicate whether I

quiring that his spouse be primarily a home- I say “oddly” because his position on vanced skill. These rules are thus the heart ditional transformation rules are require
maker. formal operations seems to preclude such of the mechanism for predicting specific All the transformations are defip
Level 9 abstract systems produce a much parallels. sequences of development. The need for structurally. Two or more skills with gi\
more flexible, differentiated relation be- Within Piaget’s framework, cognitive de- such a set of transition rules to account structures are transformed into one or mc
tween two identity concepts. For instance, velopment virtually ends with formal opera- for developmental change has been recog- skills with a new type of structure. 1
the person can relate two aspects of his tions: Adolescents entering the formal nized for a long time by many develop- induction of a specific structural transforn
own and his spouse’s identity, such as operational period have achieved fully mental psychologists (e.g., Beilin, 1971; tion always involves both organismic a
career and parental identities, and thus con- logical thinking, and there is little more for Brainerd, 1976; Flavell, 1963; Kessen, 1966; environmental factors. At least two
sider in a more differentiated way what his them to do, except perhaps to extend their Van den Daele, 1976). The rules are also ganismic factors are involved: The pers
own identity requires of his spouse’s logical thinking to new content areas (Piaget, intended to apply to changes in the organiza- must initially have the skills required
identity and what his spouse‘s identity 1972). Many people have been dissatisfied tion of behavior during learning or problem application of the transformation and mi
requires of his own identity. with this conception of formal operations solving (Fischer, 1975, 1980: Leiser, 1977). be capable of applying the transformati
Finally, at Level 10, systems of abstract (e.g., Arlin, 1975; Gruber & V o n k h e , 1976: The transformation rules and the skill rules to those skills. For example, if a p
systems. this person can coordinate two or Riegel, 1975; Wason, 1977), but there has structures of the levels should be able to son has the necessary initial skills but th
more abstract identity systems. He might been no alternative position for analyzing explain most of the developmental se- are already at her optimal level, then she v
relate his own and his spouse’s career and development beyond early adolescence. quences documented in the research litera- not be able to apply the transformation i
parental identities now (one Level 9 system) Consequently, major age differences in the ture. In addition, many new sequences can combining those skills to reach the nt
with their career and parental identities acquisition of various of Piaget’s formal be predicted that have not yet been investi- higher level.
10 years ago when they were first married operational tasks have been interpreted pri- gated. In this section on the transformation Likewise, at least two environmen
(a second Level 9 sqstem). The result is a marily as resulting from performance rules, however, I will refrain from reviewing properties are necessary. First, the envirc
higher-level conception of what their joint factors, not from developmental changes empirical support for the rules, so that I ment must have properties such that if t
career and parental identities have been like (Inhelder & Piaget, 1955/1958; Martarano, can present the concepts briefly and directly. initial skills are transformed, the resulti
during their marriage. 1977; Neimark, 1975). According to skill In a later section, several studies testing new skill will work. Second, the speci
Although I know of no rigorous tests of theory, many of these age differences may predictions based on the rules will be de- environmental situation must have prc
this or any other developmental sequences well arise because the tasks require dif- scribed. erties such that it will induce the person
in abstract skills during adolescence and ferent levels of abstraction. The five transformation rules are inter- use the initially separate skills in juxtapo
adulthood, several investigators have re- Piagetian scientific tasks and the rarefied coordination, compounding, focusing, sub- tion, thus leading her to explore the relatio
ported data that generally support the pre- atmosphere of theory construction are not stitution, and differentiation. Intercoordina- between the initial skills and construct t
dictions of development from Levels 7 to 10. the only places that skills should develop tion and compounding specify how skills transformed skill (see also Schaeffer, 197
Some of the most detailed findings involve through Levels 7 to 10. Most adults prob- are combined to produce new skills. Inter- The transformation therefore requires be
developments in the history of science. Both ably master at least a few skills beyond coordination describes combinations that organism and environment; transformatio
Miller (Note 5) and Gruber (1973; Gruber & Level 7, like the hypothesized identity con- produce development from one level to the cannot be attributed to either organism
Barrett, 1974) have described developments cepts. Other skills that probably belong to next (macrodevelopment), and com- environment alone.
of scientific theory that seemed to them to Levels 7 to 10 include moral judgment, the pounding describes combinations that Two of the transformation rules, intc
roughly follow Piaget’s description of cog- managerial skills of the director of a corpora- produce development within a level (micro- coordination and compounding, invol
nitive development from the pre-operational tion or a school system, the skills required development). Focusing and substitution combinations of two skills to produce a ne
period to the formal operational period. to write an effective essay or novel, and the specify smaller microdevelopmental steps more complex skill. Many psychologir
Miller illustrates this parallel for the de- skills involved in programming and operat- than compounding. Focusing deals with have talked about combinations of skills
velopment of quantum mechanics, and moment-to-moment shifts from one skill to a mechanism to explain the development
Gruber for the development of Darwin’s another, and substitution designates certain more complex skills, especially in the litex
theory of evolution. If these scientific a Within the present theory, Piaget‘s pre-operational, cases of generalization of a skill. The fifth ture on skill acquisition (e.g., Bruner, 197
theories were developing through Levels 7 concrete operational, and formal operational periods rule, differentiation, indicates how sets be- 1973; Fitts & Posner, 1967) and the Piageti:
to 10, their progression would resemble the are explained by Levels 4 to 7. come separated into potentially distinct sub- literature (e.g., Cunningham, 1972; Hur
 ~~

498 KURT W. FISCHER

-
THEORY O F COGNITIVE DEVELOPMENT

1975; Piaget, 1936/1952).The first two trans- ments. When her skills with the gadget are cause it provides a specifically apgropriate 13P * 3H1 + I3H ’SI =
f”*&- to specify e m m € yat L e d 5, sfie can uncterstand t w o M p t b n of ttre pnzcas of c o m b i ”
how such skill combinations occur. separate representational mappings in- [3P ++ 3H ++ (.
of two skills at one level to form a new skill
volving the vertical and horizontal segments at the next level. Intercoordination means She thus produces a more complex Level
lntercoordination of the cord. She can understand the mapping reciprocal coordination. Two lower-level skill that allows her to control the relatioi
from the horizontal segment, TH, to the skills become coordinated with each other among all three actions.
Intercoordination specifies how the vertical segment, Tv,using the horizontal and thereby produce a new higher-level What behavioral consequences does SUC
person combines skills to develop from level length to predict the vertical length; and she skill. In this article, the term intercoordina- a skill have? When the child is actually
to level, all the way from Level 1 to Level can understand the mapping from the ver- tion is reserved explicitly for this process. bed, it is difficult to assess whether she hi
10. The process is analogous to the com- tical segment, TV,to the horizontal seg- The term coordination is used to refer in the skill, because the context alone prl
bination of atoms to form a molecule. At ment, jCH,using the vertical length to pre- general to all instances of a person’s relating duces all three components. As a result, sl
the beginning of the process of interco- dict the horizontal. As a result, the child’s two or more actions whether or not he or she can show all three actions together withoi
ordination, the child has two well-formed behavior with the cord at Level 5 seems is going through the process of interco- controlling the relations among them. Tl
skills, a and b , at a specific Level 1. The paradoxical to an adult. The child seems to ordination. existence of this environmentally elicite
two skills are functioning separately from understand how the horizontal and vertical conjunction of the three actions, on tf
each other until some object or event in the segments of the cord change and how the other hand, shows how the environment ci
environment induces the child to relate the changes relate to each other, yet she does Compounding
induce combination of the actions into
two skills to each other. The child then not recognize that the total length of the cord
works out the relationship between the two must always remain the same because the The second rule for combination of skills, complex skill. At least once or twice evei
skills with that object or event and so changes compensate for each other (Wil- compounding, specifies the most important day, the child is faced with a situatic
gradually intercoordinates the skills. When kering, 1979). microdevelopmentat transformation. In where the three actions go together. Almo
the intercoordination is complete, the two The paradox is resolved when the child compounding, two skills, a and b , at Level L inevitably, then, she will explore the re1
skills, a and b , from Level L have been intercoordinates the two mappings to are combined to form a more complex skill, tions among the actions and ultimatei
transformed into a new skill, d , at Level produce the Level 6 skill for conservation c . at the same Level L.y The process is produce the compounded skill in Equation.
For assessment of this compounded skil
L + 1, which includes them. The process of length in the gadget, as follows: diagramed as follows:
pretend play provides a better situatic
is diagramed as follows:
[jCH- T”].[5CC- 5cHJ= a+b=c. (3) than going to bed, because in pretend pla
a .b = d . (1) [BC,,V c-) ICH.,l. (2) the child must actively put the three actior
The addition symbol signifies compounding. together. When she can control all thrc
The multiplication symbol signifies inter- The child with this Level 6 skill can under- Just as with intercoordination, compound- actions in a single system, then she ca
coordination (Table 2). stand how vertical and horizontal length ing is induced by both organismic and en- pretend to go to sleep-holding the pillov
The essence of the process of interco- interrelate, instead of merely how vertical vironmental factors. First, the person must placing her head on it, and closing her eyt
ordination lies in what seems to most adults relates to horizontal and how horizontal possess skills a and b , and second, some (Watson & Fischer, 1977). Other develol
to be a paradox. A child is given a task relates to vertical (Verge & Bogartz, 1978). object or event in the environment must in- mental sequences involving compoundir
that normally requires a Level L under- That IS, she has constructed a skill for the duce him or her to combine them. have been tested and supported by Bertei
standing, but her skill for that task is only at total length of the cord (composed of hori- An example will clarify the compounding thal and Fischer (1978), Watson and Fischc
Level L - 1. Consequently, she seems to zontal and vertical components), and that process. Consider an 18-month-old who has (1980), Hand (1980), and Fischer an
have all the knowledge that is needed to skill allows her not only to predict how one Level 3 skill relating holding a pillow, Roberts (Note 3).
perform the task, yet cannot d o it. Only horizontal and vertical vary but also how 3P,to placing her head on the pillow, 3H, and The process of compounding is not nece:
when she intercoordinates the relevant skills their covariations sum to a constant length. a second Level 3 skill relating placing her sarily limited to combining just two simp1
a t Level L - I to form the new skill at The formal descriptions of Levels 1 to 7 head on the pillow, 3H, to going to sleep, 3S. skills. It also has more complex forms, wit
Level L will she be able to perform the task. in Table 3 indicate graphically how inter- (Going to sleep is originally an involuntary combinations of larger numbers of skill:
Note, however, that this process of inter- coordination occurs at each level. Re- response, but like blinking and breathing, it For example, the child might compoun
coordination is gradual and continuous. The peatedly as one moves down the table, the can be brought under partial voluntary the Level 3 skill in Equation 4 so that
fact that it involves qualitative change does combination of two skills a t one level pro- control. Many children at this age expend a included four or five actions instead of on1
not in any way imply that the change is duces a new kind of skill a t the next level. lot of effort trying to control sleeping and the three. lo Indeed, such successive compounc
abrupt or discontinuous. The diagrams in Table 3 also show how each circumstances that accompany it. Usually,
The development of conservation of skill at one level still includes the two skills the specific response that is signified by 3S is Compounding of another sort might also occur. Tf
length in the gadget (Piaget et al., 1968) from the previous level. The formal de- voluntarily closing the eyes or saying skill d at Level L + 1 might be combined with skill
provides a clear illustration of this paradox. scriptions thus show both the origin of each “sleep.”) With these two skills, the child at Level L to form the more complex skill e at Lev
As stated before, the child must have a higher-level skill in the lower-level skills still cannot control all three actions at once: L + 1. Research is required to determine whether bot
Level 6 skill to understand that the length of and the emergence of a new type of skill at types of compounding exist.
holding the pillow, placing her head on the Two points should be noted about this comple
the cord conserves despite changes in the the higher level. pillow, and going to sleep. T o do so, she kind of compounding. First, the number of skil
lengths of the vertical and horizontal seg- I chose the term intercoordination be- must compound the two Level 3 skills: compounded cannot be endless because compoundin
500 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5

- 5L],
ing may ultimately account for much of the
pf3€e*e f - k t e f , as € s a ilks-
trate later.
CaB hadie @*
task determine the limit of what the person
kt that task. & any
one moment, she cannot bring to bear all of
It allows predictions of certain kinds of
Be%*& erderiffgs. €m*z
that can be solved with, at a minimum,
a ti%&
[5W - ’CV (1
which involves control of all three variabl
at once instead of only two.
Compounding describes relatively large her skills; normally, she can deal with only two skills at Level L and a shift in focus
According to this analysis, then, the ch
microdevelopmental steps. The next two one skill at a time. Focusing describes the from one of the skills to the other. This task
will show the following microdevelopmenl
transformation rules, focusing and sub- person’s shifts from one skill to another is more complex than a task that can be
sequence of skills: first Formula 8 or 9, th
stitution, describe smaller microdevelop- within the level or levels at which she is solved with, at a minimum, one skill at Level
Formula 10, then Formula 11, and fina
mental changes. functioning in the task. For instance, the L. If the two tasks are within the same
Formula 12. Similarly, the focusing ri
hypothetical child knows a lot about the task domain, then the first, more complex
task is predicted to develop after the second predicts many microdevelopmental E
Focusing: Moment-to-Moment Behavior gadget, because she has mastered all 12 quences, such as transitional steps betwe,
mappings. Nevertheless, her understanding task.
Focusing deals with one kind of shift in For example, suppose that the gadget is acquisition of the simple Level 3 skills I
of the gadget is severely limited by the fact the left-hand side of Equation 4 and acqui
what is commonly called attention. It that she can focus on only one mapping at partially covered, so that only two variables
describes not only a type of developmental are visible at a time. The child first deals tion of the compounded Level 3 skill on t
a time. right-hand side of Equation 4. See Hal
change but also a type of moment-to- Say that at a certain moment she is con- with only the weight, 5W, and the vertical
moment change in behavior. In a specific (1980) and Watson (1978) for tests of a
sidering the set, 5W, of weights. She cannot segment, ’CV, using the skill
task or context, a person will normally ditional sequences involving focusing.
deal with all six of the mappings for weight
have a collection of skills available, and but must focus on just one, such as the [5W - ‘CV]. (8)
those skills will generally be related to each mapping of weight, 5W, onto length of the Substitution
other because subgroups of them will share spring, T. A few moments later, she shifts Once she has used this skill, the cover is
one or more sets. For example, recall the changed so that she can see only the vertical The transformation rule of substitutit
focus to a second, related skill, the mapping
hypothetical child who has a complete Level segment and the spring, which requires the deals with one type of generalization:
of horizontal length of the cord, ’CH, onto
5 understanding of the gadget (without skill skill at Level L is mastered with one tas
length of the spring, 5L,and then she shifts
compounding): She understands 12 dif- focus to the mapping of length of the and then the person attempts to transfer
ferent two-set mappings, all possible pair- [Tv- ’L]. (9) to a second, similar task. The rule appli
spring, ‘L, onto horizontal length of the
ings of the four sets involving the gadget. cord, ‘CH. These changes in focus are By shifting what is covered, the experi- when all components but one in the first ta
In this collection of skills, each set is in- diagramed as foIlows: menter can thus control the child’s change are identical with those in the second ta
cluded in six of the mappings. At a given in focus: and when that one different component c.
moment with the gadget, the child will be [’w- ’L] > [“CH -’I51 > be generalized to the second task.
using one of the mappings. When she shifts [jL-’CI;I. (7)
[jw- ‘CV]> [jcv- 5LI. (IO) Levels I1 and 111, the component must be
focus, she shifts from one specific mapping For the child to do this task as described set; at Level IV (which is Level I of the ne
to a second closely related mapping that Clearly, changes in focus can produce in Formula 10, she must have both Level 5 tier), it can be a set o r a system. The sk
shares at least one set with the first map- very complicated sequences of behavior. In skills. The focusing rule therefore predicts with the substitute component will 1
ping.” A shift in focus from skill e to skillf assessing a person’s skills with a task, care that the skills in Formulas 8 and 9 will mastered after the original skill and befo
I S represented symbolically as follows: must be taken to separate mere changes in develop before the change in focus in any skills of greater complexity in the san
e >f. focus from the actual control of relations Formula 10. A developmental sequence of task domain. Substitution is diagramed
between sets. The shifts in focus indicated this type has been demonstrated by Gottlieb, follows:
The symbol for “greater than” thus signifies in Formula 7 do not demonstrate control by Taylor, and Ruderman (1977).
a shift in focus. When a shift in focus can the child of the compounded skill [’W - Sub d = dl, (1
With the covering procedure, the experi-
be consistently controlled by the child, the 5L - ‘CH- 5L- jCd, although under menter can teach the child to change focus or for a specific level,
transformation is diagramed: the proper environmental circumstances consistently. The child will thus learn a new Sub [‘A - ’B] = [‘A -‘Ell]. (1
they can be transitional to the formation of skill involving a change of focus:
Foc ( e , f ) = [e >f]. (6) such a compounded skill. The set ’B, is the substitute set.
The levels of the collection of skills that a Focusing is not, however, merely a FOC([5W- ‘Cv], [jCv- ‘L]) = The skill for pretending to go to sIet
person has available to her on a specific statement of a methodological difficulty. r(5w - ~c,) > (’cy- 5 ~ 1 1 , (I I) provides an example of the application
this rule. After the child develops the sk
l 1 Notice that according to this definition, focusing
which will allow her to do the task even in Equation 4, she extends that skill to
is probably seriously limited by memory. Limitations
should be much less severe when the skills are inter- could not happen at Level I , because skills at Level I when all three variables are uncovered. substitute object. Instead of holding h
coordinated into a higher-level skill. Second, a mere each include only one set. However, the theory pre- This controlled-focusing skill is slightly pillow and pretending to go to sleep, sl
string of behavior, occurring once or twice, does not dicts that Level 1 sensory-motor sets are generated more advanced developmentally than the substitutes a piece of cloth for the pillow
indicate that the person has compounded all the by a prior reflex tier, which specifies the components simple change in focus in Formula 10. It is
behaviors in the sequence into a single skill. Just as of a Level 1 set and thereby indicates how focusing Sub [3P e 3H tf 3S]
with the little girl in bed. one must be certain that applies to Level 1 . The nature of this tier wifl be also transitional to the compounded Level
the person actually controls the relations between sets. suggested later. 5 skill = [3P, fz 3H tf (1
 502 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 50

