THE PSYCHOLOGY OF THINKING, ANIMAL PSYCHOLOGY, AND

THE YOUNG KARL POPPER

MICHEL TER HARK

In the 1920s, Karl Popper wrote two large manuscripts on psychology that he never pub-
lished. In his autobiography, Unended Quest, he attempts to reduce the importance of his
work in psychology as much as possible, and in his philosophical work he is an antipsy-
chologist. However, in this article, it is argued that Popper’s early psychology has been piv-
otally important for the development of his philosophy. In particular, it is shown that
Popper’s views on psychology underwent a radical shift, one that paved the way for his
characteristic deductive stance in philosophy. Popper’s views shifted from an inductive and
associationistic psychology toward a noninductive psychology of problem solving. Tracing
the historical background of Popper’s early work reveals how he integrated various parts
of the psychology of Karl Groos into his analysis of the childish phenomenon of dogmatic
thinking and how he shortly after appropriated various elements of the animal psychology
of Hans Volkelt and Herbert Jennings in his biological approach to (dogmatic) thinking. In
the monumental works of Otto Selz, however, Popper finally found the roots of a nonin-
ductive and biological approach to the growth of individual and scientific knowledge.
© 2004 Wiley Periodicals, Inc.

Although Sir Karl Popper is considered one of the most influential philosophers of the

twentieth century, little is known about his early years. In his autobiography, Unended Quest
(1974), he tells us that he started his career studying and teaching pedagogy at the Pedagogic
Institute in Vienna. He also attended lectures of Karl Bühler, formerly related to the so-called
Würzburg School of Denkpsychologie, at the Psychology Institute in Vienna. He learned more
from Bühler than from any other teacher and in 1928 he wrote his PhD dissertation in the psy-
chology of thinking (Denkpsychologie) under Bühler’s supervision. Soon he found that some
of his results in psychology had been anticipated by Bühler and by a psychologist loosely re-
lated to the Würzburg School, Otto Selz.

By his own account, this discovery was one of the reasons why he abandoned psychol-

ogy and turned to philosophy. The shift would be quite radical, for in his first major work, The
Logic of Scientific Discovery (1934/1959), the psychology of scientific discovery would be
rigorously dispelled from the area of the philosophy of science. Yet in his work from the 1960s
onward, Karl Popper increasingly began to organize his wide-ranging interests in philosophy
of science, epistemology, and metaphysics around the idea that we acquire knowledge by a
process of trial-and-error elimination. This method, Popper contended, is essentially a three-
stage model, which he took to apply to animal learning as well as to the upper reaches of sci-
entific research: forming a problem or expectation, trying out a number of solutions of the
problem, and eliminating or discarding false solutions as erroneous. A key feature of Popper’s
theory of trial-and-error elimination is its insistence on problems or expectations taking
precedence over observations. The place accorded to sense perceptions in the empiricist tra-
dition is reversed, for rather than being the origin of knowledge, their role is limited to the

Journal of the History of the Behavioral Sciences, Vol. 40(4), 375–392 Fall 2004
Published online in Wiley Interscience (www.interscience.wiley.com). DOI 10.1002 /jhbs.20024
© 2004 Wiley Periodicals, Inc.

M

ICHEL TER

H

ARK

is a professor in the history of philosophy and its relation to the development of

the behavioral and cognitive sciences at the University of Groningen in the Netherlands. He has pub-
lished various articles on the history of Wittgenstein’s philosophy of psychology with special emphasis
on Gestalt psychology and William James (Beyond the Inner and the Outer, Kluwer Academic
Publishers, 1990). He has also published articles on Karl Popper’s roots in psychology and has just fin-
ished a book on Popper and psychology: Popper, Otto Selz and the Rise of Evolutionary Epistemology
(Cambridge University Press, 2004).

375

second and especially the third stage. Observations are always preceded by expectations,
points of view, questions, or problems, which, as a “searchlight,” illuminate a certain area,
thereby enabling the organism or the scientist to know what to observe in the first place
(Popper, 1945, p. 260).

1

The “bucket theory” of knowledge of the famous British philosophers

Locke, Berkeley, and Hume (and of a host of experimentalists in the pioneering era of psy-
chology), Popper concludes, is a myth. Evidently relishing his position as the opponent, he
triumphantly claimed: “My theory of knowledge is thus quite revolutionary: it overturns
everything my predecessors have said up to now. We are active, we are constantly testing
things out, constantly working with the method of trial and error” (Popper, 1999, p. 53).

