Beyond Amplification: Using the Computer to Reorganize

Mental Functioning
Roy D. Pea

To cite this version:

Roy D. Pea. Beyond Amplification: Using the Computer to Reorganize Mental Functioning. Educational
Psychologist, 1985, 20(4), pp.167-182. ⟨hal-00190536⟩

HAL Id: hal-00190536

[Link]
Submitted on 23 Nov 2007

HAL is a multi-disciplinary open access archive L’archive ouverte pluridisciplinaire HAL, est des-
for the deposit and dissemination of scientific re- tinée au dépôt et à la diffusion de documents scien-
search documents, whether they are published or not. tifiques de niveau recherche, publiés ou non, émanant
The documents may come from teaching and research des établissements d’enseignement et de recherche
institutions in France or abroad, or from public or pri- français ou étrangers, des laboratoires publics ou
vate research centers. privés.

HAL Authorization
. Educat~onalPsychologist I
1985, Vol 20, No. 4, 167-182

. Beyond Amplification:
Using the Computer to Reorganize
Mental Functioning
Roy D. Pea
Center for Children and Technology
Bank Street College of Education

Computers are classically viewed as amplifiers of cognition. A n alternative

conceptualization is offered of computer as reorganizer of mental func-
tioning. Software analyses illuminate the advantages of the latter approach
for new visions of the potential cognitive benefits of computers. A new re-
sult emerges: Because the cognitive technologies we invent serve as instru-
ments of cultural redefinition (shaping who we are by changing, not just
amplifying, what we do), defining educational values becomes a foreground
issue. The demands of an information society make an explicit emphasis o n
general cognitive skills a priority. The urgency of updating education's
goals and methods recommends an activist research paradigm: to simulta-
neously create and study changes in processes and outcomes of human
learning with new cognitive and educational tools.

The computer, that symbolic workhorse and this technology offer the processes of educa-
supreme number-cruncher, has lately become a tion? What is unique about its workings as a
central topic of thought and discussion for edu- tool for the intellect? How does or should its
cators and psychologists. Brought by the ad- uses in societv influence what is done in schools
vent of inexpensive and relatively powerful soft- with computers? How can our research inquir-
ware for personal computers, now within the ies contribute to the understanding and effec-
budgetary considerations of most if not all tive design of these major changes to the face of
school systems, the uses of computers have education? How might information technol-
raised many seminal questions about the future ogies redefine the very possibilities of educa-
of education and for the research community in tion?
psychology and education. What exactly does In this paper, I analyze some of these issues,
shaped by what has been learned and the
Parts of this paper were presented at the annual meetings emerging issues after 5 years of research at the
of the American Educational Research Association, Divi- Center for Children and Technology at Bank
sion C , Invited Symposium o n "Computers as Tools of the Street College in New York and related studies
Intellect: Lessons from the Past and a n Agenda for the Fu- elsewhere. Among our research projects have
ture," April 1985, Chicago, IL. My colleagues at the Center
been studies of the develowment of wroblem
for Children and Technology have contributed immeasura-
bly t o these arguments. Kurt VanLehn and Jim Kaput solving and planning skills in Logo program-
helped reveal t o me the drastic changes in mathematical ming (Kurland & Pea, 1985; Pea & Kurland,
thinking and education arising with problem-solving soft- 1984; Pea, Kurland, & Hawkins, 1985), of the
ware. Seth Chaiklin helped in thjnking about the moral is-
cognitive demands and consequences of learn-
sues (and with editing). I also thank Jerome Bruner, Richard
Shavelson, and Noreen Webb for t h e ~ rvaluable sugges- ing programming (Clement, 1984; Kurland,
tions. I would like t o express appreciation t o the Spencer 1984; Kurland, Clement, Mawby, & Pea, in
Foundation, the National Institute of Education, the press; Mawby, in press), of classroom uses of
Alfred [Link] Foundation, Carnegle Corporation, Xerox tool software such as data-base management
Corporation, and IBM University Relations for their gener-
ous support of our research at t h e center and for providing
systems a n d word processors (Freeman,
opportun~tiesto examine these issues throughout our re- Hawkins, & Char, 1984), of how teachers' in-
search projects. T h e opinions expressed here d o not terwretive frameworks for software are linked to
necessarilv reflect the policies of anv of these institutions or how they reorganize classroom learning with
individuals. new technologies (Hawkins, 1985; Hawkins &
Requests for reprints should be sent t o Roy D. Pea, Cen-
ter for Chlldren i n d Technology, Bank Street College of Sheingold, in press; Sheingold, Hawkins, &
Education, 610 West 112th Street, New York NY 10025. Char, 1984), and formative research for the cre-

Copyright 1985 by Lawrence Erlbaum Associates, Inc.

167
 168 ROY D. PEA

ation of a multimedia instructional package Historical Perspectives on Cognitive

(software, television, print) of an open-ended Technologies
nature for student learning in mathematics, sci-
ence, and technology (Char, Hawkins, Woot- Long before computers appeared, there were
ten, Sheingold, & Roberts, 1983). remarkable extensions of human intelligence
This paper highlights the potential yield in through the use of technical instruments. I take
thinking of roles for educational computing dif- as axiomatic that intelligence is not a quality of
ferently than we typically do. Computers are the mind alone, but a product of the relation
commonly believed to change how effectively between mental structures and the tools of the
we do traditional tasks, amplifying or extending intellect provided by the culture (Bruner, 1966;
our capabilities, with the assumption that these Cole &Griffin, 1980; Luria, 1976, 1979; Olson,
tasks stay fundamentally the same. The central 1976, 1985; Olson & Bruner, 1974; Vygotsky,
point I wish to make is quite different, namely, 1962, 1978). I call these tools cognitive tech-
that a primary role for computers is changing nologies. A cognitive technology is provided by
the tasks we do by reorganizing our mental any medium that helps transcend the limita-
functioning, not only by amplifying it. There- tions of the mind, such as memory, in activities
fore, although the predominant use of comput- of thinking, learning, and problem solving. The
ers in education today is with software that technologies that have received perhaps the
aims to make more efficient long-familiar drill most attention in this respect are written lan-
and practice activities in basic skills, especially guage (Goody, 1977; Greenfield, 1972; Olson,
in math and language arts, I plan to focus on 1977; Ong, 1982; Scribner & Cole, 1981) and
more recently emerging trends and o n future di- systems of mathematical notation, such as alge-
rections in using software as tools for sup- bra or calculus (Cassirer, 1910, 1957; Kline,
porting and reorganizing children's thinking. 1972).
This paper has four parts. The first sketches But consider computers as cognitive tech-
these two different conceptualizations of the nologies. Computers are universal machines for
transformational roles of noncomputer cogni- storing and dynamically manipulating symbols,
tive technologies (e.g., written language) in hu- which appear to serve as the currency of human
man intelligence and cognitive change, as am- thought (Greeno & Simon, in press). Capable
plifier and reorganizer of mind. From these of real-time programmable interactions with
conceptualizations, one is likely to draw differ- human users, computers may provide the most
ent implications for how computer technologies extraordinary cognitive technologies thus de-
may contribute to human thinking and to the vised. How can past research on noncomputer-
processes of education. T h e second part ana- based cognitive technologies guide our defini-
lyzes several examples of software as cognitive tion of priorities for future research with com-
technologies, and the findings illuminate the puters as cognitive technologies in education?
advantages of the reorganizer approach for con- Cognitive technologies, such as written lan-
ceptualizing potential benefits of computers in guages, are commonly thought of as "cultural
human thinking. The third major section traces amplifiers" of the intellect, to use Jerome
deeper consequences of the reorganizer ap- Bruner's (1966, p. xii) influential phrase. They
proach for educational uses of computer-based are viewed as cultural means for empowering
cognitive technologies. In particular, because human cognitive capacities. Greenfield and
the cognitive technologies we invent can serve Bruner (1969) observed that cultures with
as instruments of cultural redefinition (shaping technologies such as written language will
who we are by changing what we do), selecting "push cognitive growth better, earlier, and
values for educational goals becomes a crucial longer than others" (p. 654). We find similar up-
issue. I argue that cognitive demands wrought beat predictions for computer technologies
by the information society make a curriculum embodied in a widesuread belief that thev will
explicitly emphasizing general cognitive skills inevitably and profoundly amplify human men-
an important priority. In the fourth part, I sug- tal Dowers (Pea & Kurland. 1984).
gest that the urgency of updating educational This amplifier metaphor for cognitive tech-
aims and methods recommends an activist re- nologies has led to many research programs,
'
search paradigm for simultaneously creating particularly o n the cognitive consequences of
and studying changes in processes and out- literacy and schooling (e.g., on formal logical
comes of human learning with new cognitive reasoning) in the several decades since Bruner
and educational technologies. and his colleagues published Studies in Cognitive
 BEYOND AMPLIFICATION 169