The set 3P1,holding the cloth, is substituted Diff A = AH,AN. (17) At the same time, with a gadget like the tion of the cerd, but because of fatigue E
fur the originat set ?, hoTding the p i o w . *&6fte, iR W k h €he mrd is st8 e " d r r y s e t is nowfunctioning at Levr
The extension of the pretending skill to the The development of conservation of divided into vertical and horizontal seg- 5. She can use a skill that would not be poi
cloth develops after the original skill (on the length in the gadget illustrates how dif- ments, he or she could tell that the vertical sible for someone who has never develope
left-hand side of Equation 15) and before ferentiation occurs when a new skill i s and horizontal segments were each different this skill to Level 6. She might use th
any more complex skills (Watson & Fischer, formed. A child with Level 5 mappings for
from the total cord in the modified gadget. coordinated lengths of the two segments c
1977). the gadget understands generally how the Likewise, certain experimental training pro- the cord to make coarse, qualitative Level
length of the vertical segment relates to the cedures can produce such separation or dis- predictions about the length of the spring:
length of the horizontal segment and vice crimination (Denney, Zeytinoglu, & Selzer,
Differentiation versa but does not yet understand conserva- W Y , H - 6LI (IS
1977).
The final transformation rule for explain- tion of the total length of the cord. Another The interaction of task and level helps to
way of stating this confusion is that in this So far, I have emphasized general issue
ing development is differentiation, in which resolve a paradox in the developmental about differentiation because they are in-
what was initially a single set becomes task, the child has not adequately differ- literature. In some experiments, young chil-
entiated the total length of the cord from the portant for understanding how differentii
separated into distinct subsets. Differentia- dren confuse variables like the several types tion and combination work together in ski
tion is probably always a product of one of lengths of the horizontal and vertical seg- of cord length in the gadget, but in other
ments. When asked about the total length theory. But differentiation can also be usel
the other transformations, especially inter- experiments children of the same age easily as a developmental transformation rule
coordination or compounding. As Werner of the cord, the child confuses it with the separate variables that seem at first to be the
length of the horizontal or vertical segment. That is, it can be used to predict steps in
(1948, 1957) has argued, differentiation and same as the ones they confused (Kemler & developmental sequence. In the spring
integration always occur together. In skill Although this kind of lack of differentiation Smith, 1978; Smith & Kemler, 1978). In-
may seem odd to an adult, it occurs com- and-cord gadget, a skill for coarsely pre
theory, differentiation and integration (com- deed, the same child can show both kinds of dicting vertical length from horizonta
bination) are thus complementary, whereas monly in children and indeed is charac-
skills-ones demonstrating a global, syn- length is less differentiated than a skill fo
in many other approaches they are opposed teristic of earlier cognitive levels (Smith &
cretic whole that confounds several vari- predicting the same thing more exactly; ani
(e.g., Kaye, 1979; McGurk & MacDonald, Kemler, 1977; Werner, 1948). ables and ones using virtually the same vari-
The lack of differentiation in the gadget is the coarser skill will develop earlier than thi
1978). ables separately (Peters, 1977). In the tasks more differentiated one. In a sorting task
Differentiation can therefore be either resolved when the child intercoordinates the
where she uses the variables syncretically, the skill for putting different shades of rei
microdevelopmental or macrodevelop- two Level 5 mappings to form the Level 6 the child must deal with a number of related
system for conservation of the total length into a single category is more differentiate(
mental, depending on which other trans- variables at the same time, and her skill level than the skill for putting identical shades o
formation is involved. For macrodevelop- of the cord, as shown in Equation 2. The
is not sufficiently advanced for her to sepa- red into a single category, and the more dif
ment, the degree of differentiation is so great intercoordination produces differentiation
rate the variables. But in the tasks where ferentiated skill will develop later (Fische
that a set at Level L should be considered of the set for total length, 6Cv,H,from the
she separates them, she does not need to & Roberts, Note 3).
a different set when it reaches Level L + 1. sets for vertical length, T,, and horizontal deal with all of them simultaneously: able to
At higher levels, earlier global sets are length, VH:
work with first one variable and then an- Ordering the Results of Transformations
divided into distinct new sets that serve Diff ('CL, 'CH) 'CH, other, she can easily separate them.
= bCb, T,,H. (18)
in place of the earlier sets. (The superscripts This separation is, of course, not the same With five different transformation rules
to the left of the capital letters designating In the set for total length, the child com- as the differentiation that is required to co- some principles are needed for ordering thr
sets- see Tables 2 and 3-indicate the bines covariations in vertical and horizontal ordinate all the variables in a single skill. results of the different transformations intc
level of the set and thereby serve as a lengths into a concept of total length. Note For instance, with the Level 6 conservation developmental sequences. First of all, for i
reminder that a set differentiates as it de- also that the sets for vertical and horizontal skill in the gadget, the child must differ- clear-cut prediction of a sequence to obtain
velops to higher levels). Because of the lengths can be differentiated more finely at entiate covariations in vertical and hori- all skills must be in the same task domain
formation of these new sets, the person Level 6 than at Level 5: The child can deal zontal lengths and combine them into a con- Given that they are in the same domain
controls an ever larger repertoire of sets as with smaller variations in length in each of cept of total length. The three types of length the following principles allow ordering o
development proceeds. The expansion of the sets. are not merely separated; they are also steps:
the number of sets leads to a corresponding The specific variables that are separated integrated. 1. If one of the transformations is appliec
increase in the number of skills, since the in a child's behavior are a function not only The relation between differentiation and to a skill or skills, the skill resulting fron
newly differentiated sets can become of the level but also of the particular task. cognitive level has many other implications the transformation will develop after thc
separate components in new skills. For a child with skills at a given level. for analyzing development, according to initial skills.
The process of differentiation is dia- changes in the task alone can produce skill theory. For example, when a person 2. Starting with specific skills at Level L.
gramed as follows: separation. For example, if the cord in the has at some point developed a skill to Level a skill resulting from an intercoordinatior
gadget (Figure 1) were straightened out, a L but is now using the skill or some of its transformation will develop after a skiL
Diff d = d,,,, (16) child with the Level 5 skills in Equations 2 components at a lower level, the sets will resulting from microdevelopmental trans-
where the subscripts indicate subsets in the and 18 could easily control the set for the still show the effects of the earlier differ- formations, because the skill resulting from
skill d. Differentiation of a specific set A is total length of the cord in the modified entiation at Level L. Suppose that a child intercoordination will be at Level L + 1.
designated: gadget, since it would be only a single set. has developed the Level 6 skill for conserva- 3. When Principles 1 and 2 do not provide
504 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 51

other (Brown, 1965; Mead, 1934). For systems are essentially

an ordering, then for skills at the same level, Guidelines for Task Analysis . . role-specific b
Htefe wirk H t e f e fe+s*tleAq? &*. ffls€iSKx3,Hte se€istf fefe e€m&Ie€ requires h V i c mor . Fur exampte,
4. If more than one skill is involved in a Use of the theory to explain development the complementary role of child, and the doctor gives a patient inoculations (01
behavior (e.g., because of the focusing requires a behavioral analysis of perform- social role of doctor requires the comple- sensory-motor system) and examines h
transformation), that behavior will develop ance on the specific task or tasks in ques- mentary role of patient. Which skill struc- ears (a second system), and a patient tak,
later than the behaviors specified by each tion: What exactly must a person do to per- ture is required, then, for a child to under- the inoculation and poses for the ear e
skill separately (by Principle 1). A skill that form each task? stand a social role, such as a doctor ex- amination.
compounds the several skills into one will This kind of behavioral analysis is not as amining a patient? 3. What are the relations between se
develop later than the same skills connected simple as it may seem. The situation is The guidelines for task analysis fall within that the child must control (among tl
by a change in focus. If the focusing skills analogous to that of a behaviorist trying to two general categories: guidelines for deter- various possible relations shown in Table 3
involve a greater number of distinct com- determine which specific operants or re- mining what the person must control and Once the first two questions have bet
ponents than the compounded skill, then as sponses comprise performance on a given guidelines for designing and interpreting answered, determination of the relatioi
a rule of thumb the focusing skills will task. Finding the operants is no easy particular tasks. between sets is often simple. For the doctc
develop later. matter (Breland & Breland, 1961; Schick, role, the set for doctor must have at lea
Notice that by these principles, many 1971). Control a mapping relation with the set for patient
pairs of skills cannot be ordered develop- On the other hand, many investigators [iR D - 'SF']. (2
mentally. Of course, skills that have the have been highly successful in analyzing At feast three major questions are in-
same type of structure but different com- behavior into its natural units (see de Villiers volved in analyzing what the person must With a mapping, the child can relate tf
ponents cannot be ordered because they are & Herrnstein, 1976; Marler & Hamilton, control. doctor role to the patient role, which is i
in different task domains. But in addition, 1966). Premack (1965), for instance, found 1 . Does the skill require sensory-motor, that is necessary to meet the minimal CI
many skills within the same domain can- that simple observation of the actions that representational, o r abstract sets? For terion of relating a social role to its compl,
not be ordered- for example, two skills tend to recur regularly in an animal's be- understanding a social role, sensory-motor ment.
that are the same except that one has a havior allowed him to infer a long list of sets are clearly not sufficient, since the role
substituted set and one has a differentiated natural operants that formed a hierarchy of involves more than the child's own actions. Tasks
set. reinforcers. And Fischer (1970, 1980) found Representational sets are necessary, be- Thus far, the skill analysis for the role I
With the descriptions of the transforma- that changes in patterns of responding over cause the role involves the characteristics doctor has proceeded as if the skill could 1.
tion rules, all the major concepts of the trials in common learning situations demon- and actions of people independent of the considered independently of a particul;
theory have been presented. The next step strated the formation of new, higher-level child. Abstract sets are not needed, since task. But in fact, the analysis must take tl-
is to demonstrate how the theory functions behavioral units. social roles as defined here require only particular task into account. At least thrc
as a tool for explaining and predict- The skill structures specified in the theory concrete characteristics and actions of major issues are involved in designing an
ing development. After describing a set are intended to reflect the natural units of specific people (a doctor relating to a pa- interpreting specific tasks.
of guidelines for analyzing tasks and relating behavior (both thought and action), in- tient) rather than intangible attributes. 4. What is the particular task, and whi
them to the constructs of the theory, I will cluding its hierarchical character, with Understanding the general definition of a must the person control to perform it? Fc
present several experiments that have used higher-level units subsuming lower-level social role as involving one social category the role example, Watson and I devised
the theory to predict specific developmental ones. Determination of the validity of these and its complement would necessitate the task for assessing the child's understandin
sequences and synchronies and will then structures will, of course, require extensive control of intangible attributes, that is, of the role of doctor (Watson & Fische
explain a few of the many possible general research. abstract sets. 1980). Seated at a table, a child was show
deductions from the theory. Use of the theory to analyze behavior 2. What are the sources of variation that two rigid-cardboard, stand-up dolls (
into skill structures necessitates, first of all, the person must control in the skill? For the doctor and a child patient) and a few doctor'
Using the Theory to Predict Development a thorough knowledge of the available uni- doctor role, the child would have to control instruments. The experimenter acted OL
verse of skill structures defined by the levels two representational sets, not only the set the doctor's examination of the patient an
The theory can be used to predict and and transformation rules. Given that one for a person acting as doctor, RD, but also then asked the child to act out a simila
explain various developmental phenomena, has this knowledge, then task analysis can the set for a person filling the complementary story. The child was not asked to copy th
including developmental sequences and be facilitated by using a set of guidelines role of patient, Sp. Both of these sets are story precisely, so that no requirement c
synchronies, certain effects of the environ- that have been helpful for me and my col- required because according to the definition exact imitation would interfere with th
ment on developing skills, individual dif- leagues. To illustrate the use of these guide- of social role, a role must be related to its child's demonstrating her knowledge. T
ferences in development, the nature of lines, I will show how each one applies to complement. show the role of doctor, the child had t
developmental unevenness, and structural an analysis of the development of an under- Also, note that, by definition, at least have the doctor doll carry out at least tw,
relations among developing skills. But all of standing of a social role during childhood two sensory-motor action systems must appropriate actions in relation to the patien
these predictions and explanations depend (Watson, 1978; Watson & Fischer, 1980). comprise each representational set. because doll. The doctor might, for example, giv
on a prior step-task analysis (Brown & For a social category to be a social role. representational sets are formed from com- the patient a shot and look in her ears o
French, 1979; Gollin & Saravo, 1970; Klahr according to role theory, it must involve at binations of sensory-motor systems. For take the patient's temperature and examin
& Wallace, 1976). least two social categories in relation to each the roles of doctor and patient, the action her throat.
506 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5