But was Popper’s theory really as revolutionary as he claimed it to be? Largely drawing

on his still unpublished manuscripts in psychology and pedagogy from the 1920s, I argue that
Popper borrowed the key elements of his “revolutionary” epistemology from two different yet
related fields of scientific inquiry: the animal psychology and theory of emergent evolution
of C. Lloyd Morgan and Herbert Jennings on the one hand and the psychology of thinking of
his supervisor in psychology, Karl Bühler, and the largely ignored but pivotally important
Otto Selz. As regards the influence of the latter on Popper, I published the first in-depth study
more than ten years ago (ter Hark, 1993). An important impetus for this article has been the
pioneering little book on Popper’s psychology of learning by William Berkson and John
Wettersten (1984). Selz is briefly discussed in their book, but he is not presented as a figure
read by the young Popper but simply as one important figure at the time bearing some resem-
blance to the problems Popper wrestled with. Berkson and Wettersten even reached a conclu-
sion opposite to mine: “Popper’s and Selz’s views are different” (Berkson & Wettersten, 1984,
p. 10). In Wettersten (1992), the historical background of Popper’s discussions on psychology
and methodology is significantly broadened, dating back to the 1830s, but little attention is
paid either to Popper’s unpublished psychological works or their background. Further, al-
though Selz’s importance is mentioned, it is not fully developed, and Selz does not get a sep-
arate section. In his recent biography of Popper, Malachi Hacohen (2000) devotes a chapter
to Popper’s psychological works, but he fails to discuss the broader context in German psy-
chology and biology, thereby giving the young Popper more credit than he deserves.

In my own recent book, Popper, Otto Selz and the Rise of Evolutionary Epistemology (ter

Hark, 2004), I have attempted both to trace the origins of Popper’s epistemology back to early
German psychology, in particular the work of Selz, and to point out the way the work of the
latter diverges from the Würzburg School of Psychology. In this article, special emphasis will
also be given to other influences besides Selz, especially the biologist Herbert Jennings and
the biologically oriented psychology of Karl Groos and Hans Volkelt. The importance of this
historical account is twofold. In the first place, it reveals a fascinating and immensely fruit-
ful interplay between the emerging liaison of biology and psychology at the beginning of the
twentieth century and an influential epistemology widely believed to have been developed in
a purely philosophical context, notably logical positivism. In the second place, it shows
Popper’s epistemology, and hence his philosophy of science, to be much more indebted to
psychology than his vigorously professed antipsychologism would suggest.

D

OGMATIC

T

HINKING AND THE

P

SYCHOLOGY OF

P

EDAGOGY

In his autobiographical essay, “Science: Conjectures and Refutations” (1963), Popper re-

counted how he came to solve the problem of induction in the early 1920s. He approached this

376

MICHEL TER HARK

1. See also “The Bucket and the Searchlight” (Popper, 1972, pp. 341–362).

through Hume. Hume recognized that when we infer inductively the existence of an unobserved
effect from an observed cause (or vice versa) on the basis of past experience, our conclusions
have no rational support or justification. Popper agreed with Hume’s answer to what he calls the
logical problem of induction, but where he parted company with him was over whether establish-
ing a theory of the actual genesis of our knowledge of the future (and the past) means that “be-
lief ” must finally rest on induction. Having refuted the logical idea of induction, the question of
how we actually obtain knowledge indeed became an urgent one for Hume. Popper called it the
psychological problem of induction. As Popper saw it, there are two answers Hume could give:
by a noninductive procedure, thereby retaining a form of rationalism, or by an inductive proce-
dure, conceding that some of our most important modes of inference are made in the complete
absence of rational insight. Hume chose the second route and his psychological explanation of
our inferences from the observed to the unobserved in terms of habit and irresistible association
boiled down to a form of inductive learning. According to him, the experience of a constant con-
junction of ideas of causes and effects in individual experience ensures that the future occurrence
of one idea (of the “cause”) would automatically evoke the other (of the “effect”) in the mind.

This inductive psychological theory of the formation of cognition, Popper contended, is mis-

taken in at least three aspects: the typical result of repetition, the genesis of habits, and the char-
acter of those experiences or modes of behavior that he describes as “believing in a law.” The
typical result of repetition, Popper maintained, is that movements that first needed attention grad-
ually become automated and, hence, are performed without any attention at all. And habits, rather
than originating in repetition, Popper claimed, begin even before repetition can play any part
whatsoever. Finally, believing in a law, or expecting a law-like succession of events, normally can-
not be explained as resulting from repetition. On the contrary, a single striking observation may
be sufficient to create a belief or expectation. Thus, explaining the formation of cognition in terms
of a noninductive procedure, with anticipations and expectations accorded the role of hypotheses
enabling the organism or the scientist to know what to observe in the first place, Popper believed
he had rescued the rationality of cognition or, as he more often put it, critical thinking.

Popper claimed to have developed this alternative, deductive psychology of knowledge

in “ ‘Gewohnheit’ und ‘Gesetzerlebnis’ in der Erziehung” (1927), a protothesis that students
following the two-year teacher-training program at the Pedagogic Institute of Vienna had to
submit at the end of their course.

2

The manuscript however is not just a protothesis. As the

subtitle, “A structural-psychological monograph,” indicates, Popper’s ambition was to write a
full monograph, which, however, remained incomplete. Monographs were a relatively new
style of writing in child psychology inspired by William Stern, in which separate psycholog-
ical functions of the child, during a very limited period of his or her development, were sub-
jected to detailed observation and explanation.