Growth (e.g., Greenfield, 1972; Olson, 1976; change in the instruments of work (e.g., a plow
- Scribner & Cole, 1981). T h e amplifier meta- rather than the hand) changes the functional
~ h o continues
r in contemporary work o n elec- organization, or system characteristics, of hu-
tronic technologies by John Seeley Brown of mans' relation to work: What humans do as
Xerox PARC, who in a recent paper describes their tasks differs, not only d o they accomplish
his prototype software systems for writing and the work faster.
doing mathematics as idea amplifiers (J. S. In efforts to integrate accounts of individual
Brown, 1984a). For example, AlgebraLand, cre- and cultural changes, the Soviet theorists
ated by Brown and colleagues (J. S. Brown, Vygotsky (e.g, 1962, 1978) and Luria (1976,
1984b), is a software program in which students 1979) generalized the historical materialism
are freed from hand calculations associated that Marx and Engels developed for physical in-
with executing different algebraic operations struments and applied it to a historical analysis
and focus o n high level problem-solving strate- of symbolic tools such as written language that
gies, which they select for the computer to per- serve as instruments for redefining culture and
-
form. Alpebraland is said to enable students "to human nature. What Vygotsky (1978) recog-
explore the problem space faster," as they learn - acts as an instrument of
nized was that "the sign
equation-solving skills. Although quantitative psychological activity in a manner analogous to
metrics such as the efficiency and speed of the role of a tool in labor" (p. 52). A similar in-
learning may truly describe changes that occur strumental and dialectical perspective is re-
in problem solving with electronic tools, it can flected by recent studies of the "child as a cul-
be shown that more profound changes-as I tural invention" (Kessel & Siegel, 1983; Kessen,
later describe for the AlgebraLand example 1979; White, 1983). Take for instance, Wartof-
(after some historical background)-may be sky's (1983) statement of the shift in perspective:
missed if we confine ourselves to the am~lifica-
tion perspective. As Bruner has acknowledged Children are, or become, what they are taken t o be
(personal communication, November 10, 1985) by others, and what they come to take then~selvesto
cognitive tools can yield orders of magnitude be, in the course of their social communication and
interactions with others. In this sense, 1 take "child"
and thereby qualitative changes in forms of to be a social and historical kind, rather than a natu-
thought. ral kind, and therefore also a constructed k ~ n d
There is a different tradition in the study of rather than o n e given, so to speak, by nature in
cognitive technologies that may be character- some fixed o r essential form. (p. 190)
ized as cultural-historical. Influenced by the
writings of Vico, Spinoza, and Hegel, Marx and Using a Vygotskian perspective, which
Engels developed a theory of society now de- stresses the functional reorganization of cogni-
scribed as historical or dialectical materialism. tion with the use of symbolic technologies, Cole
Human nature, o n this view, rather than being- and Griffin (1980) argued that the amplifier
a product of environmental forces, is of our metaphor has important shortcomings. (Later,
own making and continually "becoming." Hu- I further illustrate such shortcomings, as re-
mankind is reshaped through a dialectic of re- vealed by uses of computers as electronic
ciprocal influences: O u r productive activities spreadsheets and problem-solving aids in math-
change the world, thereby changing the ways in ematics.) Specifically, they discussed how sym-
which the world can change us. By shaping na- bolic technologies qualitatively change the
ture and how our interactions with it are medi- structure of the functional system for such men-
ated, we change ourselves. As the biologist tal activities as problem solving or memory.
Stephen Jay Gould observes (l980b), such "cul- These fundamental changes are likely to go
tural evolution," in contrast t o Darwinian bio- unnoticed if one thinks only with the amplifier
logical evolution, is defined by transmission of metaphor. Cole and Griffin highlighted, in par-
skills, knowledge, and behavior through learn- ticular, how Luria enriched the term "function"
ing across generations and has been our nature- for psychology. We are often inclined to assume
transcendent innovation as a suecies. one-to-one correspondences between functions
O n this cultural-historical perspective, labor and structures (e.g., planning is the function of
is seen as the factor mediating humans and na- the frontal cortex). In contrast, Luria speaks of
ture. By creating and using physical instru- the function of respiration, not as the function
ments (e.g., machinery) that mediate in less and of particular tissue, but as an entire functional
less direct ways our interactions with nature, we system consisting of many components, such as
come to reshape human nature. Note how a the motor, sensory, and autonomic nervous
170 ROY D. PEA

systems. Cole and Griffin (1980) noted that for is a postdictive discipline rather than a predict-
Luria, "functional systems are distinguished not ive discipline.
only by the complexity of their structure, but
also by the flexibility of the roles played by con-
stituents" (pp. 347-348). Software as Cognitive Technology and
Cole and Griffin illustrated how Vygotsky, the Reorganization Metaphor
in similar fashion, saw shifts in functional sys-
tems for thinking as the sine qua non of devel- Let us now turn to the advent of pro-
opmental change: grammable electronic symbolic technologies
and see whether the concerns about the amplifi-
1 have attempted t o demonstrate t h a t t h e course of
child development is characterized by a radical alter-
cation metaphor raised by a cultural-historical
ation in the very structure of behavior; at each new perspective have heuristic value for our present
stage the child changes not only her response but purposes. Does the reorganization metaphor
carries out t h e response in new ways, drawing o n serve well in its place? Specifically, how might
new instruments of behavior a n d replacing o n e psy-
computer-based cognitive technologies such as
chological function by another. (Vygotsky, 1978,
pp. 72-73)
software fundamentally restructure the func-
tional system for thinking?
By contrast, Cole and Griffin (1980) note A cultural-historical analysis of computer
how use of the term "amplify" means to make software qua thinking tool is illuminating. 1 dis-
more powerful, and to amplify in the scientific cuss three cases in detail and mention several
sense "refers rather specifically to the intensifi- others in which software has qualitatively
cation of a signal (acoustic, electronic), which changed both the content and flow of the cog-
does not undergo change in its basic structure" nitive processes engaged in human problem
(p. 349). As such, "amplify" leads one to unidi- solving. In particular, the what and the when of
mensional, quantitative theorizing about the the constituent mental operations that a person
effects of cognitive technologies. As evidence of contributes to the computer-aided problem-
such tendencies, Cole and Griffin discuss how a solving efforts have undergone substantial
pencil can be thought of as amplifying the change.
power of a sixth grader's memory for a long list
of words when only the outcome of the list
length is considered. But it would be distortive, Example 1: Electronic Spreadsheets
they suggest, to go on t o say that the mental
process of remembering that led to the outcome A n illustration of computer technologies
was amplified by the pencil because "remem- that can reorganize, and not merely amplify,
bering" in the two cases refers t o two qualita- mental functioning is the electronic spread-
tively different activities. T h e pencil did not sheet, such as Visicalc and Lotus 1-2-3 (e.g.,
amplify a fixed mental capacity called memory; Levy, 1984). Several million copies have been
it restructured the functional system for remem- sold since they appeared in late 1979. Electronic
bering, and thereby led to a more powerful out- spreadsheets are software programs for micro-
come (at least for the purpose of remembering computers. T h e screen images physically resem-
more items). ble paper ledger sheets, with cells organized in
Olson (1976) makes similar arguments, fol- rows and columns. But the resemblances end
lowing Ong (1971) and Havelock (1973, 1978), there. In an electronic spreadsheet, one can
about restructurings of thinking processes cre- place a number, a calculation, or a formula in
ated through written language. For example, the formula area of any spreadsheet cell, which
logical analysis of arguments for consistency/ can subsequently be edited, copied, or moved.
contradiction becomes possible because mem- The results of calculations in the formula area
ory limitations for oral language are mitigated, appear as the content of the cell. The most dra-
and print (rather than oral narrative) provides a matic difference from static paper spreadsheets
means to store and communicate cultural is that one can change cell entries and see the
knowledge. It is important t o note that the spe- repercussions of that change recalculated im-
cific restructurings of cognitive technologies are mediately throughout the spreadsheet. Many
rarely predictable; they have emergent proper- lines of thought can be simultaneously acti-
ties that come to be discovered only through vated in the form of dynamic "living" plans, and
their use. In this sense, as Dilthey (1976) urged, their outcomes compared in terms of crucial
human history, like evolution (Gould, 1980a), variables. This what-if property has dramatic
 BEYOND AMPLIFICATION 171