In analysis of a particular task, sources assess that ability or concept (Bertenthal Predicting Developmental Sequences doctor, nurse, and patient (Step 3) but mc
u f v a r k t i m W i f t O f t e n i J f x ” ~ 8z Fisc!m, WW; €w=il,m. Beginning from a task analysis, one can
zn+v-*--mfewm:
that are not evident if one erroneously For the doctor role, the definition is that FOC([’Rn - ’Sp], VRD- TN]) =
attempts to consider the skill independently one agent must show doctorlike behaviors use the transformation rules to predict a
of a task. In the present case, the task in relation to a second agent, which must developmental sequence. The sequence can i(%D-’sp) >(%D-~T~)]. (2
brings no change in the basic mapping skill show reciprocal patientlike behaviors. A be either macrodevelopmental or micro-
developmental or both, and it can have The child can make the doctor deal with t
as diagramed in Formula 20. But the com- minimal task for this concept is the doll-play patient and then make the doctor deal wi
ponents of the representational sets are a task, with just two dolls, the doctor and the virtually any number of steps, depending on
the number of transformations that are used. the nurse, but does not integrate doctc
little more complicated than they appeared patient. Many children who can demon- nurse, and patient all together in the appr
in the analysis of Question 2. Because the strate the doctor role in this task will not There is no one true sequence that all
children will always show, because the priate role relations. This behavior is mo
child must manipulate the dolls, each repre- show it in a more complex task: If the
exact sequence that a child demonstrates advanced developmentally than the doct
sentational set must include a minimum of experimenter’s story involves, for instance, role at Step 2 because the child must posse
not just two but three sensory-motor a mother bringing her child patient to the will be determined to a great extent by the
particular tasks that he or she experiences. two complete Level 5 skills. The behavi
systems. For each representational set, the doctor’s office and consulting with the
Previous studies attempting to test detailed is less advanced than the compounded sk
child must manipulate the appropriate doll doctor and nurse while the patient is being at Step 3 because although it contains tl
in addition to performing at least two role- examined, many of the children will demon- developmental sequences (mostly predicted
from Piaget’s work) have shown a singular same components, they are not unified in
specific actions, such as giving an inocula- strate an apparent inability to understand a single skill.
tion and an ear examination. the role of doctor (Watson & Fischer, 1980). lack of success (e.g., Hooper, et al., 1979;
Kofsky, 1966). Tests of sequences predicted Another microdevelopmental step can I
One problem that can arise in interpreting 6. To go beyond an analysis of an in- predicted by use of the substitution rule:
particular tasks is that incorrect task analy- dividual task and predict a developmental from skill theory, however, have been
ses in the developmental literature may sequence, one must keep all tasks in the highly successful (Bertenthal & Fischer, Sub P R O - ‘Th - =
interfere with determination of what a per- sequence within the same task domain. With 1978; Hand, 1980: Tucker, 1979; Watson &
son actually must do to perform a task. For Fischer, 1977, 1980; Fischer & Roberts, [’RD- ’TN,x - ’Sp]. (2
the doctor role, for example, the levels and
instance, for Piaget’s final object-perman- transformation rules can be used to produce Note 3). The child shows the same behaviors as fi
ence task, where the child must find an Starting from the task analysis for the Step 3 but replaces the nurse doll with
an ordering of developmental complexity.
object that has been put through a series of doctor-role skill, one can predict many substitute object that does not normally
with tasks more (or less) complex than the
invisible displacements, most investigators developmental steps (Watson & Fischer, the nurse role, such as a plain adult ma
basic doctor-role task. But if those tasks use 1980). Table 4 shows just a few of them.
have assumed that the task requires the different procedures or varying roles (such as doll. This skill is more advanced develo
cognitive recreation of the invisible dis- Application of the compounding rule to the mentally than the skill on the left of Equ
mother-child), the theory cannot predict a doctor-role skill (Step 2) expands the doctor
placements by the child (Piaget, 1937/1954; precise developmental sequence. The many tion 22.
Uzgiris & Hunt, 1975). Recently, this inter- role to include a second complementary Thus, Equations 20, 21, and 22 lead 1
environmental and organismic factors that role, that of nurse, jTN, thus producing a
pretation has been questioned (Jennings, produce unevenness mean that develop- prediction of a four-step development
1975; Harris, Note 6), and several investi- more complex Level 5 skill (Step 3). The sequence. First, the child develops the bas
mental sequences can only be predicted child starts out with the two simple Level 5
gators have shown that the task does not unambiguously when as many sets as pos- doctor-role skill in Equation 20 (shown i
produce cognitive manipulation of rep- skills on the left of the transformation Step 2 in Table 4), then the skills resultir
resentations of invisible displacements sible are kept the same from one develop- equation in Table 4: one relating the doctor from the indicated transformations in tk
(Bertenthal & Fischer, in press; Corrigan, mental step to the next. To make clear pre- role to the complementary patient role and following order: Equation 21, Step 3 I
dictions from the task analysis of the doctor the other relating the doctor role to the
in press). Table 4, and Equation 22.
5. What is the minimal task that would role, the same demonstration procedure complementary nurse role. When those Besides these and many other micri
demonstrate the skill in question? If the should be used at every step, the same dolls two skills are combined by compounding,’* developmental predictions, macrodeveloj
skill is a specific concept, for example, should be included, and the doctor-patient they produce the skill on the right of the mental predictions can be made, of courst
one must first specify exactly what is meant relation should remain the basis of every equation: The child can make the doctor The intercoordination rule specifies tran
by the concept and then determine the step. The more microdevelopmental the deal with both the nurse and the patient in formations from level to level. Revers
easiest task that would demonstrate it. predicted sequence, the more essential such a way that the doctor takes into ac- of the intercoordination rule decompost
Without specification of a minimal task, it is that the content and procedure remain count the nurse’s role relation to the patient the doctor-role skill (Step 2 in Table 4) In1
erroneous inferences may be made about the same from one step to the next. (symbolized by the mapping of TNand its two component Level 4 skills: the simp
the child’s ability (Shatz, 1977). Task com- Even with these six guidelines, doing a ?SPin the compounded skill). representational sets for doctor, ‘RD, an
plexities that are basically irrelevant to the task analysis is no trivial matter. Unfortu- Besides the steps shown in Table 4, many
ability in question will overload the child nately, it still involves a degree more art than other microdevelopmental steps can be pre-
I would like. Yet once a task analysis is in dicted. With application of the focusing rule, Several alternative pairs of simple Level 5 ski1
cognitively and prevent him from showing could be combined to produce the same compoundc
his ability. The skill level at which a person hand. predictions based upon it follow fairly for instance, an intermediate step can be Level 5 skill relating doctor. nurse, and patient role
can control an ability o r concept is a func- easily from the levels and transformation predicted that is less developmentally ad- For example, [>RD--T,]could be compoundc
tion of the complexity of the task used to rules. vanced than the compounded skill relating with [I ,- +S,]
Table 4
A Devclopmc~ntolSequence of Socicil Role Playing

Cognitive Role-playing
Step level skill Example of behavior Skill structure Transformation rule

I 4: Representa- Behavioral role The child pretends that a doctor ['RD] Intercoordination:
tional sets doll uses a thermometer and
a syringe. ['RDj.['Si,]= Step 2

2 5 : Representa- Social role The child pretends that a doctor Compounding:

tianal doll examines a patient doll, ['R D -5 s 1.1 + ['Ru - 'TNI= Step 3
mappings and the patient doll makes
appropriate responses during
the examination.

3 Social role with The child pretends that a doctor [SRD - 5 TN - 5SS,] Focusing:
two compli- doll examines a patient doll
mentary roles and is aided by a nurse doll. Foc(["R,:-"S r l ,
Both patient and nurse
respond appropriately.

4 Shifting between The child pretends that a doctor -

[t5RR, %,) > ( W , -TN-V P ) ] Intercoordination:
family role and doll is the father of the patient
~RD--Si,j.~R'R,--S,j=Ste~ 5
doctor role doll, and then he or she
switches to having the doctor
doll fill only the doctor role-
examining the patient doll
with the help of the nurse
doll, as in Step 3.

5 6: Representa- Intersection of The child pretends that a doctor tBRw ++ "SI.,( 1 Compounding:"
two roles and doll examines a patient doll
tional
systems their comple- and also acts as a father to L"Rii,, * "SP,~
I + ['RDJI* "VM,W]
ments the patient, who is his son or
daughter. The patient doll
+ L'Vv,.,, * "Si,(1 = Step 6
acts dppropriately as both
pdtient and offspring
(Tuble conlitiric> )
-*

Role-playing
Step skill Example of behavior Skill structure

6 Intersection of The child pretends that the [%F." ++ 'VM,W ++ "SP,,Ih

three roles and doctor doll is doctor, father,
their comple- and husband, relating to the
ments patient doll, who is a patient
and the man's offspring, and
to the woman doll, who is the
patient's mother and the
man's wife.

Note. In the formulas, the italicized capital letters stand for the child's representation of a particular doll as an independent agent: R for the doctor
doll, S for the child doll, T for the nurse doll, and V for the woman doll. The subscripts designate the role or roles that the child represents for each
doll, as follows: C = child; D = doctor; F = father; H = husband; M = mother; N = nurse; P = patient; W = wife.
For most steps, several alternative forms of skills involving the same sets could be combined to produce essentially the same new skill. An
example of such alternative forms is given in Footnote 12.
I,In Step 6, the doctor doll carries out three roles, whereas only two roles are listed for each of the other two dolls. Formal distinctions can be
made between closely related roles so that these two dolls would also each carry out three roles. For example, the woman doll could be not onlv
mother and wife but also adult-responsible-for-child-in-the-doctor's-office. In practice, however, children' will usually ignore such subtle rolk
differentiations.

/'

/'