3

Popper subjected the childish phenomenon of

dogmatic thinking to a detailed conceptual and empirical analysis resulting in a complex ex-
perience, the “experience of regularity” (Gesetzerlebnis, virtually untranslatable, incorporat-
ing the double suggestion of lived experience and normative requirements), and consisting of
three more elementary experiences: attitude (Einstellung), finding (Setzung), and standing by
one’s opinion (Festhalten). Children, according to the young Popper, expect regularities
everywhere and seek to find them even where there are none. They typically stick to their ex-
pectations even when inadequate and when they ought to accept their inadequacy. In many

THE PSYCHOLOGY OF THINKING

377

2. See my “Between Autobiography and Reality: Popper’s Inductive Years” (2002) for a detailed discussion of
Popper’s thesis.

3. The best example of such monographs is Die Persönlichkeit des dreijährigen Kindes (1926) by Stern’s pupil Elsa
Köhler.

cases, the resistance to (critical) demands of modification prompts or is even due to what he
calls “the fear of the unfamiliar” and the need for assurance or certainty.

With this analysis, Popper hoped to contribute to the pedagogical problem of delimiting

more precisely “where and when it is necessary to put limits to the free activity of pupils”
(Popper, 1927, p. 12). A proponent of the movement, headed by Eduard Burger, propagating
the self-activity of children, Popper nevertheless observed that a teacher may be forced to
apply the dogmatic way of teaching as long as a child is not yet capable of independent work.
Overlooking such limits can lead to grave problems, such as that of the child’s incapacity, due
to insufficient knowledge, to meet the demands of drawing certain consequences. In this way,
the teacher teaches everything “but he raises no character” (Popper, 1927, p. 13). Providing
the boundary between the “stage of habit” and the “stage of self-activity” with “an exact psy-
chological foundation” is a central problem of his monograph. This limited acceptance of
dogmatic education notwithstanding, Popper’s predominant attitude is negative. Following
Burger, he described the specific goal of education as free self-determination. Education must
be shaped by the goal of raising the child to the level of the individual requirements of such
autonomy, and failing to meet these requirements was to be less than fully a man of “charac-
ter.” The means to attain the level of free self-determination, Popper argued, were provided by
Burger’s principles of “activity and spontaneity,” which emphasize the emancipation from
prejudice in the creation of an autonomous character.

In this pedagogical context, Popper’s familiar epistemological distinction between dog-

matic and critical thinking makes its first appearance:

Free thinking, [. . .], is critical thinking—dogmatic thinking is unfree: By free thinking
one can only mean thinking “without prejudices”, that is, thinking which judges state of
affairs, without presupposing the result [. . .] of judging; critical thinking is also think-
ing guided by reasons; it is an active and spontaneous form of thinking, in contrast to
dogmatic thinking, which does not touch the accepted (adopted) “judgement.” (Popper,
1927, p. 10)

A superficial consideration of the concept of habit, he went on, reveals how problematic it is
to combine an education through habit with an education that conforms to the ideals of free
thinking; for inherent in the concept of habit, Popper claimed, is “a strong passive element, a
moment of automatism, the loss of insightful, active thought and its replacement by ‘auto-
matic’ association” (Popper, 1927, p. 11).

These educational implications of the psychology of dogmatic thinking, however, re-

ceived no further treatment in Popper’s thesis, thereby testifying to its unfinished character.
Another and related sign of Popper’s only limited interest in the empirical and practical con-
sequences of his analysis of dogmatic thinking was the following remark, also in the pref-
ace: “The psychological and also philosophical consequences of the psychology of the
Gesetzerlebnis proved to be so far-reaching that they surpass the pedagogical application,
although they too transcend largely the problem, in importance; therefore they will also be
indicated in the work at hand” (Popper, 1927, p. 4). Despite the fact that Popper did not say
explicitly what these consequences are, it is clear from the same preface that he must be re-
ferring to the concept of habit and its role in educational theories, for as he says a few lines
further: “It is known that the concept of ‘habituation’ plays a decisive role in educational
theories; by means of the investigation at hand it will have to undergo a revolutionary revi-
sion” (Popper, 1927, pp. 3–4). The newly constructed concept of Gesetzerlebnis is the con-
ceptual fruit of this revision since it is said to arise from “an analysis of what one tradition-
ally takes to be ‘habits.’ ”

378

MICHEL TER HARK

These ambitious claims seem to foreshadow the later rejection of Hume’s inductive the-

ory of the formation of cognition, yet there is no evidence that Popper really reached this
conclusion in his thesis of 1927, as he urged his readers to believe in the autobiographical
essay in Conjectures and Refutations. On the contrary, the most characteristic feature of
Popper’s later alternative theory of knowledge, the noninductive idea that theories or expec-
tations precede observations, was not only conspicuous by its absence but also contradicted
by his explicitly endorsing the inductive method himself. It was only the first element of his
later theory, the idea that repetition leads to automated processes, that was present in the the-
sis. Yet this part of the story was far from original and in fact boiled down to a theory of the
mechanization of acquired activities in the manner of the psychologist Karl Groos. Indeed,
a comparison of Popper’s theory of habit and Karl Groos’s theory of habit reveals similari-
ties too great to be accidental.