consequences for the cognitive activity of budg- Finally, at the company level, the micro-
eting (and financial modeling), in ways to be computer electronic spreadsheet has decentral-
described. ized financial planning. Everyone is doing it.
Before 1979, ledger sheets representing finan- The number of mediating links between plan-
cial quantities, formulas relating these quanti- ning and testing financial models has been re-
ties, and their change over time were either duced rather than increased by the computer
recalculated by hand after every change or technology, and executives report feeling more
modeled with main-frame programs under the in control of their futures.
control of data-processing departments. Execu-
tives doing the financial planning were not in-
volved in these operations. Microcomputer Example 2: Software for Problem Solving
budgeting has become a highly creative means in Mathematics
for generating and testing various scenarios in
complex financial situations for what could be, Similarly, mathematics educators have be-
given different hypothetical assumptions. The gun to argue that the use of symbolic manipula-
effort rewired to formulate such scenarios in tion programs such as muMath, MacSyma, and
the past and to update them regularly made TK!Solver for doing algebra leads to a profound
such exploration unfeasible, except (in limited shift in the functions and structure of mathe-
fashion) by main frames under control of data- matical thinking from mechanical operations
processing departments, not the executive. to problem-solving operations (Conference
In terms of the reorganization metaphor, the Board of the Mathematical Sciences, 1983; Fey,
tool has restructured the mental work of budg- 1984; Maurer, 1984a, 1984b; National Science
eting. The what has changed: T h e predomi- Board, 1983). For example, the microcomputer
nant constituent mental operations for the fi- program muMath can easily do complex equa-
nancial planner are now planning and hypo- tion solving, including solution of numerical
thesis testing by means of interactive devel- and literal equations, factoring of polynomial
opment and testing of different models for expressions, evaluation of definite and indefi-
budgets. The when, or the temporal sequenc- nite integrals, differentiation of elementary
ing, of mental operations in the functional sys- functions, solution of equation systems, and
tem for budgetary thinking has also changed: simplification of equations, even those with
Now the planner can opportunistically and radicals (Kunkle & Burch, 1984; Wilf, 1982).
flexibly test hypotheses in the model virtually What implications does "the disk with a college
wherever and whenever he or she wants. For education" have for how students think math-
example, any hypothesis o n relationships be- maticallv?
tween cells can be tested by modifying formulas T o take o n e example, a student using
and observing the recalculated results. muMath spends time primarily in algorithm de-
Beyond the quantitative amplification of sign and search (solution-path-finding) of ap-
efficiency -some estimate time-saving ratios of propriate operators rather than acting as a me-
80: 1b u s i n e s s planners now run vast numbers chanic for calculating numerical expressions.
of complex experiments in this cognitive play- The central role of search in using computer-
ground for playing hypothesis-comparison based mathematics problem-solving software
games, including many more variables than be- may become clearer in AlgebraLand.
fore. They also have a better understanding of Consider linear equation solving. Search is
the interdependencies of their component oper- not a central concept in algebra instruction to-
ations than before this tool was available. day, but a central insight of cognitive science is
This tool has also qualitatively changed the that learning problem-solving skills in math
organization of budgetary justification and ar- fundamentally involves search, that is, knowl-
gumentation. Electronic spreadsheets are now edge about when to select what subgoals, in
commonly used, unlike anything before, to what sequence. In most classroom instruction
quantitatively justify business decisions in in algebra equation solving, the teacher selects
group discussions by on-line comparisons with the operator to be applied to an equation (e.g.,
alternatives, and the dynamic what-if capacities add to both sides), and the student carries out
of such systems make it possible to display im- the arithmetic. The pedagogical flaw in this
mediately the consequences of different ap- method is that students d o not know when to
proaches to a problem that may be suggested select the various subgoals (Simon, 1980) when
during the meeting. solving equations alone, even if they know how
172 ROY D. PEA

to execute subgoals (e.g., to d o the arithmetic shows each operation used to transform an ex-
once the divide operation has been selected). pression. T h e Search Space Window records all .
Originally created several years ago at Xerox the solution steps the student explored. This
PARC bv. 1. S. Brown, K. Roach. and K.
>
space is represented graphically as a search tree
VanLehn, and currently being revised by C. that displays solution paths with all the back-
Foss for work with middle-school students, tracking points and problem-solving moves
AlgebraLand is an experimental system for help- made while trying to solve the equation. In this
ing students learn algebra from doing problems case, the student took three different ap-
(I. S. Brown, 198413). A picture illustrating proaches to solving this equation, reflected in
some of the features of the system to be dis- the three branches from the original equation.
cussed is dis~laved - 1.
. , in Figure Each intermediate expression that resulted
The task for the student in this figure is to from applying the Do Arithmetic operator ap-
solve the equation for N (shown in the Solve for pears in boldface for perceptual clarity.
N window on the figure's right side). Algebraic AlgrbraLand performs all the tactical, algebraic
operators listed in t h e Basic Operations operations and arithmetic calculations. Stu-
window o n t h e b o t t o m right side (e.g., dents only select the operator and its scope of
Combine-Terms, Add to Both Sides, Distrib- application, effectively eliminating errors in
ute) can be selected to apply to the whole equa- arithmetic or application of operators, and leav-
tion or to one of its subexpressions. After se- ing them free for the real mental work of search
lecting the operation and where to apply it, the
- -
and operator evaluation.
student can execute it. This creates a second al- Operators are also provided for exploring so-
gebra expression. lution paths. There is an UNDO operator that
The Record Window (upper right) shows the returns the equation to its immediately prece-
steps taken to a solution. Its left column lists all ding state and a G O T 0 operator (not on the
the intermediate expressions; its right column menu) that returns to any previous state. Stu-

Figure I . The AlgebraLand screen.

 BEYOND A M P L I F ~ C A T I O N 173

dents can also back up a solution path by ap- but students can now analyze and learn from an
plying the inverse of a forward operator (e.g., se- explicit written history of their problem-solving
lecting Divide after they have just applied moves in searching for the path of operators.
Multiply).
- . AlgebraLand, with its focus on problem-solving
Because the windows show every operator strategies as the crucial human component to
used and everv state the eauation was trans- equation solving, thus provides students with
formed into, students have valuable opportuni- the opportunity to become familiar with the
ties to learn from specific tracks of their prob- idea of search, to understand the importance of
lem solving, and they play with possibilities. search in a specific case, and to learn how to im-
They can explore the search paths of their solu- prove their search. In systems such as these, the
tion space, examine branch points on one stem tracks of the process of education can be high-
where an operation was used that led down an lighted and learned from in a dynamic, interac-
unsuccessful path, and o n another stem, try an tive way never possible through more static in-
operation started down a path toward solution. structional media (Bruner, 1960). And with the
Then they can study features of the equation at rich derivational base of possible operations
the branch point which may have originally and paths of search, new frontiers of play
recommended the optimal operation to be emerge, which could be supported by gaming
used. Guided by this hypothesis for what options.'
worked, they can test it out in future equation The consequences for math education and
solving. These learning activities are not possi- for what mathematical thought requires that
ble with traditional methods for learning to result from these new cognitive technologies are
solve equations. The cognitive technology of- remarkable: Students can and need to learn,
fered by AlgebraLand affords new opportunities among other problem-solving skills, how to
for different forms and types of learning search effectively. A n d although estimation
through problem solving that were not avail- skills are still central, error-free computation of
able in static, noncomputer-based symbolic sequences of operations on numbers and for-
technologies for solving equations. mulas is n o longer as important a mental activ-
In summary, the computer environment ity in mathematical problem solving.
AlgebraLand emphasizes a procedure diametric-
ally opposed to the traditional instructional
method described. With AlgebraLand, the stu- Example 3: Writing With Outliners and
dent chooses when to apply operators, and the Idea Organizers
comuuter carries out the mechanical wroce-
dures to transform the equation. Students are Two dramatically different kinds of com-
thus challenged by the problem of search for ~uter-basedwriting technologies have appeared
and discovery of a path of operations that will recently, each designed to better serve the
- of solv-
lead from the uroblem state to the goal exteriorization and revision of thinking proces-
ing for the unknown. Note how without the ses that written language allows (cf. Pea &
graphic availability of the search map, the Kurland, in press).
problem-solving process would be ephemeral, The first type of writing tool is the outlining
especially when a student's cognitive processes program, and it ~ r o v i d e sa rich technology for
are regularly diverted to applications to interactively creating and revising a structured,
ouerators. top-down plan of a written document. Several
Learning effective search skills in algebra commercially available examples for micro-
equation solving is not a trivial task. The cogni- computers are ThinkTank (Living Videotext)
tive technology of AlgebraLand reorganizes the and Framework (Ashton-Tate). Their essential
learning in a way that appears to highlight more property is the capacity they afford the writer of
fundamental skills to be learned-the func- portraying a n outline at different levels of detail
-
tional svstem of mathematical thinking for the without revising its contents. With this facility,
equation-solving task. Similar reorientations one can quickly flip (usually in a keystroke or
are evident in recent artificial intelligence two) between different perspectives on the doc-
tutors in the programming language LISP ument, analyze its part-whole relationships,
(Anderson & Reiser, 1985) and geometry
proofs (Boyle & Anderson, 1984). T h e required 'I am grateful t o Jerome Bruner for directing my attention
constituent operations are redirected. Calcula- ro the new playgrounds of mind we are finding and explor-
tion of arithmetical operations is eliminated, ing through t h e ludic possibilities of interactive media.
174 ROY D. PEA