/
 ~~~ _ _ _ _ _ _ _ ~ ~ ~~

510 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5

Predicting Developmental Synchronies the object-permanence tasks, it was pre- grees of synchrony can be tested. Consider Longitudinal testing should therefo
AcrDss TmE Domaims &ae&ame woutct atso txegfn to aewords -seqt”s x,y,d z ,wkere p r Q m e El&r irLQY- tliFa& t d
to ask about or refer to objects that were x and y are hypothesized to involve virtually velopmental sequence than cross-section
Because of the importance of environ- not present, especially in the object-per- the same skills, and z is hypothesized to testing. Jackson, Campos, and Fisch
mental factors, precise predictions of micro- manence tasks.I3 For example, he would involve different skills. Within a given age (1978) tested this prediction by comparii
developmental sequences can be made only start using “all gone” and “more” appropri- range, sequences x and y should correlate the effects of longitudinal and cross-se
within a task domain, where most of the ately for the absent objects. This corres- together more highly than either of them tional procedures on development throug
components are the same o r very similar pondence was predicted from skill theory correlates with sequence z. Similarly, if a an eight-step sequence of object perm
for adjacent steps in a sequence. The pre- because with Level 4 skills, the child con- particular environmental condition such as nence. Longitudinal testingproduced a larj
diction of synchronies across task domains trols representational sets and therefore can practice is hypothesized to increase the practice effect, as predicted: two to thrc
is much more complicated, because few or understand that objects are agents of action synchrony between two developmental se- steps in the eight-step sequence.
n o components are shared across domains. independent of him: He can control the quences, then the correlation between the Because of this practice effect, long
Yet predictions about synchrony are clear. representational set o r sets for objects in sequences under that condition should be tudinal testing should produce an inflate
First, because unevenness is the rule in those tasks even when he cannot perceive higher than the correlation under other en- estimate ofthe synchrony between develol
development, the degree of developmental the objects. He should, therefore, be able vironmental conditions. Indeed, according ment in two different task domains. Usuall!
synchrony between two task domains will to speak spontaneously about objects that to skill theory, an experimenter should be in a group of children who have not exper
seldom be high. It will usually be moderate disappear in those tasks or in similar situa- able to control the degree of synchrony enced longitudinal testing, most of th
for familiar domains, because each inde- tions. Corrigan’s findings supported this that he or she will obtain by simple en- children will have differential experienc
pendent skill develops with age and this prediction of precise correspondence vironmental manipulations. with the skills in any two domains. Const
relation with age produces some correlation between object permanence and use of “all quently, in cross-sectional testing, th
between the two skills. gone” and “more.” Role of the Environment synchrony between the two development;
Second, manipulation of environmental Testing for developmental synchronies sequences will not be high, except in th
factors such as degree of practice will between task domains is unfortunately more According to skill theory, environmental case where the sequences actually belon
drastically alter the degree of synchrony. complex methodologically than it at first factors play a central role in determining to the same skill domain. (Recall that a ski
For instance, sequences in two highly appears. The correlation produced by age the relative degree of synchrony between domain is composed of a group of tas
alone is a difficulty that is too often ignored. developmental sequences, and they also domains that develop in close synchrony
practiced domains should show nearly per-
fect synchrony, as will be explained later. If the age range tested is wide, the cor- affect the specific developmental sequences On the other hand, the extensive practic
Third, whenever developmental se- relations can be substantial. The develop- that people show. Some of these predictions that occurs in much longitudinal testin
ment of classification skills between 1 and are presented below, primarily for the virtually eliminates this differential experr
quences in two different domains intersect
so that a skill in one domain becomes part 7 years, for example, correlates highly with effects of specific testing procedures, in- ence and elevates the skills in both tas;
of a skill in the other, the development of shoe size, r(68) = 3 5 , p < .001 (Fischer & cluding the differences between longitudinal domains to the person’s optimal level. Con
the skill in the first domain will predict Roberts, Note 3). and cross-sectional procedures and the sequently, even when the skills are in fac
the development of the skill in the second. Skill theory suggests several ways of effects of the specific tasks used to test from independent skill domains, longi
This correspondence will be precise, with overcoming this problem in testing for de- developing skills. tudinal testing will usually produce a higl
virtually every child that develops through velopmental synchronies. First, when synchrony between them-and a higl
the two intersecting sequences showing the precise predictions can be made about Effects of Testing correlation.
predicted correspondence. exactly which developmental steps in- the Comgan’s study of language develop
Corrigan (1977, 1978, 1979, 1980) has Longitudinal and cross-sectional pro- ment and object-permanence developmen
found support for these predictions about cedures should produce very different tested this prediction (Comgan, 1977,1978)
synchrony between task domains for the tested directly rather than i patterns of synchrony across task domains, As I reported above, she found that for i
relationship between the development of co_xelations. as a function of the effects of practice. group of infants tested cross-sectionally
object permanence (finding hidden objects) Second, predictions about relative de- Because skills must be practiced to be the correlation between the developmenta
and the development of language. First, the mastered, a skill that is practiced regularly sequences was only .36. But for threc
general correlation between object perma- should develop faster than a skill that is infants who were tested longitudinally ovei
nence and language development in a group ‘ I This hypothesis may seem at first to contradict practiced less often. In most longitudinal the same age range, the correlations were
of infants between 10 and 26 months of age
the earlier discussion ofthe problems with the Piagetian studies, children are effectively given re- much higher: .75, .78, and .89 for the in-
task analysis of the invisible-displacements tasks. peated practice with the skills being in-
was only moderate, T (29) = .36, p < .01. There is no contradiction for two reasons: First, dividual infants. Liben’s (1977) study of the
Further analyses indicated that this correla- Comgan used a more complex testing procedure that vestigated, because they perform the same effects of training and practice on memory
tion was produced entirely by the relation seemed to provide a better assessment of the use of or similar tasks session after session. In improvement and Jackson et al.’s (1978;
of performance in each task domain to age. representation in object-permanence tasks than the most cross-sectional studies, on the other comparison of cross-sectional and longi-
Piagetian procedure. Second, an object-permanence hand, children are not given regular practice tudinal procedures also corroborate the
Second, one point of precise correspond- task can require representation for reasons other
ence between the two skills could be pre- than those embodied in the Piagetian task analysis
with the skills being investigated, because prediction.
dicted. When the child reached Level 4 for (Fischer & Jennings, in press). they are tested on each task only once. These findings thus support the argument
512 KURT W.FISCHER THEORY OF COGNITIVE DEVELOPMENT

that cross-sectional tgstilg in general pro- volving focusing, compounding, or differ- microdevelopmental sequences for object e n c e g the environment would induce la
vides a better test of the natura& Occurrirrg mtktI0n witt mt VP= ift tkeir b&w* fcefftftiKe fer example, €m- i"Jt"ce s in &evetoy".
svnchronv between task domains than if thev are not exDosed to the suecific tasks rigan, 1977, 1978; Opaluch, 1979; Uzgiris Individual differences can take sew
longitudinal testing. Developmental psy- # or siiuations that will induce those par- & Hunt, 1975; Wise, Wise, & Zimmerman, forms. People differ in rate of developmc
chologists commonly disparage the useful- ticular skills. 1974). Some move through the hierarchy of le%
ness of cross-sectional methodology in Even macrodevelopmental steps in- These effects of testing procedures on much faster than others. People differ
developmental research, but the ability to volving intercoordination may be skipped variations in sequences and on synchronies their profiles of cognitive skills-catalog1
predict developmental sequences makes for particular sequences. Recall, for in- across sequences are more than a mere of which skills have attained which levc
cross-sectional testing a powerful develop- stance, the Level 7 skill for the abstract methodological nuisance. They are a reflec- And most interestingly, people differ in
mental tool (Fischer, Note 7): Specific concept of conservation, that is, the concept tion of the general importance of environ- paths through which they develop.
parallel sequences can be predicted in differ- of quantities that d o not change because mental factors as determinants of cognitive Many cognitive-developmental psycf
ent task domains, and a separate task can be they are composed of two constituent development. ogists have assumed that all people n
devised for each step in each sequence. quantities that compensate for each other. mally develop through the same devell
Then, with cross-sectional testing of every Suppose that a person develops this con- Why Unevenness Must be The Ruie mental path in any single domain, but
person on every task, scalogram analysis cept of conservation without ever having cently a large number of researchers h:
can be used to test the validity of the se- developed the Level 6 skill for conservation If environmental factors are as important begun to argue that individual difference:
quences, and the synchrony between se- of length. Perhaps he coordinates a Level 6 as I have argued in determining sequence some o r al18developmental paths are I
quences can be compared step by step skill for conservation of amount of clay with and synchrony, then indeed unevenness norm (e.g., Braine, 1976; Nelson, 19
(Bertenthal & Fischer, 1978: Watson, 1978; another Level 6 skill for conservation of must be the rule in development. The level, Rest, 1976).
Watson & Fischer, 1977). number and so generates the abstract con- or step within a level, that an individual Skill theory predicts that individuals \.
Variations in testing procedure will affect cept of conservation without ever dealing attains on a task is affected by so many frequently follow different paths of deveh
not only the degree of synchrony but also with conservation of length. When he is environmental factors that he o r she could ment and that these differences will take
the particular developmental sequences then tested for conservation of length in not possibly perform at the same level or least two forms. First, different individu
that people show. Many developmental the spring-and-cord gadget, he can gen- step on all tasks. Jackson et al. (1978), in will develop in different skill domains. C
psychologists assume that every skill eralize the Level 7 skill for abstract con- their study of object permanence, examined person will develop basket-making skills t
domain shows only one true developmental servation to conservation of length in the three different potential sources of uneven- not reading skills; another will develop bc
sequence. one set of stages of a fixed num- gadget, and thus he will have developed the ness: practice, task, and content. All three basket-making skills and reading skills, t
ber (e.g., Kohlberg, 1969). Skill theory skill for conservation of length without sources produced unevenness. As de- not skills for drawing maps.
predicts, to the contrary, that the develop- ever having gone through Level 6 for that scribed earlier, the difference between the Second, different individuals will follc
mental sequence that a person progresses particular skill. He will have effectively longitudinal and cross-sectional groups different developmental paths in the sal
through will vary depending upon the as- skipped Level 6 in the developmental se- showed strong unevenness due to practice: skill domain (Fischer & Corrigan, in pres
sessment tasks and procedures used, as well quence for conservation of length. two to three steps in an eight-step sequence. The developmental transformation ru
as analogous environmental factors that Many of these irregularities and varia- Similarly, the specific task used to assess predict a large number of different possit
occur naturally, outside the experimental tions in developmental sequences will be re- obj,ect permanence created substantial un- paths in any single domain. The sprir
context (Fischer & Corrigan, in press). duced or eliminated by repeated testing with evenness: two steps in the eight-step se- and-cord gadget illustrates how individu;
The variation in sequences as a function similar tasks. Suppose, for example, that an quence. Finally, the content (the type of can take different paths within the sat
of testing is especially obvious for micro- 8-year-old child has many other Level 6 stimulus searched for) often produced small domain. The way that an individual mo%
developmental sequences. The develop- skills but has not been induced to develop but reliable unevenness, especially with the from Level 6 ski!ls for the gadget to a Leve
mental transformation rules can be used to conservation of length. Exposure to the task cross-sectional procedure: Both the type of skill integrating all four variables of t
predict a large number of microdevelop- for conservation of length with the gadget object and the familiarity of the object pro- gadget (weight, length of spring, vertic
mental steps. For example, use of the sub- will normally induce him to develop con- duced unevenness ranging up to one step in length of cord, and horizontal length) u
stitution rule on Steps 2, 3. and 4 in Table 4 servation of length (Hooper, Goldman, the eight-step sequence. vary depending upon the particular Leve
would have produced six microdevelop- Storck, & Burke, 1971). Because of effects skills that he combines. The results of ts
mental steps instead of three. Yet if chil- like this, performance in later testing ses- Individual Differences possible alternative paths are shown
dren are not exposed to the specific tasks sions will commonly fit a sequence better Table 5. In the first path, an individL
corresponding to each predicted step, many than performance in the initial session Just as environmental factors make un- begins with two Level 6 skills: the syste
of the steps will not appear in their behavior. (Tucker, 1979). evenness the rule within an individual, so relating weight, 6W, to the length of t
If their environment never induces the use The effects of specific tasks and testing they ensure that different individuals will spring, 6L,and the system for conservatio
of a substitute object, for example, they procedures may explain many of the dis- show different patterns of cognitive de- relating total length of the cord at two d
will never show these three new substitution agreements in the developmental literature velopment. Of course, hereditary factors ferent times, ;CX,"and $Cv,H.As shown
steps for Table 4, nor any of the other pos- about sequences in a given skill domain. also contribute to individual differences in Table 5 , the individual forms a Level 7 sk
sible substitution steps in the development For example, different investigators, using development (as well as to unevenness); for the entire gadget by intercoordinatii
of social role playing. Likewise, steps in- different procedures, have found different but even without those hereditary differ- these two Level 6 skills.
514 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5

In the second path, a different individual allows direct access to the concept of con- levels in the same domains. The environ- usually by 7 or 8 years of =e, that t
begins w& a BIffereftt &Level Q servatitw of &e W k@ of the d, meatzd diversity of kttRtaR exec, its t"to€ehy d o a l l o t chmgewhen a b
systems, each involving weight fiW:weight whereas the second skill requires that the well as the genetic diversity of the human of clay is elongated into a sausage, flatten
and length of the spring, %; weight and conservation be inferred from the coordina- species, ensures the occurrence of major into a pancake, or changed into some 0th
vertical length of the cord, 6Cv;and weight tion of two representational systems. individual differences in development. shape. At the same age, however, they si
and horizontal length of the cord, TH. He On the other hand, the two final Level 7 Skill theory thus makes several general believe that the weight of the clay ball do
too combines these skills to form a Level 7 skills are equivalent for most purposes. predictions about the effects of environ- change when the shape changes. TypicaU
skill for the entire gadget; but to do so, he Both of them interrelate the same four vari- mental factors on sequence and synchrony they will not develop the skill for consen
must go through more developmental trans- ables, and both of them reflect accurately in development and provides tools for ana- tion of weight until 9 or 10 years of ag
formations, as shown in Table 5, and he the relations among the variables in the real lyzing some of these effects. In addition, Within Piaget's framework, this consiste
ends up with a different skill from the in- gadget. Individuals using the two skills will the structures defined by the theory suggest sequence is puzzling because both types
dividual who followed the first path. come to mostly the same conclusions about a number of general corollaries about struc- conservation are said to require exactly t
The first path is more efficient than the the variables in the gadget. tural relations and how they determine se- same kind of concrete operational schem
second one: It requires fewer transforma- Similarly, for virtually every skill at every quence and synchrony. Two factors (height and length) covary
tions, and the final skill (Step 2 of Path 1 one of the levels, different individuals can such a way that changes in one compensa