4

Karl Groos (1861–1946) studied philosophy in Heidelberg. His main interest was in the

psychological analysis of esthetic experience. In his study of the esthetically appreciating in-
dividual, play and the instinct of imitation stood on the foreground. Both aspects illustrate
Groos’s biological approach to psychology. As a professor in Basel and later Tübingen, he lec-
tured on pedagogy. In 1901, he published one of the earliest experimental studies on thinking
(Groos, 1901–1902), the core of which would return in his pedagogical studies. In the later
editions of his Das Seelenleben des Kindes (1923), he elaborated upon the similarity between
his theory of thinking and the work of Otto Selz.

Like C. Lloyd Morgan and Herbert Jennings, Groos was concerned with the distinction

between instinct and habit. But unlike them, he approached this problem also from an educa-
tional point of view. In this emphasis, Groos aligned himself with William James, who had
maintained that the fundamental mechanism for the acquisition of new activities is habit or
exercise (Gewohnheit, Übung). A specific modification of this general mechanism is what
Groos called the urge for repetition (Wiederholungsdrang). Tendencies to repeat certain activ-
ities, more prominent among children than among adults, lead to the acquisition of habits. A
further modification of James’s general principle consisted of a “habitual attitude”
(Einstellung); people can make themselves ready for a certain soon-to-be-learned activity.
With practice and “frequent repetition,” the habitual attitude follows increasingly “involuntar-
ily, indeed even under circumstances unconsciously” (Groos, 1923, p. 52). This essentially as-
sociative process ultimately leads to the complete mechanization of the newly acquired activ-
ity: “By this is meant that with increasing exercise consciousness, which played a leading role
during the first acquisition of the activity, fades into the background” (Groos, 1923, p. 52).
Thus having become fully mechanized, acquired activities complete innate drives and in-
stincts by what is aptly called a system of “acquired reflexes.”

Another indication of Popper’s indebtedness to Groos is to be found in the paragraph on

the phenomenology of the Gesetzerlebnis. The first and most important element of the com-
plex experience of dogmatic thinking discussed there in detail by Popper is the “attitude to-
wards the unfamiliar.” Despite the fact that Popper did not mention Groos by name here, his
use of the expression “fear of the unfamiliar,” echoes Groos’s “fear of what is unhabitual.” The
(habitual) readiness for certain things to happen in specific circumstances, Groos contended,
is typically not a conscious experience. Rather, only when something unexpected occurs are
we surprised, and this very surprise betrays the idea that we were prepared for what is habit-

THE PSYCHOLOGY OF THINKING

379

4. Popper did mention Groos explicitly both in the text and in the bibliography. In the text, he referred to Groos’s
theory of exercise (Übung) in the context of a discussion of the role of playing in the development of the child’s in-
tellectual capacities.

ual. On the other hand, if something that is “unhabitual” and unexpected pops up, this will
draw one’s attention involuntarily. When this surprise is accompanied by a reflexive contrac-
tion of the body, Groos spoke of a shock (Chok). The shock, accompanied by strong aversive
feelings, he called “the fear of the Unheimlichen.”

The following remark by Popper shadows this theory of habit and what is unhabitual:

“The fear of the unfamiliar, the feeling of displeasure, [more easily] crosses the threshold of
consciousness than the feeling of pleasure, the pleasure about what is known, since repetitions
numb the irritability” (Popper, 1927, p. 54). Like Groos, Popper maintained that what is fa-
miliar, what is habitual, typically does not catch one’s attention. Against this background, his
choice of the term attitude was also significant and reminiscent of Groos. And like Groos,
Popper maintained that what is unknown or unfamiliar crosses more easily to the level of con-
sciousness. Finally, like Groos, Popper maintained that this consciousness typically takes the
form of fear or anxiety.

The evidence then for Popper’s reliance on Groos is undeniable. Although the latter’s

idea that repetitions lead to acquired activities becoming automated, rather than to something
new, differed from the traditional Humean psychological theory, by precisely according a
dominant role to repetition and association, it was far from revolutionary in the sense in which
Popper took his later theory to be. Since Popper’s rejection of and alternative to the second
claim—that habits do not originate in repetition, as Hume would have it, but begin in the
sense of expectations, before repetition can play any part whatever—is the core of his alter-
native noninductive psychological theory, it follows that he had not taken his anti-Humean
stance in 1927. Further support for this conclusion comes from his earlier described largely
negative attitude toward dogmatic thinking. An important (later) insight contributing to a de-
ductive theory of knowledge was Popper’s point to the effect that there can be no critical
thinking without a preceding phase of dogmatic thinking, a phase in which an expectation is
formed so that error elimination can begin to work on it. This view implies a much more pos-
itive appreciation of the role of dogmatic thinking than was to be found in Popper’s thesis, for
rather than claiming that dogmatic thinking is a necessary stage before critical thinking can
emerge, the young Popper, wholeheartedly in the spirit of the school reform movement, was
worried about the social effects of an education through habit: “It will be clear from the start
that ‘habit’ etc. as a means of education may have only a narrowly confined scope if it will
not run counter to the tasks of the educating generation” (Popper, 1927, p. 11).