and make and test revisions for their goodness workbench" planning programs; (b) interactive
of fit accordingly. Users report greater exper- computer programming, particularly in explor- .
imentation with alternate organizational atory programming environments such as
schemes and vastly more attention during cy- InterLisp-D ( S h e i l k Masinter, 1983); (c) using
cles of revisions to how the details of their text computer data-bases (including icon-based
contribute to the purpose of the whole. graphic data-base systems, e.g., Filevision for
Notecards (I. S. Brown, 1984a) is a mini- the Apple Macintosh) and graphing software as
computer tool created at Xerox PARC with a tools for exploratory data analysis, for organ-
different orientation: It encourages bottom-up izing data, and for framing and testing conjec-
discovery and definition of relationships among tures of patterns among variables in the data
ideas that the writer may initially have in mind (Conference Board of the Mathematical Sci-
only haphazardly or which d o not yield easily ences, 1983; Steen & Albers, 1981; Tufte, 1983;
to top-down structuring early in the writing White, 1981); and (d) using simulated micro-
process. Through cycles of shuffling and filing worlds to explore principles of Newtonian me-
of notecards according to categories one can de- chanics (disessa, 1983) and systems of mathe-
fine, one can progressively discover idea struc- matics (Abelson & diSessa, 1981) in intuitive
tures during writing, which are based o n ideas rather than formal terms.
collected from texts, and their annotation and Further examples, less accessible today to
linking by various relations (e.g., the rhetorical schools because they tend to run o n super-
relations of evidence, comment, argument), micros or minicomputers, but equally dramatic
which can then be reorganized into a map in their cognitive implications for reorganizing
around which text can be generated. VanLehn mental processes, are: (a) powerful simulation
(1985) has described in vivid detail his experi- programs, often incorporating highly realistic
ences with the powers of Notecards as a tool or graphics, for exploring the workings of complex
reorganizing the process of rhetorical analysis of systems, such as electrical systems (SOPHIE; J.
a complex text. He describes how, by explicitly S. Brown, Burton, & deKleer, 1982)or physical
tagging the nature of relationships between ar- plants (STEAMER; Hollan, Hutchins, &
guments and evidence with Notecards, he found Weitzman, 1984); and (b) artificial intelligence
loopholes in the intricacies of his own competi- programs such as expert systems and know-
tive argumentation for specific assumptions in ledge-based intelligent tutors. Expert systems
his highly complex artificial intelligence model ( D a v i s & L e n a t , 1981; Feigenbaum &
of learning to subtract (VanLehn, 1983). McCorduck, 1983; Hayes-Roth, Waterman, &
In both cases, structurally distinctive features Lenat, 1984) are programs that emulate reason-
of the writing technologies provided the possi- ing processes of experts in a field, which are to-
bilities for reorganizing one's writing processes day used primarily by adults to support and
and for trying out different activities during guide complex problem solving. For example,
writing. The closing of the temporal gaps be- they dovetail with the decision-making proces-
tween thought and action, between hypothesis ses of humans in medical diagnosis (Shortliffe,
and experiment, that these technologies enable 1976), design of new chemicals, computer-
and the rapid cycles of propose-test-revise that assisted design and manufacturing (e.g., Stefik
they thereby allow (much like the bases of & deKleer, 1983), industrial scheduling, and
spreadsheets and mathematics software) appear automated factories (Mason, Brady, Holler-
to have deep qualitative effects o n how problem bach, & Lozano-Perez, 1983). Knowledge-
solving occurs, which are not anticipated or based intelligent tutors (Sleeman & Brown,
captured by the amplifier metaphor. 1982) build detailed models of student under-
standing and embody in their interactions a
theory of tutoring. Available examples support
Other Examples learning in geometry (Boyle & Anderson,
1984), programming (Anderson & Reiser,
Beyond the salient cases I have described, in 1985), arithmetic (Burton & Brown, 1979), ge-
which human problem-solving activities are re- ography (Carbonell, 1970), medical diagnosis
organized and not just amplified by computer- (Clancy, 1983), electronic troubleshooting
based cognitive technologies, other noteworthy (Brown et al., 1982), and other technical fields,
examples that would admit complementary and the field of machine learning now addresses
analysis include: (a) complex planning aided by fundamental issues that dovetail the concerns
project management software and "cognitive of developmental and instructional psychology
 BEYOND AMPLIF~CAT~ON 175

(e.g., Mitchell, Utgoff, & Banerji, 1983; Van- modated to these latest cognitive technologies.
Lehn, 1983). Issues concerning the broader rele- Instead, it has assimilated the computer to its
vance of these types of cognitive technologies earlier fact-oriented agenda. For the most part,
for the future of human learning and develop- computers are not being used to extend and
ment are discussed in Pea (1985). redefine the powers of the child's intellect and
expressive powers. The prevalence of fact-
oriented computer-assisted instruction in
Summary schools today is well documented. A major rea-
son for this may be a commitment to the ampli-
In all of the cases described, computer tech- fier metaphor. With efficiency and speed in
nology has come to provide cognitive power achieving already-defined (and easily measura-
tools that improve the process of bringing ble) educational objectives as the goal, drill and
thought into communicable expressions in practice software offering more exercises in less
such significant ways that, once the tool is un- time is a logical choice. Although many educa-
derstood and used regularly, the user feels want- tors have begun working to remediate this situ-
ing if it is not available because it has opened up ation (e.g., Shavelson et al., 1984; Sheingold,
new possibilities of thought and action without Martin, & Endreweit, in press) less effort has
which one comes to feel at a disadvantage. It be- been devoted to thinking about the ways in
comes an indispensable instrument of mental- which computers can help serve as cognitive
ity, and not merely a tool (Minsky, 1983; Si- technologies to reorganize both the individual
mon, 1977). The cognitive power of a tech- mental life of students (and teachers) and the
nology is defined relative to a user's perspective: broader context of the educational environ-
What is a power tool today may be mundane to- ment.
morrow. It is therefore far more than an en- Many schools now offer or have mandated
hancement in efficiency of mental operations or courses in computer literacy-often courses
an increase in problem-solving skills that soft- about computers rather than courses about
ware may offer. The quantifiable products of using computers- and in 1984 over a million
problem solving have indeed been enhanced, as precollege students in the United States took
the amplification metaphor would lead us to computer programming, learning primarily the
observe, but the software has also restructured syntax and semantics of a specific programming
the thinking activities involved in such a major language. More rarely did they learn the
way that computer users come to discover new problem-solving and thinking skills that can be
methods of thinking about their mental tasks exemplified through the symbolic medium of
and unanticipated ways of using the tech- programming (Pea & Kurland, 1984).
nologies. Thus, there are emergent properties of What are some alternatives? Before em-
computer-aided thought that are unrecognized barking into this question, the terrain on which
when one subscribes solely to the amplifier we travel must be noted. The reorganization
metaphor. perspective, unlike the amplifier perspective, is
It is also noteworthy that, almost without ex- noncommital with respect to whether the con-
ception, the paradigm cases of cognitive tech- sequences of the reorganization of mental activ-
nologies are ones designed for and used by ities are positive or negative, developmental or
adults, who we typically suppose to have devel- regressive. Here, as in the study of "child devel-
oped the skills that become reorganized with opment," as Kaplan (1983) indicates, develop-
the tool's support. But what of the children for mental progress is not to be conflated with the
whom this assumption is not valid? This issue is march of history; "development" is a value con-
addressed shortly. cept, not a descriptive one. In contrast, the am-
plifier metaphor seems to carry with it the idea
that faster and more efficient is better-the cog-
Implications of the Reorganization nitive technology offering a means more ade-
Metaphor for Cognitive Technologies in quate to the task at hand. But this begs the
Education question that the reorganization perspective
makes problematic: What shape do we want the
How is education responding to these devel- effects of computers in education to take? For
opments? Are its methods and aims keeping example, television has opened up new global
pace with these transformations of the students' channels of visual communication and tremen-
world? Education has not, by and large, accom- dous educational potentials; at the same time,
 ROY D. PEA