in Table 5) relates the four variables to- take different developmental paths within a Structural Corollaries for changes in the other.
gether without redundancy. The second skill domain, and usually the end products According to skill theory, conservation
path not only goes through more trans- of the different paths will be skills that are I shall not attempt to provide an ex- weight develops after conservation of su
formations but also produces a skill (Step 3 equivalent for most purposes. That is not to haustive list of structural corollaries but in- stance because it requires a compoundc
of Path 2 in Table 5) that is full of redun- say, however, that individual differences stead will present a few illustrations of Level 6 skill that subsumes the skill for co
dancy, with the weight variable reappearing are minimal. The different paths within a potentially useful ones. servation of substance. In conservation
in every representational system. The skills domain are often significant; and more im- substance, the child must coordinate tl
also differ somewhat in the behaviors that portant, individuals normally develop in Consistent Decalage Within a Task Domain length and width of the original piece
they control. For example, the first skill different skill domains and to different skill clay, fi3L,w, with the length and width of ti
Unevenness in skills across domains transformed piece, !&,w (Halford, 197
Table 5 seems to be a fact of development. But Peill, 1975; Verge & Bogartz, 1978).15 .
Alternative Developmental Paths IO a Level 7 Skill for All Four Variables in the Gadget according to the theory, many phenomena conservation of weight, on the other ham
that are commonly classified bs instances of the child must go beyond mere amount I
Cognitive level Path 1 Path Za unevenness are in fact microdevelopmental clay and think about weight of clay. That i
Step 1: [6W f-, 2 1 , [sW f-, 6Cvl:and
sequences: The unevenness follows the he must relate the changes in the length ar
6 Representational Step 1: [6W f-, 2 1 and
systems W Y , H f-, f G . H l CRW f-, 6cki1
same pattern in virtually all children in a width of the clay to a third factor, such i
given social group, and it seems actually to the weight readings on a scale or the amoui
7: Systems of repre- Transformation: Transformation: arise from differences in the complexity of of force that he feels when he holds the c k
sentational systems [W f-, 6L].[yCy.Hf-, ~ C , . , ]= Step 2 [W f-, 6.L].[6W * T,] = Step 2a the skills. Most of the instances of horizon- in his hand, V .To coordinate all three se
Step 2: Step Za: 'W

LWJ 1
Transformation:
f-, 'L
TI.
tal decalage (unevenness within a stage or together in a single skill, he must compoun
period) studied by Piaget and his colleagues the skill for conservation of the substanc
show such microdevelopmental sequences. clay with a skill involving the weight of tf
The skill theory explanation is simplest in clay, such as the skill in which the chil
[6W f-, uC'c,].[6W f-, 'CHI
= Step 2b
cases where the skills belong to the same relates the length of pieces of clay (fc
Step Zb: 'W ++ 'Cy task domain. The skill that develops later instance, sausage-shaped pieces) to the
[ J
rw
can be derived by the transformation rules
from the skill that develops earlier. l4 Note that this type of conservation is proper
Transformation: Among the best documented cases of called conservation of substance. It has often bet
consistent decalage within a task domain is erroneously translated as Conservation of matter

1
t

= Step 3 the development of conservation of sub- conservation of volume. Both matter and volume a
stance and conservation of weight. Re- much more abstract and difficult concepts than amou
of a substance such as clay, and they develop
search has repeatedly shown that school later ages.
children develop conservation of substance A precise measure of the amount of clay, t

1 to 3 years before conservation of weight course, requires three dimensions (height, length, ar
(e.g., Hooperet al., 1971; Piaget & Inhelder, thickness), notjust two; but with early Level6 skdls tt
child does not yet understand true volume-that i
1941/1974, especially in the introduction to three-dimensional volume. His understanding c
a Steps 2a and 2b form separate skills, but they do not develop in sequence with respect to each other.
the 2nd edition; Uzgiris, 1964). In conserva- amount of clay is based on a relatively crude coordin;
according to this skill analysis. tion of substance,14 children understand, tion of just two dimensions.
516 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT

Table 6 the doll, 'G, movinn it, 2M, and looking at that it mimics. For example, B~yanti
A Few Exelfffllestrf M-kkg It, 2f:When he hqrpens to grasp the do€t, Trabasso (€97€> carefully trained presch
he can move it in front of his eyes and look children to correctly judge the larger
Actual
cognitive levels Mimicking skill at Level L Mimicked skill at Level L + I at it. He also has the related mapping of every pair of sticks in a five-stick seri
looking, moving, and grasping: When he When the children were asked about nt
Levels 2 and 3 PS, -'M, -'G2 -*MI-2Sil TS f-, 3M f,'G] happens to look at the doll, he can move his adjacent parts from the series without be
Levels 2 and 3 [?S2 -?M -2G > PG, -' M i -*Si1
P 21 [3S++ e hand to where he sees it and then grasp it. shown the specific lengths again, many
By compounding these two skills, he can them correctly inferred the longer stick
Levels 6 and 7 [4uf-, 6N f-, 6Pf,'QI construct a complex Level 2 skill that mimics thus apparently demonstrating transitive
the Level 3 skill, as follows: ference, which for their series of stic
would seem to be a Level 6 skill. Hc
pS2 - 2M2 - 2G2] + ?GI - 'MI - 'SI] ever, the training procedure was perfec
weight: mimicking skill at Level L is by no means = [2S2 - 'MZ - ' 6 2 - 'MI - 'Si]. (24) designed to teach a compounded Leve
skill that would mimick the Level 6 si
identical with the mimicked skill at Level With this mimicking skill, the child can for transitive inference.
L + 1. In general, the skill at Level L + 1 demonstrate a complexity and recombina- If the children had only been taugh
[!BL,we P L tf
, ~6F1- (23) will be much more flexible and differentiated tion in his actions that mimics the com- mimicking skill, their correct performar
than the skill at Level L, and the child will plexity and recombination of Level 3 . When should have been limited. For examp
Consequently, the child will develop con- have much better control over the relations he happens to look at the doll, he can move they should not have been able to solv~
servation of the weight of clay after con- among sets. But there will still be many
servation of the substance of clay. his hand to it, grasp it, move his hand in transitivity problem that required them
similarities between the mimicking skill and front of his body, and look at the doll there. organize the needed information about nc
This same kind of analysis should be able the higher-level skill.
to explain most cases of consistent decalage The sequence of actions is thus physically adjacent sticks on their own. In a follow-
An example from the sensory-motor reversed and looks superficially like the re- study, Bryant (1974, pp. 54-56; 1977) fou
within a skill domain, including cases where
tier will illustrate how mimicking occurs. combination of looking, moving, and grasp- exactly that: Children who after train
the differences in complexity are not ob-
By compounding Level 2 mappings, the ing that occurs so fluidly in the Level 3 skill. could consistently solve the original trar
vious, as when differences in stimulus sali-
child can mimic the flexibility and com- But in fact it is still only a chaining of actions. tivity problems could not solve similar n
ence produce decalage (Odom, 1978;
plexity of a Level 3 system, as shown in The infant can carry out the sequence, but transitivity problems that required them
Fischer & Roberts, Note 3). The skills in
each case actually differ in complexity, but Table 6. Consider the actions of grasping a he cannot reorder it into the many flexible seek out the needed information on th
psychologists have previously categorized doll, G, looking at the doll, S, and moving combinations that typify Level 3 . own. According to the mimicking corolla
them as showing unevenness simply be- the arm, M. When a child has a Level 3 A more primitive form of mimicking can all instances of mimicking should sha
cause there has been no tool for analyzing system controlling all three of these actions, also occur in this situation, as shown in similar kinds of limitations.
the skills and thus recognizing the dif- he can combine several aspects of each of Table 6. When the child has only the two When one uses skill theory to anal)
ferences in complexity. the three actions in a great variety of ways. simpler Level 2 skills shown in Equation 24, behaviors that Piaget has studied, it
For example, he can look at the doll and the stimulus context can lead him to change important to be aware of mimicking, I

use what he sees to guide the movement of focus from the first skill to the second: He pecially for many of his infant observatioi
Mimicking happens to look at the doll, moves his hand Most cases of primary, secondary, a
his arms to grasp the doll, and then once he
Besides explaining phenomena like the has grasped it, he can move it in front of his to where he sees it, and grasps it. As he tertiary circular reactions, for instanc
lag between the development of conserva- face and visually examine it. More generally. holds it in his hand, he then loses sight of seem at first to require sensory-mol
tion of substance and conservation of he can carry out plans that require him to it and so changes focus to the second skill: skills at Levels 1, 2, and 3, respective
weight, microdevelopmental transforma- consider the relations among several as- He maintains his grasp on the doll, moves But closer examination shows that ma
tions also predict another phenomenon: pects of all three actions simultaneously. his hand in front of his face, and looks. of these reactions are probably compl
mimicking, in which a complex skill or He can use his looking to guide his moving Thus, the context produces behavior that skills at the previous levels.
series of skills at Level L produces behavior all along the path of movement; he can place superficially appears to show the mimicking Mimicking is not just a laboratory curic
that seems at first to require a skill at Level the doll at any point in space within his Level 2 skill or the mimicked Level 3 skill, ity or a measurement problem, howevc
L + 1. reach; he can remember where he saw the but the child cannot actually control either of It occurs normally when the child cc
A person can mimic a skill at Level L + 1 doll a few seconds before and reach there these complex skills. structs transitional steps in the spontar
by acquiring a complex skill or series of to grasp it. And he can do all these complex Mimicking has been produced in the ous development of a skill. Indeed, mimic
skills a t Level L that includes all the sets things smoothly and planfully, without trial laboratory by a number of ingenious experi- ing skills probably lay the foundation i
that comprise the higher-level skill. The and error. mental psychologists (e.g., Case, 1974; the child's development to the next leve
mimicking skill will usually result from the At Level 2, the infant can mimic this Harris & Bassett, 1975; Siegler, 1976),
transformations of compounding or focus- Level 3 system by compounding the three and some of these studies nicely support Parallels Between Tiers
ing (or both), as illustrated in Table 6. I use actions (Table 6). First suppose that he has the skill-theory argument that the mimicking A third corollary involves the relatio
the word mimic intentionally because the a sensory-motor mapping relating grasping skill is different from the higher-level skill between tiers. Because the general Level
 5 18 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5

&3m(W&Espeat at gverytier,Bekrurisr Jkkux testssf the gredicted paEiild k- men?: The infantile reflexes seem to be directly to the prediction of a reflex-tier
should show structural paraIIeIs between tween LeveIs I to 4 and Levek 4 to 7 are f e .3”&es fer the ii&d units as+* by €Mkgef(B??,Nete 8). m l
3
tiers. If a specific developmental sequence clearly necessary, but besides generating from which skills are constructed. Un- the large quantity of other research on tl
occurs at the sensory-motor tier, for tests of skill theory, the parallel serves fortunately, almost no research has been newborn (Haith & Campos, 1977) could 1
example, then in the proper environment a another important function in research. It done that can be used to test the existence interpreted in terms of such a tier, but no]
similar sequence should appear at the repre- offers a source for new hypotheses. For of these levels, and consequently I have of it seems to provide a direct test of tl
sentational tier, but it should of course in- every phenomenon that is discovered in treated this tier as a corollary rather than predicted four levels of reflex developmer
volve changes in the structure of representa- sensory- motor development, a similar as a more firmly established part of the Bullinger describes how the tonic neck r
tions rather than changes in the structure of phenomenon can be searched for in repre- theory. flex becomes gradually coordinated with tl
sensory- motor actions. sentational development, and vice versa. The infant or fetus begins with single looking reflex and eventually develops intc
I know of only two sets of studies that Likewise, developments at the sensory- reflexes, combines the single reflexes into looking skill that is independent of the ton
provide data relating to precise parallels motor and representational tiers suggest reflex mappings, then combines the map- neck reflex.
between tiers. One shows a parallel in the similar developments at the abstract tier. pings into reflex systems, and finally com- In a sense, there are really two ton
representational tier to a sensory-motor Other investigators have proposed a bines the reflex systems to form systems of neck reflexes. In one, the infant turns k
sequence; conversely, the other shows a general parallel between sensory-motor reflex systems, which are single sensory- head to the right and raises his left han
parallel in the sensory-motor tier to a development and later development (Piaget, motor sets (Level 1 in Table 3). and in the other, he turns his head to tl
representational sequence. 1937/1954, 1967/1971; Mounoud, 1976; The term reflex is used in a number of left and raises his right hand. The ton
In the sensory-motor tier, the infant Siege1 & White, 1975; Werner, 1948). Piaget different ways in the psychological litera- neck reflexes and various looking reflex
develops skills for finding hidden objects. (1941) even gave a special name to parallels ture. Some psychologists reserve the term show a significant physical dependenc
By Level I11 of the sensory-motor tier, across his developmental periods: vertical for behaviors that are not subject to operant When the young infant is producing a givt
he can follow the visible displacements of decalages (distinguished from horizontal control and that are often assumed to be tonic neck reflex, he can look only to tl
an object and look for it where it last disap- decalages, which are “parallel” develop- controlled by the peripheral nervous sys- side of his midline where his head is turne
peared (see Piaget, 1937/1954, Stage 5 in ments within the same period). Greenfield tem, like the knee-jerk reflex. I use the term For example, when his head is turned to tl
chapter 1). Because of the parallel between and her colleagues have searched for struc- instead in the sense that it is used by ethol- right, he can look at stimuli within t.
tiers, a similar skill for search should de- tural parallels between language and manip- ogists (e.g., Hinde, 1970) and many psychol- visual field to the right of his midline, but 1
velop at Level 111 of the representational ulative play (Goodson & Greenfield, 1975; ogists (e.g., Piaget, 1936/1952): It refers to cannot look at stimuli to the left of midlin
tier. Drozdal and Flavell (197.5) have de- Greenfield, Saltzman, & Nelson, 1972; what might be called preprogramed be- To look at stimuli to the left, he must pr
scribed exactly such a behavior, which they Greenfield & Schneider, 1977). None of havior- species-specific activities that duce the other tonic neck reflex, in whit
call “logical search behavior.” By 7 or 8 these investigators, however, has provided seem to be biologically programed into the his head is turned to the left. Bulling
years of age, most children could represent a system for analyzing and predicting the nervous system (Teitelbaum, 1977). For describes how the infant gains control
the probable displacements of a lost object parallel structures, and consequently it has example, Zelazo, Zelazo, and Kolb (1972) this relation. My description of Bullinger
and so look for it where it had probably been impossible to test the validity of sug- have worked with the stepping reflex, a results includes an interpretation in terms
been lost. (See also Wellman, Somerville, gested parallels. Skill theory, with its sys- complex response that can be elicited in the the four reflex levels.
& Haake, 1979.) tem for analyzing the structure of skills, newborn infant and that seems to be or- At Level I, single reflex sets, the infa
Mounoud and Bower (1974/197.5) report a may allow more precise tests of proposed ganically related to the voluntarily con- produces single reflexes, like each of tl
parallel in the opposite direction: from the structural parallels. trolled walking that develops toward the tonic neck reflexes and each of the varioi
representational tier to the sensory-motor Besides specific structural parallels be- end of the first year after birth. Other ex- looking reflexes; but he cannot contr
tier. At Level I11 of the sensory-motor tween tiers, skill theory also predicts new amples would be the sucking reflex, which is any relations between reflexes. Bulling
tier, infants developed a skill that was ap- tiers, because Level IV of each tier pro- elicited by stimulation of the lips, and the found that infants from 1.5 to 45 days of a!
parently parallel to the skill for conservation duces a new kind of set. Both before and tonic neck reflex, in which the infant turns produced the tonic neck reflexes and varior
of weight at Level I11of the representational after the three specified tiers, the cycle of his head to one side and raises his arm on looking reflexes, but usually could not co
tier: When a familiar object made of a four levels can occur again with different the opposite side. Even complex behaviors trol any relation between tonic neck reflc
malleable substance like clay was altered types of sets. There must, of course, be like looking are reflexes within this meaning: and looking.
from its usual shape, the infants grasped it as some limit on the recurrence of the cycle, The sophisticated rules for visual scanning At Level 11, rejlex mappings, the infa