P

SYCHOLOGY OF

T

HINKING AND THE

B

IOLOGICAL

T

URN

In his second writing on psychology, his PhD thesis “Zur Methodenfrage der

Denkpsychologie” (1928), the second and decisive step in the transformation of the Humean
paradigm in psychology and epistemology was not taken either, although the incipient influ-
ence of anti-associationist psychologists like Hans Volkelt and especially Otto Selz was
clearly discernible. Another pivotally important formative feature of Popper’s dissertation was
the attempt to establish firm connections between biology and psychology by such diverse au-
thors as Jennings, Volkelt, Bühler, and Selz. It was also Bühler who set the agenda of Popper’s
dissertation. Popper’s main goal was to defend Bühler’s pluralistic methodology as put for-
ward in his recently published Die Krise der Psychologie (1927). In this most important book,
based upon his famous course in general psychology at the University of Vienna, Bühler dis-
cussed the methodological problems implicit in the many different approaches to psychology
at the time. The “crisis” of which Bühler spoke therefore was not indicative of a painful lack
of ideas and theories, but rather of an embarras de richesse. The drawback of this multiplic-

380

MICHEL TER HARK

ity of theories, as Bühler saw it, was that some of the schools, notably behaviorism, the hu-
manities, and psychoanalysis, were inclined to overplay their hands and to claim a monopoly
of truth. Bühler’s “crisis” has to be understood in this sense. In its constructive phase, the
book sought to establish the partial truth of each of the different schools and to show how their
different points of view were complementary. No school was rejected in toto, and Bühler
sorted out and selectively accepted distinctive contributions of the various schools. His main
point of criticism was that no school is capable of doing justice to what are the three basic
characteristics of human conduct: lived experience, meaningful behavior, and their relation to
culture (Gebilden des objektiven Geistes). A unitary science of psychology, Bühler main-
tained, is the science of the triad experience-behavior-culture. One of Popper’s concerns was
to demonstrate the indispensability of Bühler’s pluralistic methodology for Denkpsychologie,
and in this context, a separate problem is the evolutionary theory of cognitive development.
For in the course of his discussion of Denkpsychologie, it became clear that he was no less in-
terested in the evolutionary theory of cognitive development.

Popper’s preoccupation with an evolutionary theory of cognitive development was also

reflected in the sequence in which he dealt with Bühler’s three aspects. Rather than starting
to discuss the aspect most relevant to Denkpsychologie—the aspect of experience—it was in
fact one furthest removed from the study of thought but firmly rooted in the evolutionary sci-
ences that receives credit not only for being the first in line but also for receiving the most
sustained treatment: the behavioral aspect. Although the discussion was rather exploratory
and at some points even unbalanced, Popper’s later emphasis on the method of trial and error
and the related campaign against empiricist epistemology found its roots here. The fundamen-
tal question of his discussion of the behavioral aspect, even though not explicitly posed,
seemed to be whether thinking is to be explained in terms of sense experiences (the receptive
system, as Popper called it) or in terms of behavioral processes (the reactive system). In the
discussion that follows, we will see Popper still wavering between, on the one hand, a
Machian account with its overriding priority of the receptive system, and, on the other, the
proposal of Volkelt and Selz, aptly summarized by Popper as “Nihil est in intellectu quod non
antea fuerit in sphaera motorica” (Nothing is in the intellect which has not been before in the
motoric sphere) (Popper, 1928, p. 54).

Popper’s main interest was in what could be called subjectively functional (zweckhaftest)

behavior. This type of behavior plays an important role in what he aptly called “behaviouristic
Denkpsychologie” (Popper, 1928, p. 49). Objectively efficient behavior is behavior that is highly
adapted to a specific situation; yet, were an animal to show maladjusted behaviour (for instance,
a dog barking at a locomotive), it does not immediately follow that it is completely pointless. If
it can be shown that the animal’s behavior would have been functional in another situation, it is,
although not objectively efficient, subjectively functional (zweckhaft). According to Popper,
Mach would have explained maladapted yet subjectively functional behavior in the following
way: the situation in which the animal behaves (objectively) ineffectively means the same to the
animal, as that in which the same reaction would have been functional (cf. Mach, 1896, p. 416;
Popper, 1928, p. 51). The animal has made a “mistake”; it has “confused” the one for the other.
The main defect of this explanation is that it does not explain further what is meant psycholog-
ically by the phrase “means the same to the animal.”

This explanation, Popper believed, was provided for by Hans Volkelt who, in his much-

neglected Über die Vorstellungen der Tiere (1914), sought to understand how the animal sees
and conceives its environment, in a way which differed from both behaviorism and unre-
strained anthropomorphism. Volkelt’s main question was how it could be explained that the an-
imal, placed in the same situation, and receiving the same visual information, behaves adap-

THE PSYCHOLOGY OF THINKING

381

tively to the situation the one time and the other time not. Volkelt proposed to explain this dif-
ference in behavior in terms of “complex-qualities.”