some believe that this medium has hampered longer to have knowledge in one's own mem-
written language literacy because so much of ory, but to be able to effectively search for, find,
the children's time is spent listening and and use the information one needs for particu-
watching rather than engaging in literacy activ- lar purposes.
ities. And Plato's familiar critique of written There are profound consequences of this par-
language in the Phaedrus as a technology that adigm shift for what we d o in education and for
will weaken our memories makes clear the dark thinking about appropriate roles for computer-
side even of writing as a technology. We are al- based cognitive technologies. Although the
ways in a situation in which we must consider current uses of computers in education are lead-
the positive and negative outcomes of a new ing to documentable reorganizations of mental
technology. activities and of the contexts of learning, they
Thus, we must go beyond the recognition are often unproductive ones when measured
that cognitive technologies can reorganize men- against the purposes of helping students acquire
tal functioning to arguments for the specific transferable knowledge that will be useful over
ways in which they should d o so, arguments the long term. That is why we need an explicit
that are theoretically and empirically grounded layer of value analysis-which educational
in our best guesses and our best psychological goals are most central, with respect to what
analyses about what people will need to know purposes-to inform the choice and design of
during their lifetimes. cognitive technologies for education.
Education, whether formally o r informally With our predominantly fact-oriented cur-
acquired, is by its very nature a moral activity, ricula, we are hardly preparing our children for
in which choices are made to direct the paths of the lifelong learning the information age re-
learning to socially valued goals. What should quires. Regardless of our media, our aim should
be the aims of learning and development, and no longer be the hopeless task of pouring an
how can education support these processes? ocean of facts through a straw into the child's
Which of our current learning objectives- memory in hopes of the well-bucket coming up
many of them historical remnants of curricula full when it is needed. Instead, we can work to
defined in the 19th century-are n o longer val- help children learn for themselves how to seek
uable and which new ones are? There are some out, organize, and use information for different
aspects of our students' world that demand our purposes. From this orientation, education is
attention, and that appear to warrant a novel envisioned as a process of enabling independ-
approach to these issues. ent, critical, and unique thinkers to take initia-
tives individually and collab~rativel~ to pose
and solve problems, and to apply and develop
Looking to the Information Society their learning and thinking skills while accom-
plishing these tasks. This will require assem-
What we have arrived at is the question of bling a new vision for education in an age of
what education is going to be in this informa- technology that recognizes the causal powers of
tion age. Our children now live in an informa- the individual (Harre, 1984; Harre & Madden,
tion society increasingly dependent on com- 1975). It appears that knowledge of facts will
puter-stored information and knowledge, and still be useful, but as usable materials for think-
on the use of computer tools for transactions ing about events and problems and to help
with that information to try to understand and guide actions, not as ends in themselves nor as
manage its complexities. A defining feature of inert memory entries to be accessed at the time
this new society has been the information ex- of assessment and then forgotten.
plosion: Over 500 different computer data-
bases are available, and 3 million new reports
are published each year (Kerr, Braithwaite, Me- A n Explicit Cognitive Skills Emphasis Is
tropolis, & Sharp, 1984). Knowledge obsoles- Central
cence is a central in most fields, and
America's corporations spent 60 billion dollars For the reasons described, it seems that a pro-
last year reeducating their employees. In- ductive approach for cognitive technologies in
deed, the Nobel laureate Herbert Simon has education will begin to: (a) define the cognitive
pressed the point that in this information age skills children will require to be in control of
"knowing" has become redefined as a verb of ac- their own learning and information manage-
cess rather than possession. T o know is no ment and (b) design and create new tech-
 BEYOND AMPLIFICATION 177

nologies to help support the attainment and use A skills-oriented approach does not mean,
of these skills. The learning of such skills would however, as some recent thinking skills
thus become an explicit rather than a tacit ob- programs presuppose (e.g., deBono, 1985;
jective of education. Whimbey & Lochhead, 1980), that these skills
Among other aims that we see as central in can be effectively taught (i.e., for subsequent
the forms of information literacy called for to- use) devoid of detailed emphases on domain-
day are: specific applications. Surely method without
content is blind. Schoenfeld's research (1985a)
1. A new emphasis o n cognitive skills of in- on teaching and student learning of general
formation management (Hawkins, heuristics such as "draw a diagram" for mathe-
Mawby, & Ghitman, in press), including matics problem solving makes this point
problem posing/question definition (S. I. clearly: O n e must ask what kind of diagram.
Brown & Walter, 1983), flexible strategies Similarly, Soloway (in press) demonstrates the
for information retrieval, information centrality of domain-specific knowledge for
schematization and inference, textual learning general problem-solving heuristics for
summarization and intertextual inte- writing Pascal computer programs, such as
gration. "break the problem into parts" (Descartes' "di-
2. A renewed emphasis o n written commu- vide and conquer" heuristic). He finds that
nication and critical inquiry skills (e.g., without prior experience in solving problems in
evaluation of sources of information and that domain one cannot identify the sub-
claims to knowledge). ~ r o b l e m that
s it makes sense to break the prob-
3. Metacognitive and self-regulatory skills lem up into! The application of the general heu-
(A. L. Brown, 1978) such as planning ristic needs to be guided by its prior historical
ahead, comprehension monitoring (Wag- applications in the specific knowledge domain
oner, 1983), cognitive resource manage- under consideration.
ment or control (Schoenfeld, 1985b), and Thus, it appears that general skills can be an
learning how to learn (Dansereau, 1985; instructional emphasis, but that they must be
Weinstein & Underwood, 1985). learned through content-driven examples (cf.
4. Strategies for creative thinking and prob- A. L. Brown, 1985; Glaser, 1984). It seems very
lem solving (e.g., brainstorming; problem likely that effective computational tools can be
decomposition; and proposing, testing, devised for learning and practicing such skills
and debugging approaches to a problem) through problem solving across different con-
and systematic decision-making- methods tent domains.
(e.g., decompositional approaches t o
comparing utilities of choices, e.g., cost-
benefit analysis) that crosscut knowledge Software Needed for Promoting
domains. Transferable Cognitive Skills
5. Cooperative group problem solving
(Slavin et al., 1985) and negotiation skills. M a n y forward-looking educators a n d
schools have begun to use the thinking tools
Why are these types of skills important?They used by adults to solve problems in such disci-
are important because they appear to character- lines as business, history, math, and science-
ize the cognitive performances of expert prob- software for graphing, data-base management,
lem solvers in a great variety of disciplines, as word-processing, and spreadsheet software.
the artificial intelligence literature and cog- The difficulties of integrating adult versions of
nitive science studies attest (e.g., Barr & these tools (designed for different users and dif-
Feigenbaum, 1982; Brown, Bransford, Ferrara, ferent purposes) into the curriculum have come
& Campione, 1983; Greeno & Simon, in press) to be realized. Versions of these tools specifi-
and because they are high-yield skills that cally designed for children have begun to ap-
promise utility throughout the life span, unlike pear, including the widely used Bank Street
a traditional fact-oriented curriculum (Boyer, Writer (word-processing program; Kurland, in
1983). These broad families of skills also cross- press) and the Quill writing system (Rubin &
cut the too often segregated domains of the tra- Bruce, in press).
ditional curriculum, and we would hope that For example, in school studies conducted by
new cognitive technologies developed to sup- Char and colleagues at our center (Char, Free-
port them could be used throughout schooling. man, & Hawkins, 1985; Hawkins, Char, &
178 ROY D PEA