if its weight remained the same, even though since it cannot go on infinitely; but that described by Haith (1978) seem to be pre- maps one reflex onto another and thi
in other situations they routinely adjusted limit will have to be determined by future programed properties of the looking reflex begins to control relations between reflexe
their grasp to fit the differing weights of research. or reflexes. To distinguish these reflex be- For example, he should be able to produc
various objects. That is, they seemed to haviors from peripherally controlled re- the head-right tonic neck reflex in order
assume that the familiar object’s weight re- flexes like the knee jerk, I will call them look at a stimulus to his right. At Level I1
Rejlex Tier reflex skills or sets, because skill theory
mained the same even though its shape had reflex systems, the infant relates two ma
changed. This sensory-motor conserva- The tier before the sensory-motor tier predicts that they normally develop into pings to each other, integrating the tu
tion was not present in young infants and could be called the reflex tier and might well sensory-motor skills. tonic neck reflexes with the two lookii
emerged by about 18 months of age. provide the starting point for skill develop- I know of only one study that relates reflexes (left and right) in a reflex systen
 520 KURT W FISCHER THEORY OF COGNITIVE DEVELOPMENT 5:

H e shovldlherefore be able to shift from structural corollaries, as well as other Th_e study by McCall et al. (1977) on shifts in patterns of skills with developmer
one tonic neck reflex to the other as neces- generaf effects of the environment on skiff shifts m the pfh of irtfafft skitff shews Tise g e € k e F d skiks aE@w8Re to flredi
sary to look anywhere within his left and development and, of course, many other one method for inferring such transition not only broad statistical changes but al:
right visual fields. Bullinger describes the specific developmental sequences and syn- points. These researchers found instabilities many other general skill patterns, such
development of control by the infant over chronies. Rather than enumerating more in the correlation patterns of infant tests the probability of possession of a s p e d
the relation between the tonic neck reflexes such predictions, however, I would like to that correspond generally to what is pre- skill in all people in a large, cultural
and the looking reflexes, but he does not turn to some general implications of skill dicted by skill theory. When a shift to a new homogeneous population. One can predic
discriminate between the predicted Level I1 theory for conceptions of cognition, learn- optimal level occurs, an increased uneven- for example, the average age at whii
and Level 111 skills. Infants usually showed ing, and development. ness in the levels of performance will virtually all children of a given culture w
some control of relations between the two appear in the individual child. The reason have attained a specific level of a skill th
types of reflexes at 45 to 80 days of age. A Few Implications of the Theory for this greater unevenness is that the speed is important for that culture: Virtually
At Level IV, systems of reflex systems, of increase in optimal level becomes larger American middle-class children will ha
the infant coordinates two Level I11 sys- Any theory worth its salt should do more attained an understanding of the social rc
at these times and the child can initially
tems into a higher-order system and thus than answer the original questions it was of doctor (Level 5: Step 2 in Table 4) 1
apply this new capacity to only a few skill
generates a single sensory- motor set devised to answer. It should have implica- 5 years of age. One can specify the n o m
domains. Consequently, many correlations
(sensory-motor Level I). He should be tions for other important questions. Several range in which American middle-cia
across domains decrease. The periods of
able, for example, to relate the tonic-neck- of the most interesting implications of skill children will normally be moving onto a nr
theory involve central topics in cognitive correlational instability thus reflect times of
reflex-and-looking system with another maximal change. McCall e t al. found four cognitive level for skills that are importa
reflex system involving posture and looking, psychology: the nature of the big picture of to them, as shown in Table 7. Also, te:
such periods of instability during the first
thus showing highly flexible looking be- cognitive development, the analysis of can be made of the levels predicted by sk
2 years of life, exactly as is predicted from
havior that is relatively independent of cognitive development and learning across theory versus those predicted by 0th
the four sensory- motor levels. (They
specific postures: He has generated a new skills, and the relation between behavior theories (e.g., Bickhard, 1978; Case, 19;
found these periods of instability before
kind of set, the single sensory-motor and thought.
they knew about skill theory.) Halford & Wilson, 1980; Isaac & O’Connc
action of looking. Bullinger found such Presumably, similar instabilities could 1975; Mounoud, 1980; Mounoud & Haue
flexible looking behavior commonly in
The Big Picture of Development
be found for all the higher levels as well. Note 9).
infants 80 to 120 days old. Skill theory emphasizes careful analyses For example, Kuhn (1976) finds instabilities
In this way, development through the of specific tasks and predictions of specific in ability-test correlations in early adoles- Application to Other Skill Domains
reflex tier produces a single sensory-motor sequences and synchronies in circum- cence, when people are presumably moving
set. Note, however, that such a set involves scribed task domains. But it also goes be- to optimal Level 7, single abstractions. As the social-role example implies, t
not only one reflex system but two or more, yond these specifics to predict the general Epstein (1974a, 1974b, 1978) reports spurts “big picture” to which skill theory appli
because a Level IV skill involves the co- nature of major shifts in cognitive develop- in mental age and brain growth that seem to is not limited to the standard cognith
ordination of at least two Level 111 systems. ment-how skills are changing across the correspond with the emergence of Levels 5 , developmental tasks (mostly Piagetian tas
In the Bullinger example, the child co- board as the person develops. 6, 7, and 8. and IQ-type tasks). It has the promise
ordinates the tonic-neck-reflex system with Although particular skills do not show There is a difficulty, however, with using applicability across many different SI
another postural system in such a way that abrupt or discontinuous change, major age as the dimension along which one looks domains and consequently the potential 1
the postural adjustments go almost un- statistical shifts in populations of skills do for instability. After infancy, developmental integrating theoretical analyses in areas tl
noticed, but in other cases the two systems occur (Feldman & Toulmin, 1975). In skill canalization decreases (McCall, 1979; have usually been treated as theoretica
are more obvious. For example, an infant theory, the child’s optimal level increases Scarr-Salapatek, 1976), and consequently distinct. Skill theory may be applicable
can coordinate a reflex system for sucking with age, and the speed of the increase is people probably no longer change to a new areas as diverse as language developme
with a reflex system for looking, and thereby Faster when the child is moving into a new optimal level at the same approximate age. social development, and learning.
he can look while he is sucking. This kind level (Fischer & Bullock, in press; Fischer, This variability in the age of shifting should The skill levels should apply to any sk
of analysis can provide a mechanism for Note 7). Together with environmental in- increase dramatically at higher levels. Also, that develop, since they characterize 1
predicting and explaining the composition duction, these spurts at each level will at higher levels, the prevalence of uneven- general information-processing system
of sensory-motor sets, especially the types produce major changes in the profile of skill ness within an individual should become human beings. Applying the theory to a n ,
of co-occurring behaviors that can be levels. Transition periods between “stages” much greater. This problem with age can skill domain will not be an easy matter,
globally combined in the single, poorly dif- can therefore be defined as times when an be eliminated if good measures of skill course, because it will require careful
ferentiated sensory-motor sets described increase in optimal level is producing a levels are used. Then people can be grouped scriptive analysis of the specific skills tl
earlier. major shift in the population of skills, with not by age but by their optimal level, and develop in that domain. This kind of care
Skill theory produces, then, at least these many skills gradually moving to the new the distribution of optimal levels within a analytic research has only recently beco
four structural corollaries: the reflex tier, optimal level. To the extent that the new sample will demonstrate whether spurts and c o m m o n in cognitive-developmen
parallels between tiers, mimicking, and con- optimal level applies broadly across a wide instabilities exist (Fischer & Bullock, in psychology.
sistent decalage within a task domain. The range of skills, the shift in the skill profile press; Fischer, Note 7). The first step in applying skill theory
theory should also be able to predict other should be dramatic and easy to detect. Skill theory thus predicts general types of new spheres such as language develc
 522 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 5:

Table 7 lem solving. These chzinxes should be pre- action, since thought is literally built fro
Age Periods ctf Which € e d s First O~velop~ ctictabfe by the micrmk&pmenta€ tfm- settsory-f&&F &kills. A b , s e n s a 3
formation rules of the theory. For example, motor development does not cease at tl
Cognitive level Age perioda end of the sensory-motor tier but co
in the microdevelopmental sequence in
I: Single sensory-motor sets Several months after birth which children pretend about going to sleep, tinues at higher levels.16
2: Sensory-motor mappings Middle of first year the successive steps in the sequence are Representational and abstract skills pr
3: Sensory-motor systems End of first year and start of second year essentially steps in the generalization of duce and direct sensory-motor action
4: Systems of sensory-motor systems, which are single
representational sets an action: Children pretend to go to sleep, This relation between representation ai
Early preschool years
5: Representational mappings Late preschool years then pretend to put a doll t o sleep, then action is illustrated by the example of tl
6: Representational systems Grade school years pretend to put a block to sleep, and so forth child’s understanding of the spring-and-co
7: Systems of representational systems, which are single (Watson & Fischer, 1977). Similarly, many gadget at Level 5. When the child undt
abstract sets Early high school years stands the mapping of weight (represenl
8: Abstract mappings Late high school years
microdevelopmental sequences typically
9 Abstract systems Early adulthood categorized under cognitive development tional set W ) onto the length of the spri
1 0 Systems of abstract systems Early adulthood could equally well be categorized under (representational set %), her control
~~
learning or problem solving (e.g., Fischer each representational set is based on se
=These periods are merely estimates for middle-class Americans. For Levels 9 and 10, existing data do & Roberts, Note 3). sory-motor sets. With her Level 5 skl
not allow accurate estimation.
Likewise, adults solving a complex she can therefore directly control t
problem or rats learning to run a maze show various weights to manipulate the length
ment or social development must therefore only thing special about these Piagetian systematic changes in the organization of the spring. She is not left sitting in a corn
be an analysis of some of the specific cognitive skills is that their development their behavior (Duncker, 1935/1945; merely thinking about how weight relat
skills that develop in language and in social was investigated first- before the develop- Fischer, 1975; Siege1 & White, 1975). These to length. Behaviors studied in our lab01
relationships (see, e.g., Harter, 1977). ment of the language skills or social skills changes can be treated as microdevelop- tory also illustrate this relationship betwe
Starting with these specific skills, the theory that they are supposed to explain. Interac- mental sequences, and therefore skill theory representational and sensory-motor st
can be used to predict how they will de- tions between some Piagetian skills and should be able to predict and explain them (Bertenthal & Fischer, 1978; Watson
velop through the skill levels, as was demon- some language skills or some social skills (Fischer, 1974, 1980). Fischer, 1977, 1980; Fischer & Rober
strated earlier by the prediction of a de- will undoubtedly occur in development, Skill theory, then, may help to integrate Note 3).
velopmental sequence for social-role skills but they will be highly specific interactions, such apparently diverse research areas as The inclusion of sensory-motor ski
(Table 4). not general relationships in which one type learning, problem solving, social develop- in representational skills is especially e
Notice that language skills, social skills, of skill will be a general prerequisite for ment, language development, and cognitive dent in language. Speech and gesture, whi
and skills in Piagetian tasks are all “equal” the other. And interactions will occur in development. It also has important implica- are both sensory-motor skills, are essent
in skill theory (as they are in the approach both directions, not just from Piagetian tions for another major research problem- components of the representational skills
of Vygotsky, 1962). Many recent ap- skills to language or social skills, but also the relation between behavior and thought. language (e.g., Fischer & Corrigan,
proaches to language development and vice versa. The earlier discussion of syn- press; MacWhinney, 1977).
social development have postulated that cog- chrony explained the kinds of relationships Behavior and Thought In addition, the control of sensory-mo
nitive skills are somehow more fundamental that should be expected: (a) a low general skills by representational skills extends 1
than language skills or social skills. For synchrony across domains, (b) high general A classic problem for most cognitive yond the direction of sensory-motor sk
example, the development of some Piagetian synchrony only when the skills in the approaches has been that their constructs that are already present. Higher-level sk
measure of cognitive development, such as specific domains being tested are all main- typically d o not explain how thought is also direct the acquisition of new low
object permanence, is hypothesized to be tained at the children’s optimal level, and turned into action (see Hebb, 1974). AS level skills. Jacqueline’s “bimbam” sk
the one prerequisite for the appearance of (c) specific interactions only when a par- some wit said, they leave the organism described earlier, provides an exam]
language (see Corrigan, 1979; Fischer & ticular skill in one domain becomes a com- sitting in a corner thinking. (Piaget, 1946/1951, Observation 64). Wh
Corrigan, in press). Similarly, researchers ponent of a particular skill in the other Skill theory provides a possible way out she first combined two Level 3 sensor
in social development use conservation or domain. Note that the kind of specific of this dilemma. Thought (representation motor systems into the Level 4 bimb.
some other Piagetian measure to explain the interaction to be expected is what be- and abstraction) develops out of behavior representation for fluttering, her skill cc
emergence ofimportant social skills, such as havioral analyses of transfer have always (sensory-motor action), and the skills of trolled just two things that fluttered: hersc
perspective-taking and morality. The predicted: Specific components of one skill thought hierarchically incorporate the skills when she rocked back and forth on a pit
Piagetian skill is again elevated to a special become components of a second skill of action that they have developed from. of wood, and leaves, when she made th
status, as if it were more fundamental (e.g., Baron, 1973; Mandler, 1962; Reed, That is, representational skills are actually
than the social skills. Emst, & Banerji, 1974). composed of sensory-motor skills; and
l6 In Piaget’s theory, the nature of the relation
According to skill theory, there is nothing In addition to large-scale developmental likewise, abstract skills are actually com-
tween sensory-motor action and representation is
particularly fundamental about object changes, skill theory is also applicable to posed of representational skills and there- clear, but it seems that sensory-motor developn
permanence, conservation, or any other changes in behavioral organization that are fore sensory-motor skills. Consequently, stops at the end of the sensory-motor period fc
Piagetian measure of development. The usually categorized under learning or prob- there is no separation between thought and Piaget, 1946/1951, p. 75).
 524 KURT W . I FISCHER THEORY OF COGNITIVE DEVELOPMENT 52:

flutter. Then, through compounding and at this time it does not deal adequagly r e h t k n between recall success and skill competence versm performance, AlthOUBl
s&k&km, she extanled the skiH to new with skill bo&. kvef (see Watson & Fisctrer, W77). €tdoes there is s m e overtap between the twr
objects, such as curtains, that she could A task domain involves a series of tasks not specify, however, how the process of issues, they are not the same. The extremi
make flutter or that fluttered in the breeze. that are all very similar to each other, accessing skills relates to individual differ- formulation of the competence-perform
For each object to which she extended the typically sharing a basic group of com- ences and task differences in memory per- ance model assumes that a structure i:
skill, she constructed or included a new ponents but differing in the additional com- formance. present but that there is some perfoi-manct
Level 3 sensory-motor system involving ponents that are required to perform the Skill theory in its present formulation limitation that prevents it from being full!
the fluttering of the new object, and this tasks. A skill domain, on the other hand, does not use the information-processing realized in behavior (Chomsky, 1965). Thr
skill thus became a new sensory-motor involves a number of task domains that framework. It is a structural theory that access question, on the other hand, entail:
Zomponent of the Level 4 “bimbam” skill. share similar skills and therefore develop has its roots in the classical tradition of no such assumption, because skill theor:
In the same way, representational skills in approximate synchrony. cognitive psychology (see Catania, 1973; does not posit powerful structures that havr
at higher levels are constantly used to con- At present, skill theory determines skill Fischer, 1975). In recent years many psy- difficulty eventuating in behavior. Thl
struct new sensory-motor skills. Develop- domains in a primarily empirical way. When chologists have come to equate cognitive access question is simply: What are thr
ment from Levels 4 to 7 produces skills developments in two task domains show a psychology with the information-processing processes that determine which skill ai
that subsume more and more sensory- degree of synchrony that cannot be ac- approach. This equation ignores the fact individual will use in a particular task at :
motor actions and at the same time control counted for by environmental factors such that a long and venerable tradition of cog- given moment?
finer and finer differentiations of sensory- as practice effects, then the two task do- nitive psychology existed decades before
motor actions. Consequently, skill theory mains are said to belong to the same skill the information-processing approach was Concluding Comment
should be able to predict the development domain. To deal with skill domains in a more invented.
of complex sensory- motor skills like satisfactory way, skill theory will ultimately On the other hand, skill theory is not Whatever their form, theories are tool:
jriving a car, using a lathe, or operating require concepts for specifying the glues inconsistent with the information-pro- for thought (Hanson, 1961). The essentia
a balance scale-skills that develop after that tie task domains together. These con- cessing approach. Indeed, I would hope that test of a theory is whether it is a good tool
the first 2 years of life. Research does sup- cepts will presumably lead to a graduated some parts of it could be reformulated in This theory is intended to be a useful too
port the argument that orderly develop- notion of skill domain rather than an all- information-processing terms. Such a for- for understanding cognitive developmen
mental changes occur in sensory-motor or-none notion: Task domains will vary in mulation might provide more precision in and facilitating the process of theoretica
skills during both childhood (e.g., Green- terms of the proportions of skills that they some parts of the theory and thereby help integration that is essential to progress ii
field & Schneider, 1977; Ninio & Lieblich, share. to overcome some of the theory’s limita- psychology (Elkind & Sameroff, 1970
1976) and adulthood (e.g., Hatano, Miyake, tions, including the treatment of accessing Haith & Campos, 1977). The theory prom
& Binks, 1977). skills. ises to provide a system for predicting an(
In addition to making numerous specific Accessing Skills explaining developmental sequences an(
Any attempt to provide an information-
developmental predictions, then, skill The second limitation involves a matter processing formulation, however, should synchronies in any skill domain throughou
theory has significant implications for the that skill theory says little about. No pro- avoid a major pitfall that has plagued many the life span, and it also promises to inte
nature of changes in populations of skills in cesses are designated to deal explicitly with information-processing analyses of cog- grate analyses of development with treat
development, the integration of theoretical the way in which skills are accessed. A nitive development: They neglect the adap- ments of learning and problem solving
malyses of skill development and learning person may have available the skill needed tive process that is the very basis of cog- Time and research will tell u9hether thi:
in spheres that have been traditionally to perform a particular task or to show a nition according to skill theory. The cog- promise becomes fact.
treated as distinct, and the relation between specific behavior and yet in the appropriate nitive organism is constantly adapting skills
behavior and thought. But skill theory also context may fail to use that skill. Skill to the world, and this adaptation provides Reference Notes
has several limitations. theory does not deal directly with phe- the foundation for cognitive development 1. Pascual-Leone, J. A theory of constructiv,
nomena of this type, which are commonly and learning (see MacWhinney, 1978). Any operators, a neo-Piagetian model of conservation
Limitations of Skill Theory classed under the rubric of motivation. information-processing formulation of the and the problems of horizontal decalages. Pape
presented at the meeting of the Canadian Psychc
Two limitations of skill theory are the What makes a person do one thing instead theory must include this adaptive process logical Association, Montreal, 1972.
need for a more powerful definition of skill of another when she is capable of doing if it is to provide a fair representation of 2. Aebli, H. Continuity-discontinuity from thv
domains and the need to deal with the either? the entire theory. perspective of Piagetian andpost-Piagetian theory
The omission of accessing also means that A person should not be treated as a Paper presented at the fifth meeting of the Inter
processes by which skills are accessed. national Society for the Study of Behavioral De
skill theory neglects many of the phenomena disembodied brain developing in a virtual
velopment, Lund, Sweden, June 1979.
Defining Skill Domains of memory and attention that are such environmental vacuum. In some cognitive 3. Fischer, K. W., & Roberts, R. J., Jr. A develop
central concerns within the information- theories that make sharp distinctions be- mental sequence of classification skills in preschoa
Skill theory provides a mechanism for processing framework (see Estes, 1976). tween competence and performance, the children. Manuscript submitted for publication
predicting and explaining the development Skill theory should be able to predict the environment and the person’s adaptation to 1979.
3f skills in specific task domains, and it it are effectively left out. The issue of the 4. Fischer, K. W. The hierarchy of intellectual de
development of memory skills, and it has velopment: Piaget systematized. Paper presente,
also gives a general portrait of how popula- already been used as a tool for uncovering processes by which skills are accessed at the meeting of the American psycho logic^
tions of skills change with development. But some new memory phenomena, such as a should not be confused with this issue of Association, New Orleans, September 1974.
 526 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 52;

5. Miller, A. I. Piaget’s genetic epistemology and

the decetopment ofquantum mechanics $om m.7
to 1927: A study of structuralism. Unpublished
Piagetian Theory and the Helping Professions.
Los AngeIes: Uniyeriity of s o d e r n CaGfOmia,
in press.
G a l k a g k S {E&.&,
New York: Plenum, 1977.
**
of current research. In W. F. Overton & J. M.
4 d”f. gws,r, €85.
mental slages. Behavioral and Brain Sciences

Erikson, E. H. Childhood and society. New York

manuscript, Lowell Institute of Technology, 1972. Beth, E. W., & Piaget, J. Mathematical epistemology Chomsky, N. Aspects of the theory of syntax. Cam- Norton, 1%3.
6. Harris, P. L. Subject, object, and .finmework: and psycho1og.v (W. Mays, trans.). Dordrecht. bridge, Mass.: M. I. T. Press, 1965. Erikson, E. H. Youth Fidelity and diversity. In A
A theory of spatial development. Unpublished Holland: D. Reidel. 1966. (Originally published as Cole, M., & Bruner, J. S. Cultural differences and E. Winder & D. L. Angus (Eds.), Adolescence.
manuscript, Free University of Amsterdam, 1977. Etudes d’Epistimologie Ggnitique, 1961 14.)
~
inferences about psychological processes. American Contemporary studies. New York: American Book
7. Fischer, K. W. Towards a mezhod for assessing Bickhard, M. H. The nature of developmental stages. Psychologist, 1971,26, 867-876. Company, 1974.
continuities and discontinuities in cognitive de- Humnn De~elopment,1978.21. 217-233. Cornell, E. H., & Gottfried, A. W. Intervention with Estes, W. K. (Ed.). Handbook of learning ana
velopnicnt. Paper presented at the fifth meeting of Braine, M. D. S. Children’s first word combinations. premature infants. Child Development, 1976, 47, cognitive processes: Vol. 4 . Attention and memory.
the International Society for the Study of Behavioral Monographs of the Society for Research in Child 32-39. Hillsdale, N.J.: Erlbaum, 1976.
Development, Lund, Sweden, June 1979. Development, 1976, 41(l, Serial No. 164). Comgan, R. Patterns of individual communication Feldman, C. F., & Toulmin, S. Logic and the theory
8. Bullinger. A. Sensorimotriciti et cognition chei Brainerd, C. J. The role of structures in explaining and cognitive development. (Unpublished doctoral of mind. In J. K. Cole & W. J. Arnold (Eds.),
le nourisson. Manuscript submitted for publication, behavioral development. In K. F. Riegel & J. A. dissertation, University of Denver, 1976.) Disserta- Nebraska Symposium on Motivation (Vol. 23).
1979. Meacham (Eds.), The developing individual in n tion Abstracts International, 1977, 37( 10). 5393B. Lincoln: University of Nebraska Press, 1975.
9. Mounoud, P., & Hauert, C. A. Transition periods changing world (Vol. I). The Hague: Mouton, 1976. (Upiversity Microfilms No. 77-7400) ?.- Field, 3. The adjustment of reaching behavior to
as critical periods in development. Paper presented Brainerd, C. J. The stage question in cognitive- Corrigan, R. Language development as related to a r b infancy. Child Development.
at the fourth meeting of the International Society developmental theory. The Behavioral and Bruin 6 object permanence development. Journal of
for the Study of Behavioral Development, Pavia, Sciences, 1978, 1 , 173-182. Language, 1978,5, 173-189. 4
Fischer, K. W. The structure and development of
Italy, September 1977. Breland, K., & Breland, M. The misbehavior of or- Comgan, R. Cognitive correlates of language: D&] sensory-motor actions. Unpublished doctoral dis-
ganisms. American Psychologist, 1961, 16, 681-
684.
Brown, A. L., & French, L. A. The zone of potential
Development, 1979,50, 617-631. -’
ferential criteria yield differential results. Child \

Comgan, R. The use of repetition to facilitate s p n -

sertation, Harvard University, 1970.
Fischer, K. W. Structuralism for psychologists.
Contemporary Psychology, 1972, 17, 329-331.
References
development: Implications for intelligence testins in taneous language acquisition. Journal of Psycho- Fischer, K. W. Cognitive development as problem
Achenbach, T. M., & Weisz, J. R. Alongitudinal study the year 2000. Intelligence, 1979,3, 255-274. linguistic Research, 1980,3, 231-241. solving: The meaning of decalage in seriation tasks.
of developmental synchrony between conceptual Brown, R. Socialpsychology. New York: Free Press, Comgan, R. The effects of task and practice on Proceedings of the Fifrh Annual Conference on
identity, seriation, and transitivity of color, number, 1965. search for invisibly displaced objects. Quarterly Re- Structural Learning. MERGE Research Institute,
and length. Child Development, 1975,46, 840-848. Bruner, J. S. The growth and structure of skill. In view of Development, in press. ONRC Technical Report, June 30, 1974.
Adelson, J. The political imagination of the adoles- K. J. Connolly (Ed.), Mechanisms of motor skill Cunningham, M. Intelligence: Its organization and de- Fischer, K. W. Thinking and problem solving: The
cent. In J. Kagan & R. Coles (Eds.), Twelve to development. New York: Academic Press, 1971. velopment. New York: Academic Press, 1972. cognitive approach. In K. W. Fischer, P. Shaver,
sixteen: Early adolescence. New York: Norton, Bruner, J. S. Organization of early skilled action. Denney, N. W., Zeytinoglu, S., & Selzer, C. Con- A. Lazerson (Eds.), Psychology today:An introduc-
1972. Child Development, 1973, 44, 1-11. servation training in four-year-old children. Journal tion (3rd ed.). Del M a , Calif.: CRM Books, 1975.
Aebli, H. A dual model of cognitive development: Bryant, P. E. Perception and understanding in young of Experimental Child Psychology, 1977, 24, 129- Fischer, K. W. Learning as the development of or-
Structure in cultural stimulation, construction by children: An experimental upproach. New York: 146. ganized behavior. Journal of Structural Learning,
the child. International Journal of Behavioral Basic Books, 1974. de Villiers, P. A,, & Herrnstein, R. J. Toward a law 1980,3, 253-267.
Development, 1978, 1. 221-228. Bryant, P. E. Logical inferences and development. of response strength. Psychological Bulletin, 1976, Fischer, K. W., & Bullock, D. Sequence, synchrony,
Arlin, P. K. Cognitive development in adulthood: In B. A. Geber (Ed.). Piaget and knowinx: Studies 83, 1131-1153. and constraint in cognitive development. In K. W.
A fifth stage? Developmental Psychology, 1975, in genetic epistemology. London: Routledge & Dewey, J. The reflex arc concept in psychology. Fischer (Ed.), Cognitive development. San
1 1 , 602-606. Kegan Paul, 1977. Psychological Review, 18%,3, 357-370. Francisco: Jossey-Bass, in press.
Baldwin, I . M. Mental development in the child and Bryant, P. E., & Trabasso, T. Transitive inferences Drozdal, J. G., Jr., & Flavell, J. H. A developmental Fiscber, K. W., & Comgan, R. A skill approach to
the race (3rd ed., rev.). New York: MacMillan. and memory in young children. Nature, 1971. study of logical search behavior. Child Develop- language development. In R. Stark (Ed.), Language
1925. 232, 456-458. ment, 1975,46, 389-393. behavior in infancy and early childhood. Amster-
Bandura, A., & Walters. R. H. Social learning and Bullinger, A. Orientation de la tete du nouveau-ne Duncker, K. On problem solving (L. S. Lees, trans.). dam: Elsevier, in press.
personality development. New York: Holt, en presence d’un stimulus visuel. L’AnnPe Psy- Psychological Monographs, 1945,58(5, Whole No. Fischer, K. W., & Jennings, S. The emergence of
Rinehart & Winston, 1963. ., Fhologique, 1977, 77, 357-364. 270). (Originally published, 1935.) representation in search: Understanding the hider
Baron, J. Semantic components and conceptual de- ‘*fp;rworth, G. Asymmetrical search errors in in- Elkind, D., & Sameroff, A. Developmental psy- as an independent agent. Quarterly Review of Devel-
velopment. Cognition, 1973,2, 299-318. ’. Bcy. Child Development, 1976,47, 864-867. chology. Annual Review of Psychology, 1970, 21, opment, in press.
Barratt, €3. B. Training and transfer in combhat&& b.9,S . Cognitive competence. In K. Connoily 191-238. Fischer, K. W., & Lazerson, A. Human develop-
problem solving: The development of formal reason- J. Bruner (Eds.), The growth of competence. Lon- Emde, R., Gaensbauer, T., & Harmon, R. Emotional ment. New York Worth, in press.
ing during early adolescence. Developmental don: Academic Press, 1973. expression in infancy: A biobehavioral study. Fischer, K. W., & Watson, M. W. Explaining the
Psychology, 1975.11, 700-704. Case, R. Structures and strictures: Some functional Psychological Issues, 1976, IO, No. 37. Oedipus conflict. In K. W. Fischer (Ed.), Cognitive
Beilin, H. Developmental stages and developmental limitations on the course of cognitive growth. Endler, N. S . , & Magnuson, D. Toward an inter- development. San Francisco: Jossey-Bass, in press.
processes. In D. R. Green, M. P. Ford, & G. B. Cognitive Psychology, 1974,6, 544-574. actional theory of personality. Psychological Fitts, P. M., & Posner, M. I. Human performance.
Flamer (Eds.), Measurement and Piaget. New Case, R. Intellectual development from birth to adult- Bulletin, 1976,83, 956-974. Belmont, California: BrookdCole, 1967.
York: McGraw-Hill, 1971. hood: A neo-Piagetian interpretation. In R. Siegler Epstein, H. T. Phrenoblysis: Special brain and mind Flavell, J. H. The developmental psychology of Jean
Bertenthal, B. I., Campos, J., & Haith, M.M. De- (Ed.), Children’s thinking: What develops? Hills- growth periods: I. Human brain and skull develop Piaget. New York: Van Nostrand, 1%3.
velopment of visual organization: The perception dale, N.J.: Erlbaum, 1978. ment. Developmental Psychobiology, 1974, 7, Flavell, J. H. Comments on Beitin’s paper. In D. R.
of subjective contours. Child Development, in press. Catania, A. C. The psychologies of structure, func- 207-216. (a) Green, M. P. Ford, & G. B. Flamer (Eds.),
Bertenthal, B. I., & Fischer, K. W. The develop- tion, and development. American Psychologist. Epstein, H. T. Phrenoblysis: Special brain and mind Measurement and Piaget. New York: McGraw-
ment of self-recognition in the infant. Developmental 1973,28, 434-443. growth periods: 11. Human mental development. Hill, 1971. (a)
Psychology. 1978, 14. 44-50. Catania, A. C. The psychology of learning: Some Developmental Psychobiology, 1974, 7, 217-224. Flavell, J. H. Stage-related properties of cognitive
Bertenthal, B. I., & Fischer, K. W. Toward an lessons from the Darwinian revolution. Annals ofthe (b) development. Cognitive Psychology, 1971, 2,
understanding of early representation. In Pro- N e u York Academy of Sciences, 1978,309. 18-28, Epstein, H. T. Some additional data relevant to con- 421-453. (b)
ceedings of rhe Ninth Annual Conference on Chandler, M. J . Social cognition: A selective review siderations about the existence of cognitive-develop- Flavell, J. H. An znalysis of cognitive-developmental

Y
 528 KURT W. FISCHER THEORY O F COGNITliVE DEVELOPMENT 52!
sequences. Generic Psychology Monographs, 1972, In R. Held, H. Leibowitz, & H. L. Teuber (Eds ), on the reaching and retrieval behavior of youn
@5; 279-358.
New Yozk: Harper & Row, 1964. (Originally
HanfiooE ofsensory pi5ysioTogy P O I . 81 E& pi&&ed, €S9..f isants. CTiiid Deidopmem. m,&% tf2-ti7.
Flavell, J. H. Cognitive development. Englewood Springer, 1978. Leiser. D. Les reprksentations d’une procedure d
Cliffs, N.J.: Prentice-Hall, 1977. Isaac, I). J., & O’Connor, B. M. A discontinuity
Haith, M. M., & Campos, J. J. Human infancy. theory of psychological development. Human seriation chez l’adulte. Unpublished doctoral di!
Flavell, J. H., & Wohlwill, J. F. Formal and func- Annual Review of Psychology, 1977,28, 251-294. sertation. University of Geneva, 1977.
tional aspects of cognitive development. In D. Re/ations. 1975,23, 41-61.
Halford, G. S. A theory of the acquisition of con- Jackson, E., Campos, I. J., & Fischer, K. W. The Liben, L. S . The facilitation of long-term memo1
Elkind & J. H. FlaveU (Eds.), Studies in cognitive servation. Psychological Review, 1970,77. 302-3 16. improvement and operative development. Develoi
question of decalage between object permanence and
development. London: Oxford University Press, Halford. G. S., & Wilson, W. H. A category theory person permanence. Developmental Psychology, mental Psychology, 1977, 13, 501-508.
1969. approach to cognitive development. Cognitive Lunzer, E. A. Problems of formal reasoning in te:
1978.14, 1- 10.
Fleishman, E. A. Toward a taxonomy of human per- Psychology. 1980, 12, 356-411. situations. In P. H. Mussen (Eds.), Europea