5

One of the most striking features of ani-

mal behavior, he observed, is its being adapted to a very limited scope of situations, and that it
is more or less maladjusted to circumstances outside of this circle. Since for evolutionary rea-
sons of economy it seems to be the total situation (Gesamtsituation), rather than its elements,
that determines the animal’s behavior, animal consciousness somehow effects a synthesis of the
sensory material. Put otherwise, animal consciousness consists of complex-qualities. The coa-
lescence of a multiplicity of situations into complex-qualities is not the only cause of the finely
tuned conduct of animals. It is only on the further assumption that complex-qualities are coor-
dinated with specific ways of behaving, thereby becoming effective, that a full-fledged expla-
nation of the adaptation of animal behavior is accomplished (Volkelt, 1914, p. 106).

The answer to Popper’s question of what it can mean psychologically that different situ-

ations have the same meaning for the animal is that its visual impressions of both situations
are embedded in the same complex-qualities (Popper, 1928, p. 52). It was only Volkelt, Popper
went on, who provided an evolutionary explanation of such exchanges. The extent to which
complex-quality and reaction are effectively coordinated, Volkelt contended, is proportional
to the vital importance of the current situation (Volkelt, 1914, p. 98). The animal organism is
highly adapted to vitally important situations, and it is only in “vitally indifferent situations,”
situations to which it is not adjusted, that the animal can make “mistakes.” Thus, Popper’s con-
clusion: “To the objective observer this reaction is a failure, but not to the animal: the animal
acts subjectively functionally” (Popper, 1928, p. 54).

With its emphasis on evolutionarily prefabricated and coordinated reactions, Volkelt’s

theory of animal cognition clearly prepared the ground for Popper’s later rejection of the, ac-
cording to him, deeply entrenched view of the mind being a tabula rasa and sense perception
the origin of all (human) knowledge, the so-called “bucket theory of mind.” But the most dis-
tinctive feature of his alternative theory of (animal) cognition, the so-called “searchlight the-
ory of mind,” as well as extra munitions for discarding empiricism and associationism, came
from Otto Selz. A first clue to this pivotally formative role of Selz comes from the sequel to
the discussion of Volkelt when Popper referred to the former’s theory of “trying-out behav-
iour” (probierenden Verhalten). Selz’s psychology of productive thinking, in particular, his in-
terpretation of Wolfgang Köhler’s findings of the intelligent achievements of chimpanzees in
terms of his own theory of trying-out behavior, Popper observed, “is clearly strongly biolog-
ically oriented” (Popper, 1928, pp. 57–58). Discussing the aspect of experience, Popper again
mentioned Selz approvingly and even quoted in full a passage from the latter’s synopsis of his
psychology of reproductive and productive thinking in which his own evolutionary approach
was clearly contrasted with pre-Darwinian forms of association psychology:

At least for a limited area we have demonstrated that new modes of response of organ-
isms arise systematically from previously developed effective (i.e., life-supporting and
life-enhancing) modes of response, and can thus be made comprehensible. It is not
through a senseless play of associations but invariably by dint of previously developed ef-
fective (life-supporting and life-enhancing) operations that new mental modes of func-
tioning are seen to arise. . . . Perhaps our era is witnessing the beginnings of a biology of
the inner man. Psychology thus enters the ranks of the biological sciences. (Popper, 1928,
p. 69; Selz, 1922, p. xii)

382

MICHEL TER HARK

5. This concept Volkelt takes over from his teacher Felix Krüger, who formulated it in opposition to Gestalt psychol-
ogy. Complex qualities, although similar to Gestalt properties in being more than the sum of their constituting parts,
have a phenomenological immediacy absent from the standard examples of (physical) Gestalten as described by the
Gestalt psychologists. See Krüger (1926).

Finally, Popper referred to Selz’s theory of trying-out behavior in a passage that foreshadows
his later philosophy of science. But before discussing this latter development, it is necessary
to give a brief outline of Selz’s theory.

O

TTO

S

ELZ AND

P

OPPER

’

S

P

SYCHOLOGY OF

S

CIENTIFIC

D

ISCOVERY

Otto Selz (1881–1944) lived much of his life in seclusion, cherishing the tranquility he

needed to develop his epistemological, psychological, and pedagogical ideas. Only a passport
photograph has remained of him. In his scientific work, Selz was increasingly marginalized,
owing to his unremitting criticism of colleagues but also to his formidable, complex style of
writing. Closely allied to the Würzburg School, he did not shrink back from launching frontal
attacks on the ideas of some of its members. Aside from one pupil, Jules Bahle, who closely
collaborated with Selz, and the Dutch scholars Adriaan de Groot and Frans Prins, who applied
his ideas in psychology and pedagogy, respectively, he never founded a school, and after 1933,
his name disappears almost completely from the German psychological literature.