Freeman, 1984),it has been found that the pow- specific domain of decision-making-family ,
erful information-handling tools provided by chore planning-is used to introduce generali-
data-base management programs require new zable aspects of systematic decision-making
skills-in problem definition, planning for skills (e.g., goal monitoring, constraint plan-
searches of the data-base fields, and so o n - that ning, defining the space of alternative choices;
middle school children have not yet acquired, analysis of attributes of alternatives; plan evalu-
and which even highly creative teachers who ation and monitoring). Multiple examples of
deeply value critical inquiry and information t h e application of each targeted general
literacy are unsure how to teach. How can decision-making method are provided by the
technologies for education serve not only as software. In this way, the learner can at any
tools for thinking, but as tools for helping time explore or be guided to learning generally
thinking skills develop? useful aspects of methods they are learning to
There are currently no computer techno- apply in this specific case. We believe that by
logies that tutor the development of thinking combining the functions of a domain-specific
and metacognitive skills important for lifelong problem-solving tool with those of a general
learning and problem solving. Although curric- thinking skills coach, an effective program for
ula for the teaching of thinking and problem- learning complex thinking skills will emerge.
solving skills, such as those of Venezuela's Proj-
ect Intelligence (Herrnstein, Nickerson, de
Sanchez, & Swets, 1983), which was developed Toward an Activist Research Paradigm in
with the assistance of Harvard University and Educational Technologies
Bolt, Beranek & Newman, have proliferated in
the last 5 years (see reviews in Nickerson, Per- What implications may be drawn from these
kins, & Smith, 1985; Segal, Chipman, & considerations for how we do research in the
Glaser, 1985), we find no computer-based sys- area of educational technologies? As we con-
tems for achieving these aims. sider these issues, it is worthwhile recognizing
Several projects under way at Bank Street that we are not meeting an earlier problem, for
may contribute to visions of what is possible. In it has never been possible for education to be so
one, we are building and testing software tools outdated before. We need tools to achieve new
for helping children learn to engage in critical aims for information age education that are an
inquiry and construct a personal perspective order of magnitude more obviously effective
about various topics, particularly in science, than those we use today-the educational
throughout the curriculum. In a second proj- equivalent of the automobile, the light bulb, or
ect, we are building and testing a software envi- the television. Each demanded revolutionary
ronment to encourage the development and use changes in existing social, cultural, and eco-
of systematic decision-making skills, including nomic conditions; each led to virtually univer-
problem definition, analysis of alternatives, sal acceptance and to new, unimagined uses,
evaluating attributes of alternatives, and vari- reshaping human activities in consequence.
ous heuristics for comparing choices. Para- How can we get there?
mount in each case is the creation of effective As we d o our work today, the research cycle
and enjoyable tools for learning through doing, that leads from research proposal planning,
and student understanding of how to proceed writing, and funding to research activities, anal-
that will transcend the specific problem domain yses, writeup, review, and publication is
under study. Our belief is that if we create use- roughly 3 to 5 years. Then, of course, the
ful tools for thinking in these ways, the new vi- pipeline model for the impact of research o n ed-
sions of education described earlier will at least ucation, with the assumed basic research/
be possible because they are technically applied research dichotomy, includes the extra
feasible. step of translating the findings of the research
What we believe may be required are cogni- into educational practices, a difficult path to
tive technologies for education that embody an travel and one that often is not taken (Husen &
explicit knowledge transfer architecture, that Kogan, 1984). My conclusion is that there is not
emphasizes transfer activities in their very enough time and the ~ r o b l e m sare too impor-
structure. We are exploring this design ap- tant for us to replicate this research model in
proach in a current research and development our studies of educational technologies. We bet
project on cognitive skills. In the design of on the irrelevance of our work if we rely on off-
IDEA (Integrated Decision Envisioning Aid), a the-shelf software a n d limit ourselves t o
 BEYOND AMPLIFICATION 179

describing what happens when it is introduced Revolution: "the primary waste is [not money
to the classroom. or things- but1 human life, the life of the chil-
What more is needed? We need to design and dren while they are at school, and afterward be-
engineer environments for the transferable cause of inadequate and perverted preparation"
learning that an information age requires. (p. 59). Much the same applies today as we try
More specifically, to inform education effect- to reshape an education for the information
ively, theory and practice will need to be unified age.
through the invention of research-informed As in Dewev's davs,, , we are in need of funda-
electronic learning systems that work in educa- mental change, guided by research on student
tional settings. As Greeno (1985) has recently learning with emerging cognitive technologies
argued, "important advances in instructional and by communal dialogues about redefining
technology and in basic cognitive science will educational aims. Everyone is a stakeholder in
occur as an integrated activity" (p. 2). Research this reformative enterprise: Teachers, parents,
and development activities can be united in the researchers, industry and business, and policy
creation of educational software prototypes, -
makers all stand to gain or to lose. Working to- -
which are designed a n d built b y inter- gether to shape the technologies that will reor-
disciplinary teams of researchers, educators, ganize human thinking, we may be able to cre-
and developers, and progressively modified in ate a new svstem of education that resDects the
response to formative testing with students. creative spirit and flexibility of the human intel-
These prototypes can provide sophisticated lect. that builds on and discovers new worlds of
learning environments for students and simul- cognition, action, and play made possible by
taneously serve as research tools for deter- the remarkable symbolic powers of computers,
mining how skills and knowledge develop with and that yields resilient thinkers and actors,
these new cognitive technologies. I would argue readv to meet future worlds more radically dif-
that these technologies can serve as the educa- ferent than we can now even begin to imagine.
tional infrastructure linking psychological re-
search to educational practice, which Cham- References
pagne and Chaiklin (1985) suggest is necessary Abelson, H., & diSessa, A. A. (1981). Turtle geometry: The
for cognitive science studies to have significant computer as medium for exploring mathrmutics. Cam-
classroom applicability. bridge, MA: MIT Press.
Anderson, J. R., & Relser, B. (1985). The LISP tutor. Byte,
10, 159-175.
Coda Barr, A., & Feigenbaum, E. A. (Eds.). (1982). The handbook
of urtlficial intelligence (Vol. 2). Los Altos, CA: William
It may appear strange that I have primarily Kaufmann.
highlighted the potential positive effects of Boyer, E. L. (1983). High school: A report on secondarj educa-
tion in America. New York: Harper &Row.
computers as reorganizers of mental func-
Doyle, C . F., & Anderson, J. R. (1984, April). Acquis~tion
tioning. but in the absence of prototypes guided und auromuted instruction of geometry proof skills. Paper
by positive visions of what could be, it is un- presented at the annual meeting of the American Edu-
likely that we will ever learn what our educa- cational Research Association, New Orleans, LA.
tion can become. Just as the child needs tools to Brown, A. L. (1978). Knowing when, where, and how to re-
member: A problem of metacognition. In R. Glaser
think with (Papert, 1980) as he or she learns to (Ed.), Advances in instrucrionul psychology (Vol. 1, pp.
define and solve problems, so d o we as we work 77-165). Hillsdale, NJ: Lawrence Erlbaum Associates,
to reshape the aims and methods of a computer- Inc.
enhanced education responsive to the chal- Brown, A . L. (1985). Mental orthopedics, the training of
cognitive skills: A n interview with Alfred Binet. In S. F.
lenges of an information society. We need to
Chipman, J. W. Segal, & R. Glaser (Eds.), Thinking and
create a plurality of concrete prototypes of elec- learning skills: Current research and open questions (Vol. 2,
tronic learning environments to work with, pp. 119-337). Hillsdale, NJ: Lawrence Erlbaum Associ-
whose effects positive and negative, can be em- ates, Inc.
pirically examined, reshaped, reassessed, and Brown, A . L., Bransford, J. D., Ferrara, R. A., & Camp-
ione, J. C . (1983). Learning, remembering, and under-
debated, rather than the armchair-inspired cri- standing. In J. H. Flavell & E. M. Markman (Eds.), Cog-
tiques of computers in education that have nitive development (Vol. 3, pp. 77-166), of P. H. Mussen
tended to overemphasize the future representa- (Ed.), Handbook of child psychology (4th ed.). New York:
tiveness of current software (Sloan, 1985). Wiley.
Brown, J. S. (1984a). Idea amplifiers: New kmds of electronic
Seventy years ago, John Dewey (1915) criti-
learning environments. Palo Alto, CA: Xerox Pa10 Alto
cized an American education that had yet to Research Center, Intelligent Systems Laboratory.
adapt to the changes wrought by the Industrial Brown, J. S. (1984b). Process versus product: A perspective
180 ROY D. PEA