formance. American Psychologist, 1975, 30, Jackson, S. The growth of logical thinking in normal
Hand, H. H. The development of concepts of social and subnormal children. British Journal of Educa- research in cognitive development. Monographs c
1127-1149. interaction: Children‘s understanding of nice and the Society for Research in Child Developmen
tional Psychology, 1%5,35, 255-258.
Furby, L. Cumulative teaming and cognitive develop- mean. Unpublished doctoral dissertation, Uni- 1%5,30(2, Serial No. 100).
Jamison, W. Developmental inter-relationships among
ment. Human Development, 1972. 15, 265-286. versity of Denver, 1980. concrete-operational tasks: An investigation of MacWhinney. B. Starting points. Language. 197’
Furth, H. G. Piaget and knowledge. Englewood Hanson, N. R. Patterns of discovery. Cambridge. 53, 152-168.
CWs, N.J.: Prentice-Hall, 1969. Piaget’s stage concept. Journal of Experimental
England: Cambridge University Press, 1961. Child Psychology, 1977,24, 235-253. MacWhinney, B. The acquisition of morphophonolog:
Gagne, R. M. Contributions of leaming to human Hams, P. L., & Bassett, E. Transitive inferences by Monographs of the Society for Research in Chit
Jennings, S . J. H. Developing representations of
development. Psychological Review, 1968, 75, 4-year-old children. Developmental Psychologj. Dei*e/opment,1978,43( 1-2, Serial No. 174).
177-191. events: An analysis of search in Piciget’s insisible
1915, 1 1 , 875-876. displacement problem. Unpublished doctoral dis- Mandler, G. From association to structure. Psych,
Gagne, R. M. The conditions of learning. New York Harter, S . A cognitive-developmental approach to logical Revien, 1962.69, 415-427.
Holt, Rinehart & Winston, 1970. sertation. University of California at Santa Barbara,
children’s expression of conflicting feelings and a 1975. hfarler, P., & Hamilton, W. J., 111. Mechanisms I
Gelman, R. Cognitive development. Annual Review of technique to facilitate such expression in play animal behavior. New York: Wiley, 1966.
Psychology, 1978.29, 297-332. Kagan, J. The concept of identification. Psychological
therapy. Journal of Consulting and Clinical Psx- Review, 1958,65, 296-305. Martarano, S. C. A developmental analysis of pe
Goidstein, D. M., & Wicklund, D. A. The acquisition chology, 1977.45, 417-432. formance on Piaget’s formal operations tasks. D
of the diagonal concept. Child Development, 1973, Kagan, J. Structure and process in the human infant:
Hatano, G., Miyake, Y., & Binks, M. G. Performance The ontogeny of mental representation. In M. H. velopmenral Psychology, 1977, 13, 666-612.
44, 210-213. of expert abacus operators. Cognition, 1977, 5, McCall. R. B. Qualitative transitions in behavior
Gollin, E., & Saravo, A. A development analysis of Bomstein & W. Kessen (Eds.), Psychological
47-56. deivlopment from infiinc?: Image to intention. development in the first two years of life. In M. I
Laming. In J. Heilmnth (Ed.), Cognitive studies I. Hebb, D. 0. What psychology is about. American Bomstein & W. Kessen (Eds.), Psychological d
Hillsdale, N.J.: Erlbaum, 1979.
New York: BrunneriMazel. 1970. Psychologist, 1974,29, 71-79. velopment from infancy: Image to intention. Hill
Goodson, B. D., & Greenfield, P. M. The search for Kaplan, B. Meditations on genesis. Human Develop-
Hermstein. R. J. The evolution of behaviorism. ment, 1967, I O , 65-87. dale, N.J.: Erlbaum, 1979.
structural principles in children’s manipulative play: American Psychologist, 1977.32, 593-603. McCall, R. B., Eichorn, D. H., & Hogarty, P.
A parallel with linguistic development. Child Kaye, K. The development of skills. In G. Whitehurst
Hinde, R. A. Animal behaviour (2nd ed.). New York: & B. Zimmerman (Eds.), The functions of language Transitions in early mental development. Mon
Development, 1975,46, 734-746. McGraw-Hill, 1970. graphs of the Society for Research in Child D
and thought. New York: Academic Press, 1979.
Gottlieb, D. E., Taylor, S. E., & Ruderman, A. Hobhouse, L. T. Mind in evolution. London: Mac- velopment, 1977,42(3, Serial No. 171).
Kemler, D. G., & Smith, L. B. Is there a develop-
Cognitive bases of children’s moral judgments. Millan, 1915. mental trend from integrality to separability in per- McGurk, H., & MacDonald, J. Auditory-visu
Developmental Psychology, 1977,13, 547-556. Hooper, F. H., Goldman, J. A., Storck, P. A,, & coordination in the first year of life. Internation
Gottlieb, G. Conceptions of prenatal development: ception? Journal of Experimental Child Psychology,
Burke, A. M. Stage sequence and correspondence 1978.26. 498-507. Journal of Behavioral Development, 1978, I , 22!
Behavioral embryology. Psychological Review, in Piagetian theory: A review of the middle-child- 240.
Kessen, W. Questions for a theory of cognitive de-
1976,83, 215-234.
Greenfield, P. M. Cross-cultural research and hood period. In Research Relating t o Children, velopment. In H. w. Stevenson (Ed.), Concept of Mead, G. H. Mind, self, and society from the stan
Bulletin 28. Urbana, Ill.: Educational Resources development. Monographs of the Society for Re- point of a social behaviorist (C. W. Moms, Ed
Piagetian theory: Paradox and progress. In K. F. Information Center, 1971. Chicago: University of Chicago Press, 1934.
Riegel & J. A. Meacham (Eds.), The developing search in Child Development, 1966,31(5, Serial NO.
Hooper, F. H., Sipple, T. S., Goldman, J. A., & 107). Moore, T. E., & Harris, A. E. Language and thoug
individual in a changing world (Vol. 1). The Hague: in Piagetian theory. In L. S. Siege1 & C. J. Braine
Mouton, 1976. Swinton, S. S. A cross-sectional investigation of Klahr, D., & Wallace, J. G. Cognitive development:
children‘s classificatory abilities. Generic Psy- An information processing view. Hillsdale, N.J.: (Eds.), Alternatives to Piaget: Critical essays on ti
Greenfield, P., Saltzman, E., & Nelson, K. The de- Erlbaum, 1976. theory. New York: Academic Press, 1978.
velopment of rulebound strategies for manipulating chology Monographs, 1979,99, 4 1-89.
Horn, J. L. Human abilities: A review of research Kofsky, E. A scalogram study of classificatory de- Mounoud, P. Les rtvolutions psychologiques
seriated cups: A parallel between action and I’enfant. Archives de Psychologie, 1976, 44 (17
and theory in the early 1970‘s. Annual Review of velopment. Child Development, 1966,37, 191-204.
grammar. Cognitive Psychology, 1972.3, 291-310. 103- 114.
Greenfield, P. M., & Schneider, L. Building a tree Psychology, 1976,27, 437-486. Kohlberg, L. Stage and sequence: The cognitive-
Hunt, J. McV. The impact and limitations of the giant developmental approach to socialization. In D. A. Mounoud, P. L’utilisation du milieu et du corps prop
structure: The development of hierarchical com- par le btbe. In J. Piaget, P. Mounoud, & J.
of developmental psychology. In D. Elkind & J. H. G o s h (Ed.), Handbook of socialization theory
plexity and intempted strategies in children’s Bronckart (Eds.), La psychologie. Paris: E
construction activity. Developmental Psychology, Flavell (Eds.), Studies in cognitive development. and research. Chicago: Rand McNally, 1969.
London: Oxford University Press, 1%9. Kopp, C. B., O’Connor, M. J . , & Finger, I. Task cyclopedie de la Pleiade, GaUimard, 1980.
1977, 13, 299-313. Mounoud, P., & Bower, T. G. R. Conservation
Gruber, H. E. Courage and cognitive growth in Hunt, J. McV. Implications of sequential order and characteristics and a Stage 6 sensorimotor problem.
Child Developm ent , 1975,46, 569-573. weight in infants. Cognition, 1974/1975,3, 29-40
children and scientists. In hf. Schwebel & J. Raph hierarchy in early psychological development. In
B. 2. Friedlander, G. M. Sterritt, & G. E. Kirk Koslowski, B., & Bruner, J. S. Learning to use a Negoita, C. V., & Ralescu, D. A. Applications
(Eds.), Piaget in the classroom. New York: Basic fuzzy sets to systems analysis. New York: Halst
Books, 1973. (Eds.), Exceptional infant (Vol. 3). New York lever. Child Development, 1973,43, 790-799.
Kuhn, D. Relation of two Piagetian stage transitions Press, 1975.
Gruber, H. E., & Barrett, P. H. Darwin on man: Brunner/Mazel, 1975.
to IQ. Developmental Psychology, 1976, 12, Neimark, E. Intellectual development during ado1
A psychological study of creativity. New York: Inhelder, B., & Piaget, J. The growth of [ogicnl 157-161. scence. In F. D. Horowitz (Ed.), Review of chi
Dutton, 1974. thinking from childhood to adolescence (A. Parsons Lashley, K. S. In search of the engram. In Society of development research (Vol. 4). Chicago: Universi
Gruber, H. E., & Vontche, J. Rtflexions sur les & S . Seagrim, trans.). New York: Basic Books,
experimental biology symposium no. 4: Physio- of Chicago Press, 1975.
op6rations formelles de la penste. Archives de 1958. (Originally published, 1955.) logical mechanhms in animal behaviour. Cambridge, Nelson, K. Structure and strategy in learning to tal
Psychologie, 1976,64(171). 45-56. Inhelder, B., & Piaget, I. The early growth of lofic Monographs of the Society for Research in Chi
Haith, M. M. Visual competence in early infancy. England: Cambridge University Press, 1950.
in the child (G. A. Lunzer & D. Papert, trans.). Lasky, R. E. The effect of visual feedback of the hand Development, 1973,38(1-2, Serial No. 149).
 KURT W. FISCHER THEORY OF COGNITIVE DEVELOPMENT 53 1

0, A., & Ljeb!ichl A. The grammar o€h piaser, 3. k k i k t m & e x h t i m f” actotesce-ce to

-P SfneWe” m drihtren’s copying. ChZd aduIthood. Human Development, 1972,15, 1-12.
s&Ua, c. E., & p*, s. 6. cqiesivity and In K F. €Gg&& I. k I Z e a c h a EXs.J, 7%
developing individual in a changing world (Vol. 1).
rvelopment, 1976,47, 846-850. reversibility in children’s understanding of causal
F’iaget, J. L‘tquilihration des structures cognitives: sequences. Child Development, 1978, 49, 1219- The Hague: Mouton, 1976.
m, R. D. A perceptual-salience account of Problbme central du dtveloppement. Etudes Verge, C. G . , & Bogartz, R. S. A functional measure-
calage relations and developmental change. In 1222.
d‘Epistimologie Ginitique, 1975,33. Shatz, M. The relationship between cognitive pro- ment analysis of the development of dimensional
S. Siegel & C. J. Brainerd (Eds.), Alternatives F’iTet, J., Gnze, J.-B., Szeminska, A., & Vinh Bang. coordination in children. Journal of Experimental
Piaget: Critical essays on the theory. New York: cesses and the development of communication
Episttmologie et psychologie de la fonction. skills. In H. E. Howe, Jr., & C. B. Keasey (Eds.), Child Psychology, 1978.25, 337-353.
:ademic Press, 1978. Etudes d’Epistimologie GinCtique, 1%8,23. Vygotsky. L. S. Thought and language (E. Hanfmann
luch, R. E. The role of complexity in Piaget’s Piaget, J., Inhelder, B. The psychology of the chrld Nebraska Symposium on Motivation (Vol. 25).
Lincoln: University of Nebraska Press, 1977. & G. Vakar, trans.). Cambridge, Mass.: M.I.T.
ject permanence tasks. (Doctoral dissertation, (H. Weaver, trans.). New York: Basic Books, 1969. Press, 1962.
niversity of California at Los Angeles, 1978.) Siegel, A. W.. & White, S. H. The development of
(Originally published, 1966.) spatial representations of large-scale environments. Wason, P. C. The theory of formal operations- A
ssertation Abstracts International, 1979, 39, Piaget, J., Inhelder, B. Mental imagery in the child critique. In B. A. Geber (Ed.), Piaget and knowing:
60B. (University Microfilms No. 7901386) In H. W. Reese & L. P. Lipsitt (Eds.), Advances
(P. A. Chilton, trans.). New York: Basic Books, in child development and behavior (Vol. IO). New Studies in genetic epistemology. London: Rout-
rusek, H. Experimental studies of appetitional 1971. (Originally published, 1966.) ledge & Kegan Paul, 1977.
havior in human newborns and infants. In H. W. Piaget, J., York: Academic Press, 1975.
Inhelder, B. The chrlds construction of Siegler, R. S. Three aspects of cognitive development. Watson, M. W. A developmental sequence of social
svenson, E. H. Hess, & H. L. Rheingold (Eds.), quantities: Conservation and atomrsm (A. J. role concepts in preschool children (Doctoral disser-
rly behavior. New York: Wiley, 1967. Cognitive Psychology, 1976,8, 48-520.
Pomerans, trans.). London: Routledge & Kegan Skinner, B. F. The behavior of organisms. New York: tation, University of Denver. 1977.) Dissertation Ab-
ual-Leone, J. A mathematical model for the transi- P a d , 1974. (Originally published, 1941.) stracts international, 1978. 38, 4546B. (University
n rule in Piaget’s developmental stages. Acta Piaget, J., & Szeminska, A. The child‘s conception of Appleton-Century-Crofts, 1938.
Skinner, B. F. Contingencies of reinforcement: A Micjofilms No. 7801402)
ychologica, 1970,32, 301-345. number (C. Gattegno & F. M. Hodgson, trans.). Watson, M. W.. & Fischer, K. W. A developmental
, E. J. Invention and discovery of reality. New theoretical analysis. New York: Appleton-Century-
New York: Humanities Press, 1952. (Originally Crofts, 1969. sequence of agent use in late infancy. Child De-
irk: Wilev. 1975. published, 1941.) velopment, 1977.48, 828-835.
:r, A. Cerebral function in infancy and childhood. Pikas, A. Abstraction and concept formatron. Cam- Smedslund, J. Concrete reasoning: A study of intel-
lectual development. Monographs of the Society f o r Watson, M. W., & Fischer, K. W. Development of
‘w York: Consultants Bureau, 1963. bridge, Mass.: Harvard University Press, 1966. social roles in elicited and spontaneous behavior
‘s, A. M. Language learning strategies: Does Premack, D. Reinforcement theory. In D. Levtne Research in Child Development, 1964.29, (2, Serial
No. 93). during the preschool years. Developmental PSF-
: whole equal the sum of the parts? Language, (Ed.), Nebraska Symposium on Motivation (Vol. Smith, L. B., & Kemler, D. G. Developmental trends chology, 1980, 16, 483-494.
77,53, 560-573. 13). Lincoln: University of Nebraska Press, 1965. Wellman, H. M., Somerville, S. C.. & Haake, R. J.
.t, J. Le mbcanisme du dtveloppement mental Reed, G. F. Skill. In E. A. Lunzer & J. F. Moms in free classification: Evidence for a new conceptual-
ization of perceptual development. Journal of Development of search procedures in real-life spatial
les lois du groupement des op4rations. Archives (Eds.), Development in learnrg: Vol. II. Develop- environments. Developmental Psychology, 1979,
Psychologie, 1941,28, 215-285. Experimental Child Psychology7 1977,24, 279-298.
ment in human learning. New York: American Smith, L. B., & Kemler, D. G. Levels of experi- 15, 530-542.
t, J. Classes, relations, et nombres: Essai sur les Elsevier 1968
--->
enced dimensionality in children and adults. Cog- Werner, H. Comparative psychologJ of mental de-
wpements de la logistique et sur la reversibilite de Reed, S. K., Ernst, G. W., & Banerji, R. The role of velopment. New York: Science Editions, 1948.
?ens&. Paris: Vrin, 1942. nitive Psychology, 1978, 10, 502-532.
analogy in transfer between similar problem states. Suppes, P. Introduction t o logic. New York: Van Werner, H. The concept of development from a
t, J. Traiti de logique: Essai de logistique Cognitive Psychology, 1974, 6, 436-450. Nostrand, 1957. comparative and organismic point of view. In D. B.
;ratoire. Paris: Colin, 1949. Reese, H. W., & Lipsitt, L. P. Experimental child Teitelbaum, P. Levels of integration of the operant. Harris (Ed.), The concept of development. Min-
t, J. The psychology of intelligence (M.Piercy psychology. New York: Academic Press, 1970. neapolis: University of Minnesota Press, 1957.
D. E. Berlyne, trans.). New York: Harcourt, Rest, J. R. New approaches in the assessment of moral In W. K. Honig & J. E. R. Staddon (Eds.),
Handbook of operant behavior. Englewood Cliffs, Wilkering, F. Combining of stimulus dimensions in
ice, 1950. (Originally published, 1947.) judgment. In T. Lickona (Ed.), Moral development children’s and adults’ judgments of ages: An in-
t, J. Play, dreams, and imitation in childhood and behavior. New York: Hob, Rinehart & formation integration analysis. Developmental Psy-
Gattegno & F. M.Hodgson, trans.). New York: Toussaint, N. A. An analysis of synchrony between
Winston, 1976. concrete-operational tasks in terms of structural and cholog?, 1979, I S , 25-33.
rton, 1951. (Originally published, 1946.) Riegel, K. F. (Ed.). The development of dialectical performance demands. Child Development, 1974, Winner, E., Rosenstiel, A. K., & Gardner, H. The
t, J. The origins of intelligence in children operations. Human Development, 1975,18, 1-238. development of metaphoric understanding. Develop-
Cook, trans.). New York: International Uni- Ruff, H. A. The coordination of manipulation and 45. 992-1001.
Tucker, J. The concepts of and the attitudes toward mental Psychology, 1976, 17. 289-297.
sities Press, 1952. (Originally published, 1936.) visual fixation: A response to Schaffer (1975). Wise, K. L., Wise, L. A , , & Zimmerman, R. R.
t, J. The construction of reality in the child work and play in children. Unpublished doctoral
Child Development, 1976,47, 868-871. dissertation, University of Denver, 1979. Piagetian object permanence in the infant rhesus
Cook, trans.). New York: Basic Books, 1954. Sameroff, A. J. Can conditioned responses be monkey. Developmental Psychology, 1974, IO,
iginally published, 1937.) Uzgiris, I. Situational generality in conservation.
established in the newborn infant: 1971? Develop- 429-437.
:,J. Logique et Cquilibre dans Ies comportements Child Development, 1964,35, 831-842.
mental Psychology, 1971,5, 1- 12. Wohlwill, J. F. The study of behavioral development.
sujet. Etudes d’kpistkmologie Ginitiques, 1957, Sameroff, A. Transactional models in early social Uzgiris, I. C. Organization of sensorimotor intelli- New York: Academic Press, 1973.
!7- 118. relations. Human Development, 1975, 18, 65-79. gence. In M. Lewis (Ed.), Origins of intelligence: Wohlwill, J. F., & Lowe, R. C. An experimental
, J. Piaget’s theory. In P. H. Mussen (Ed.), Scandura, J. M. Structural learning. New York: Infancy and early childhood. New York: Plenum analysis of the development of conservation of
michael‘s manual of child psychology. New Gordon & Breach, 1973. Press, 1976. number. Child Development, 1962,33, 153- 167.
.k: Wiley, 1970. Scam-Salapatek, S. An evolutionary perspective on Uzgiris, I. C., & Hunt, J. McV. Assessment in in- Zelazo, P. R., Zelazo, N. A., & Kolb, S. “Walking”
, J. Structuralism (C. Maschler, trans.). New infant intelligence: Species patterns and individual fancy: Ordinal scales ofpsychological development. in the newborn. Science, 1972, 176, 314-315.
k: Basic Books, 1970. (Originally published, variations. In M: Lewis (Ed.), Origins of intelli- Urbana: University of Illinois Press, 1975.
2 ,
J.1 gence: Infancy and early childhood. New York: Van den Daele, L. D. Formal models of development. Received F e b r u a r y 22, 1980
, J. Biology and knowledge (B. Walsh, trans.). Plenum Press, 1976.
cago: University of Chicago Press, 1971. Schaeffer, B. Skill integration during cognitive de-
ginally published, 1967.) velopment. In R. A. Kennedy & A. Wilkes (Eds.),
, J. The theory of stages in cognitive develop Studies in Zong-term memory. New York: Wiley,
it. In D. R. Green, M.P. Ford, & G. B. Flamer 1975.
;.), Measurement and Piaget. New York: Schick, K. Operants. Journal of the Experimental
h w - H i I I , 1971. Analysis of Behavior, 1971, I S , 413-423.