6

In 1909, he

took his PhD in philosophy at the University of München. After his PhD, Selz went to Bonn to
do experimental investigations in the laboratory of Külpe. Both Külpe and Bühler were among
his subjects, and he probably attended some of their seminars.

7

These investigations resulted in

his first major work, his Habilitationsschrift, Über die Gesetze des geordneten Denkverlaufs.
Eine experimentelle Untersuchung (1913). With his second major work in the psychology of
thinking, Zur Psychologie des produktiven Denkens und des Irrtums (1922) (its publication
was postponed because of the First World War), Selz’s intellectual prestige was incontestably
on the rise, and in 1923, he was called to the chair of Philosophy, Psychology and Pedagogy at
the Handelshochschule in Mannheim. The practical orientation of the Handelshochschule
more or less forced Selz to work also on the practical applications of his psychology. In spite
of the abstract nature of his psychology, he succeeded in pointing out its relevance for peda-
gogy and, in particular, for attempts at fostering intellectual achievements, a project undertaken
with some of his pupils.

8

Meanwhile, the psychological institute of Mannheim headed by Selz,

who became Rector of the Handelshochschule in 1929–1930, flourished, and the first disser-
tations on Denkpsychologie and pedagogy began to appear, among them Jules Bahle’s cogni-
tive psychological investigations of musical composing (Bahle, 1930).

All this came to a sobering halt after 30 January 1933, when Hitler was appointed

Chancellor of Germany. Most expelled psychologists left Germany and migrated to the United
States. Not Selz. He led a withdrawn life in Mannheim, where with opportunities to do exper-
imental work gravely diminished, he threw himself into purely theoretical work on the Aufbau
of the phenomenal world. After the Reichskristallnacht he was caught and deported to the con-
centration camp of Dachau, from which he was relieved again in December 1938. In May
1939, he finally migrated to the Netherlands. The one desirable outcome of this shameful
episode was that Selz came into contact with the Dutch pedagogue Phillip Kohnstamm and
A. D. de Groot. Selz taught at the Amsterdam Teachers Seminar (Nutseminarium) on psychol-
ogy and pedagogy and participated in scientific discussions at the Faculty of Psychology,
hugely enriching the field of psychology. He would not be spared the horrors of the Holocaust,

THE PSYCHOLOGY OF THINKING

383

6. See Seebohm, 1970, p. 278. Although Selz’s name can be found in Boring and Watson’s list of 538 important psy-
chologists since 1600 (Watson, 1977), neither Boring’s own History of Experimental Psychology (1929) nor
Watson’s The Great Psychologists (1978) even spend a footnote on Selz.

7. Alexandre Métraux, e-mail correspondence.

8. E.g., Selz (1935) and Kindler (1929). The latter was the first PhD thesis after the Handelshochschule turned into
a university.

though. In July 1943, he was caught again by the Nazis and deported to the concentration camp
Westerbork. A postcard saying that he wanted to give courses in Westerbork was the last sign
of life. On 24 August, he was put on train Nr. DA 703 to Auschwitz. He either died in transit
from suffocation or exhaustion—he was suffering from heart problems—or was murdered by
being sent into the gas chambers.

Although his experimental set-up was in the tradition of the Würzburg School of

Psychology, headed by Oswald Külpe, Selz’s theoretical work deviated from the school in two
respects. The first difference was that Selz did not adopt even a modified version of associa-
tionism; he rejected it completely. His relentless criticism of various proponents of associa-
tionism was elaborated in detail in his first major work (Selz, 1913). The second difference
was that in Selz’s work the explanandum of psychology was shifted from the content of think-
ing to the process of thinking. Thought processes, Selz maintained, essentially consist of a se-
ries of operations designed to complete certain cognitive or motor complexes (in the case of
skills). This conceptual revision of the psychology of thinking was found in his second major
work (Selz, 1922) and in his synopsis of his two major works, Die Gesetze der produktiven
und reproduktiven Geistestätigkeit. Kurzgefasste Darstellung (1924/1980).

Rather than reducing problem solving to the scheme of stimulus-response, whether con-

ceived in a physiological or behavioral manner, the theory of Komplexergänzung instead
explained thinking in terms of a process of completing gaps in already existing conceptual or
relational complexes. Problems can take two forms: either the question is one of reinstalling a
certain relational complex, in which case no new constructions are required and the task is sim-
ply one of reverting to the existing memorial complex, or there is a genuine new problem, in
which case actualizing knowledge from memory won’t suffice. In these latter cases of produc-
tive thinking, “solving-methods,” a term newly coined by Selz, are becoming actualized.
Solving-methods’ effectiveness can mean that familiar methods are applied to new cases or that
familiar methods are the basis for developing new methods. In reproductive and productive
thinking alike, the process of Komplexergänzung is directed by “schematic anticipations” of
the solution. To take one of Selz’s own (simple) examples, a candidate at the exam not only
knows that the first letter prompted by the examinator is part of a whole, but also that this
whole is a word and that the word begins with this letter. This knowledge does not consist of
separate elements, but rather the subject knows that the prompted letter is part of a word; that
is, a relational whole, or structure, is involved rather than an aggregate of elements. Being
aware of this cognitive whole prompts the subject to anticipate schematically the answer to the
question. Thus establishing a system of provisional relations between the new elements and the
cognitive whole they fit into, schematic anticipations clearly function as a sort of hypotheses.