o n tools for communal and informal electronic learn- Dewey, J. (1915). The school and society (rev. ed.). Chicago:
ing. In S. Newman & E. Poor (Eds.), Report from the University of Chicago Press. (Original work published
learning lab: Educat~onin the electronic age (pp. 41-57). 1900)
New York: Educational Broadcasting Corporation, Dilthey, W. (1976). T h e construction of the historical world
WNET. in the human studies. In H. P. Rickman (Ed. & Trans.),
Brown, J. S., Burton, R., & deKleer, J. (1982). Pedagogical, W. Dilthey, Selected writings (pp. 168-245). New York:
natural language, and knowledge engineering tech- Cambridge University Press.
niques in SOPHIE 1, 11, and 111. In D. Sleeman & J. S. diSessa, A . A . (1983). Phenomenology and the evolution of
Brown (Eds.), Intell~gent tutoring systems (pp. 230-282). intuit~[Link] D. Gentner & A. Stevens (Eds.), Mental
New York: Academic. models (pp. 15-33). Hillsdale, NJ: Lawrence Erlbaum
Brown, S. I., & Walter, M. 1. (1983). The art of problem Associates, Inc.
posing. Phdadelphia, PA: Franklin Institute Press. Feigenbaum, E. A,, & McCorduck, P. (1983). The fifth gen-
Bruner, J. S. (1960). The process of education. Cambridge, eration: Artifmal intelligence and Japan's computer chal-
MA: Harvard University Press. lenge to the world. Reading, MA: Addison-Wesley.
Bruner,]. S. (1966). O n cognitive growth 11. In J. S. Bruner, Fey, J. T . (Ed.). (1984). Computing and mathematics: The im-
R. R. Olver, & P. M . Greenfield (Eds.), Studres in cogni- pact on secondary school curr~[Link] Park, MD: Na-
tive growth (pp. 30-67). New York: Wiley. tional Council of Teachers of Mathematics.
Burton, R. R., & Brown, J. S. (1979). A n invest~gationof Freeman, C., Hawkins, I., & Char, C. (1984). Information
computer coaching for informal learning activities. In- management tools for the classroom: Exploring database
ternational Journal of Man-Machine Studies, 11, 5-24. management systems (Tech. Rep. No. 28). New York:
Carbonell, J. R. (1970). A1 in CAI: A n artificial intelligence Rank Street College of Education, Center for Children
approach to computer-aided instruction. IEEE Transac- and Technology.
tions on Man-Muchme Systems, MMS-I 1(4), 190-202. Glaser, R. (1984). Education and thinking: T h e role of
Cassirer, E. (1910). Substance and functlon and Einstern's knowledge. American Psycholog~st,39, 93-104.
theory of relativity. Mineola, NY: Dover. Goody, J. (1977). The domestication of the savage mind. New
Cassirer, E. (1957). The philosophy of symbolic forms. Vol. 3: York: Cambridge University Press.
the phenomenology of knowledge. New Haven, C T : Yale Gould, S. J. (1980a). Senseless signs of history. In S. J.
University Press. Gould, The panda's thumb: More reflections in natural his-
Champagne, A. B., & Chaiklin, S. (1985, April). Inter- tory (pp. 27-34). New York: Norton.
disciplina? research in science learning: Great prospects and Gould, S. J. (1980b). Shades of Lamarck. In S. J. Gould
dim future. Paper presented at the annual meeting of the (Ed.), The panda's thumb: More reflections in natural his-
American Educational Research Association, Chicago, tory (pp. 76-84). New York: Norton.
IL. Greenfield, P. (1972). Oral and written language: T h e con-
C h a r , C., Freeman, C., & Hawkins, J. (1985, April). In- sequences for cognitive development in Africa, the
corporating database software Into the classroom context: An United States, and England. Language and Speech, 15,
ethnographic study. Paper presented at the annual meet- 169-178.
ing of the American Educational Research Association, Greeno, J. G . (1985, April). Advancing cognitive science
Chicago, IL. through development of udvanced instructional systems. Pa-
C h a r , C., Hawkins, J., Wootten, J., Sheingold, K., &Rob- per presented at the annual meeting of the American
erts, T . (1983). "Voyage of the Mimi": Classroom case stud- Educational Research Association, Chicago, 1L.
ies of software, vtdeo, and print materials. New York: Rank Greeno, J. G., & Simon, H. A. (in press). Problem solving
Street College of Education, Center for Children and and reasoning. In R. C. Atkinson, R. Herrnstein, G.
Technology. Lindzey, & R. D. Luce (Eds.), Stevens'handbook of experi-
Clancy, W. J. (1983). G U I D O N (Tech. Rep. No. S T A N - mental psychology (rev. ed.). New York: Wiley.
CS-83997). Stanford, C A : Stanford University, De- Harre, R. (1984). Personal being. Cambridge, MA: Harvard
partment of Computer Science. University Press.
Clement, C . (1984, August). Analogical reasoning and learn- Harre, R., &Madden, E. H. (1975). Causal powers: A theory
ing computer programming. Paper presented at the Har- of natural necessity. Oxford, England: Basil Blackwell.
vard University Conference o n Thinking, Cambridge, Havelock, E. (1973). Prologue to Greek h e r m y Cincinnati:
MA. University of Oklahoma Press for the University of
Cole, M., &Griffin, P. (1980). Cultural amplifiers recons~d- Cincinnati Press.
ered. In D. R. Olson (Ed.), The social foundations of lan- Havelock, E. (1978). The Greek concept of justice. Cam-
guage and thought: Essays in honor ofJerome S. Bruner (pp. bridge, MA: Harvard University Press.
343-364). New York: Norton. Hawkins, J. (1985). The interpretation of Logo in practice
Conference Board of the Mathematical Sciences. (1983). (Tech. Rep. No. 34). New York: Bank Street College of
The mathematical sczences a ~ r r ~ c u l uK-12:
m What 1s still Education, Center for Children and Technology.
fundamental and what is not. Washington, D C : Author. Hawkins, J., C h a r , C., & Freeman, C . (1984). Software tools
Dansereau, D. F. (1985). Learning strategy research. In 1. in the classroom (Final Report t o the Carnegie Corp-
Segal, S. Chipman, & R. Glaser (Eds.), Think~ngand oration).
learning skills: Relating instruction to basic research (Vol. 1, Hawkins, J., Mawby, R., & Ghitman, M. (in press). Prac-
pp. 209-239). Hillsdale, NJ: Lawrence Erlbaum Associ- tices of novlces and experts in critical inquiry. In R. D.
ates, Inc. Pea & K. S. Sheingold (Eds.), Children and microcom-
Davis, R., &Lenat, D. B. (1981). Knowledge-based systems in puters: Theory, research, and development from Bank
artificial intelligence. New York: McGraw-Hill. Street's Center for Children and Technology. Norwood, NJ:
DeBono, E. (1985). T h e C o R T thinking program. In J. W. Ablex.
Segal, S. F. Chipman, & R. Glaser (Eds.), Thinking and Hawkins, J., & Sheingold, K. (in press). T h e beginning of a
learnmg skills: Relating instruction to basic research (Vol. 1, story: Computers and the organization of learning in
pp. 363-388). Hillsdale, NJ: Lawrence Erlbaum Associ- classrooms. In J. Culbertson & L. L. Cunningham
ates, Inc. (Eds.), Uses of mrcrocomputers and other technology in edu-
I BEYOND AhlPLlFlCATION 181

cation. Chicago: University of Chicago Press. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Hayes-Roth, F., Waterman, D., & Lenat, D. (Eds.). (1984). Minsky, M. (1983). Why people think computers can't.
Building expert systems. Reading, MA: Addison-Wesley. Technology Rec'ieur, 86, 65-81.
Herrnstein, R. I., Nickerson, R. S., d e Sanchez, M., & Mitchell, T. M., Utgoff, P. E., & Banerji, R. (1983). Learn-
Swets, J. A. (Eds.). (1983). Project Intelligence: The devel- ing by experimentation: Acquiring and refining
opment of procedures to enhance thinking skills. problem-solving heuristics. In R. S. Michalski, J. G.
Cambridge, MA: Bolt, Beranek & Newman, Inc. Carbonell, & T . M. Mitchell (Eds.), Machine learning:
Hollan, J. D., Hutchins, E., & Weitzman, L. (1984). An artificial intelligence approach (pp. 163-189). Palo
STEAMER: A n interactive inspectable simulation- Alto, C A : Tioga Press.
based training system. A1 Magazine, 5, 15-27. National Science Board Commission o n Precollege Educa-
Husen, T., & Kogan, M. (Eds.). (1984). Educational resarch tion in Mathematics, Science, and Technology. (1983).
and policy: How do they relate? Oxford, England: Educating Americans for the 21st century. Washington,
Pergamon. DC: National Science Foundation.
Kaplan, B. (1983). Genetic dramatism. In S. Wapner & B. Nickerson, R. S., Perkins, D. N., & Smith, E. E. (1985).
Kaplan (Eds.), Toward a holistic developmental psychology Teaching thinking. Hillsdale, N]: Lawrence Erlbaum As-
(pp. 53-74). Hillsdale, NJ: Lawrence Erlbaum Associ- sociates, Inc.
ates, Inc. Olson, D. R. (1976). Culture, technology, and intellect. In
Kerr, D. M., Braithwaite, K., Metropolis, N., &Sharp, D. L. B. Resnick (Ed.), The nature of intelligence (pp.
H . (Eds.). (1984). Science, computers, and the information 189-202). Hillsdale, NJ: Lawrence Erlbaum Associates,
onslaught. New York: Academic. Inc.
Kessel, F. S., & Siegel, A. W. (Eds.). (1983). The child and Olson, D. R. (1977). From utterance to text: T h e bias of lan-
other cultural inventions. New York: Praeger. guage in speech and writing. Harvard Educational Re-
Kessen, W. (1979). T h e American child and other cultural view, 47, 257-281.
inventions. American Psychologist, 34, 815-820. Olson, D. R. (1985). Computers as tools of the intellect. Ed-
Kline, M. (1972). Mathematical thought from ancient to mod- ucational Researcher, 14, 5-8.
ern rimes. Oxford: Oxford University Press. Olson, D. R., & Bruner, J. S. (1974). Learning through ex-
Kunkle, D., & Burch, C . I., Jr. (1984). Symbolic computer perience and learning through media. In D. R. Olson
algebra: The classroom computer takes a quantum leap. (Ed.), Media and symbols: The forms of expression, commu-
Mathematics Teacher, 77, 209-2 14. nication, and education (pp. 125-150). Chicago: Univer-
Kurland, D. M. (Ed.). (1984). Symposium: Developmental sity of Chicago Press.
studies of computer programming skills (Tech. Rep. No. Ong, W. J. (1971). Rhetoric, romance, and technology: Studies
29). New York: Bank Street College of Education, Cen- in the interaction of expression and culture. Ithaca, NY:
ter for Children and Technology. Cornell University Press.
Kurland, D. M. (in press). Designing a word processor for Ong, W. J. (1982). Orality and literacy: The technologizing of
children: The Bank Street Writer. In R. D. Pea & K. S. the word. New York: Methuen.
Sheingold (Eds.), Children and microcomputers: Theory, Papert, S. (1980). Mindstorms: Children, computers, and pow
research, and development from Bank Street's Center for erful ideas. New York: Basic Books.
Children and Technology. Norwood, NJ: Ablex. Pea, R. D. (1985). Integrating human and computer intelli-
Kurland, D. M., Clement, C., Mawby, R., &Pea, R. D. (in gence. In E. L. Klein (Ed.), New directions for child devel-
press). Mapping the cognitive demands of learning to opment. No. 8: Children and computers (pp. 75-96). San
program. In 1. Bishop, J. Lochhead, & D. N. Perkins Francisco, C A : Jossey-Bass.
(Eds.), Thinking. Hillsdale, NJ: Lawrence Erlbaum Asso- Pea, R. D., & Kurland, D. M. (1984). O n the cognitive ef-
ciates, Inc. fects of learning computer programming. Neu, Ideas in
Kurland, D. M., & Pea, R. D. (1985). Children's mental Psychology, 2, 137-168.
models of recursive Logo programs. Journal of Educa- Pea, R. D., & Kurland, D . M. (in press). Cognitive tech-
tional Computing Research 1 , 235-243. nologies for writing. In E. Rothkopf (Ed.), Review of re-
Langley, P. (1985). Learning to search: From weak methods search in education (Vol. 13). Washington, DC: Ameri-
to domain-specific heuristics. Cognitive Science, 9, can Educational Research Association.
2 17-260. Pea, R. D., Kurland, D. M., & Hawkins, J. (1985). Logo and
Levy, S. (1984, November). A spreadsheet way of knowl- the development of thinking skills. In M. C h e n & W.
edge. Harper's, pp. 68-64. Paisley (Eds.), Children and microcomputers: Formative
Luria, A. R. (1976). Cognitive development: Its cultural and so- studies (pp. 193-212). Beverly Hills, CA: Sage.
cial foundations. Cambridge, M A : Harvard University Rubin, A. D., & Bruce, B. C . (in press). QUILL: Reading
Press. and writing with a microcomputer. In B. A. Hutson
Luria, A. R. (1979). The making ofmind: A personal account of (Ed.), Advances in reading and language research. Green-
Soviet psychology. Cambridge, MA: Harvard University wich, CT: JAI Press.
Press. Schoenfeld, A. (1985a). Mathemarical problem solcing. New
Mason, M. T., Brady, 1. M., Hollerbach, T. J., & Lozano- York: Academic.
Perez, T. (1983). Robot motion: Planning and control. Schoenfeld, A. (1985b). Metacognitive and epistemological
Cambridge, MA: MIT Press. Issues in mathematical understanding. In E. Silver (Ed.),
Maurer, S. B. (1984a). College entrance mathematics in the Teaching and learning mathematical problem solving (pp.
year 2000. Mathematics Teacher, 77, 422-428. 361-379). Hillsdale, NJ: Lawrence Erlbaum Associates,
Maurer, S. B. (198413). T h e effects of a new college mathe- Inc.
matics curriculum o n high school mathematics. In A . Scribner, S., & Cole, M. (1981). The psychology of literacy.
Ralston & G. S. Young (Eds.), The future of college math- Cambridge, MA: Harvard University Press.
ematics ( pp. 153-175). New York: Springer-Verlag. Segal, J. W., Chipman, S. F., & Glaser, R. (Eds.). (1985).
Mawby, R. (in press). Proficiency conditions for the devel- Thinking and learning skills: Relating instruction to basic re-
opment of thinking skills through programming. In J. search (Vol. 1). Hillsdale, N]: Lawrence Erlbaum Asso-
Bishop, J. Lochhead, & D. N. Perkins (Eds.), Thinking. ciates, Inc.
182 ROY D. PEA 7

Shavelson, R. J., Winkler, J. D., Stasz, C . , Feibel, W., Stefik, M. J., & de Kleer, J. (1983, April 21). Prospects for
.
Robyn, A. E., & Shaha, S. (1984). Teaching mathematics expert systems in C A D . Computer Design, pp. 65-76. '
and science: Patterns of minocomputer use (Tech. Rep. No. Tufte, E. R. (1983). The visual display of quantitative informa-
R-3180-NIE/RC). Sanra Monica, C A : Rand Corp- tion. Cheshire, CT: Graphics Press.
oration. VanLehn, K. (1983). Felicity conditions for human skill acqui-
Sheil, B. A,, & Masinter, L. M. (Eds.). (1983). Papers on sition: Validating an Al-based theory (Tech. Rep. No.
InterlispD. Palo Alto, C A : Xerox Palo Alto Research CIS-21). Palo Alto, C A : Xerox PARC.
Center. VanLehn, K. (1985). Theory reformulation caused by an argu-
Sheingold, K., Hawkins, J., & Char, C . (1984). "I'm the mentation tool. Palo Alto, CA: Xerox PARC.
thinkist, you're the typist": T h e interaction of technol- Vygotsky, L. S. (1962). Thought and language. Cambridge,
ogy and the social life of classrooms. Journal of Social Is- MA: MIT Press.
sues, 40, 49-6 1. Vygotsky, L. S. (1978). Mind in society: The development of
Sheingold, K., Martin, L. M. W., & Endreweit, M. E. (in the higher psychological processes. Cambridge, MA:
press). Preparing urban teachers for the technological Harvard University Press.
future. In R. D. Pea & K. Sheingold (Eds.), Children and Wagoner, S. (1983). Comprehension monitoring: What it is
microcomputers: Theory, research, and development from and what we know about it. Reading Research Quarterly,
Bank Street's Center for Children and Technology. 18, 328-346.
Norwood, NJ: Ablex. Wartofsky, M. (1983). T h e child's construction of the world
Shortliffe, E. H. (1976). Computer-based clinical therapeutics: and the world's construction of the child: From histor-
MYCJN. New York: American Elsevier. ical epistemology to historical psychology. In F. S.
Simon, H. A. (1977). What computers mean for man and Kessel & A. W. Sigel (Eds.), The child and other cultural
society. Science, 195, 1186-1 191. inventions (pp. 188-215). New York: Praeger.
Simon, H. A. (1980). Problem solving and education. In D. Weinstein, C . E., & Underwood, V. L. (1985). Learning
Tuma & F. Reif (Eds.), Problem solving and education (pp. strategies: T h e how of learning. In J. Segal, S. Chipman,
81-96). Hillsdale, NJ: Lawrence Erlbaum Associates, & R. Glaser (Eds.), Thinking and learning skills: Relating
Inc. instruction to basic research (Vol. 1, pp. 241-258).
Slavin, R., Sharan, S., Kagan, S. Hertz-Lazarowitz, R., Hillsdale, Nj: Lawrence Erlbaum Associates, Inc.
Webb, C., & Schmuck, R. (1985). Learning to cooperate, Whimbey, A , , & Lochhead, J. (1980). Problem solving and
cooperating to learn. New York: Plenum. comprehension: A short course in analytical reasoning (2nd
Sleeman, D., & Brown, J. S. (Eds.). (1982). Jntelligent ed.). Philadelphia: Franklin Institute Press.
tutoring systems. New York: Academic. White, S. (1983). Psychology as a moral science. In F. S.
Sloan, D. (Ed.). (1985). The computer in education: A critical Kessel & A. W. Sigel (Eds.), The child and other cultural
perspective. New York: Teachers College Press, Colum- inventions (pp. 1-25). New York: Praeger.
bia University. White, S. (1981). T h e new liberal arts. In J. D. Koerner
Soloway, E. (in press): Kids should learn to program. (Ed.), The new liberal arts (pp. 1-12). New York: Alfred
Harvard Educational Reuieu,. P. Sloan Foundation.
Steen, L. A,, & Albers, D. 1. (Eds.). (1981). Teaching Wilf, H. S. (1982). T h e disk with the college education.
teachers, teaching students. Boston: Birkhauser. American Mathematical Monthly, 89, 4-8.