In emphasizing operations and solving-methods, Selz in fact sought to transform psy-

chology as the science of mental contents, whether conceived as “imageless thought” (the
Würzburg School) or as intentional states (the Brentano School) into a biological disci-
pline. Indeed, in his work from the 1920s onward, he explicitly compared operations and
solving-methods with “functional body movements,” such as reflexes and skills. A person,
he argued, is “an organized system of functional cognitive modes of behaviour” (Selz,
1991). Unsurprisingly, given his pioneering role in especially German psychology, Selz’s
attempt to integrate psychology and biology remained tentative and unelaborated. Yet, in his
study of habit formation and the theory of learning, he introduced a notion that was to play
a crucial role both in philosophy (Popper) and psychology (de Groot). Routine solving-
methods, Selz explained, are gradually acquired by what he calls “trying-out behaviour”
(probierendes Verhalten) (Selz, 1924/1980, pp. 48–49). Significantly, Selz discussed this
form of problem solving for the first time in his detailed interpretation of Köhler’s famous

384

MICHEL TER HARK

study of chimpanzees at Tenerife in 1917, a study, by the way, that made use of Selz’s ear-
lier ideas on human thinking (Selz, 1922, p. 645ff.).

9

Like the method of trial and error fa-

miliar from Thorndike, Morgan, and Jennings, trying-out behavior is a form of learning, but
it is not a repetitive and associative process; rather, it is a process equally initiated and
guided by schematic anticipations. The most important difference, therefore, between blind
trial and error and trying-out behavior is that in the latter case, attempts are based on a par-
tial insight into the situation. Always showing a clear sense of direction, the organism tries
out within a pre-set, goal-determined, and limited domain of solution possibilities; in Selzian
terms, schematic anticipations codetermine the where and what of search and trying.

An immediate implication of the role of schematic anticipations during habit formation

was a revision of the then dominant theory of cognitive development, which assigned an irre-
ducible place to the intellect. According to this model, espoused above all by Bühler, intelli-
gence gradually arises from the increasing multiplicity of acquired habits or associations.
Following this tradition of developmental psychology, Selz also distinguished between instinc-
tive learning, automatic learning (habit formation), and insightful learning, yet, according to
him, even the simplest form of habit could be explained in terms of associative learning. On
the contrary, “acquired reflexes” are also guided by schematic anticipations. Thus conceived,
the study of the genesis of habits becomes relevant for the study of the development of intelli-
gence for, rather than being a qualitatively different and older stage, cognitive operations are
“a developmental integration of intellectual actions into an existing, more primitive system of
specific responses” (Selz, 1927/1964, p. 229).

10

With less than 250 copies sold two years after its publication, Selz’s Zur Psychologie des

produktiven Denkens und des Irrtums was quickly forgotten. The book was translated only
into Chinese, not into the English language, which was to become the official language for
psychology, epistemology, and the philosophy of science. Yet those who had taken the trouble
to plow their way through Selz’s bulky work were richly rewarded. Indeed, members of the
Gestalt movement, orchestrating their campaign against association psychology with consum-
mate skill, profited greatly from his holistic theory of schematic anticipations. Much later, in
the 1940s, the Dutch psychologist A. D. de Groot would apply the main ideas of Selz’s psy-
chology of reproductive and productive thinking to the analysis of problem solving during
chess, published in his Het Denken van den Schaker (1946), a book which in turn would stim-
ulate the pioneering work in artificial intelligence of Allen Newell and Herbart Simon (1958).
Reading Selz somewhere between 1926 and 1928 and using his ideas in transforming first
pedagogy and psychology and later epistemology, the young Popper therefore showed preco-
cious yet consummate discernment in interpreting new scientific developments, preceding de
Groot by almost two decades. Indeed, Popper would be the only one to appreciate the biolog-
ical or evolutionary dimension of Selz’s work and use this part especially for his own project
of an evolutionary epistemology and philosophy of science.

The beginnings of this project were also found in Popper’s dissertation. Discussing the

need of what Bühler called the aspect of culture for Denkpsychologie, Popper suddenly turned
to the method of scientific research (Forschungsseite). The importance of Denkpsychologie
for the study of objective products of the mind, notably science, lies, he believed, in its pro-
viding a model for understanding the growth of science:

THE PSYCHOLOGY OF THINKING

385

9. Köhler acknowledged his debts to Selz (1913) in The Mentality of Apes (1917/1925, pp. 150, 174). As Selz re-
called, Köhler’s results confirmed his own idea that directed thinking is ultimately based on the application of partly
insightful and partly automatic solving methods acquired in phylogenesis (Selz, 1922, p. 610, note 1).

10. Quoted from the English translation in Mandler and Mandler (1964).

Perhaps there are important parallels in the methods and operations of the scientific and
the “pre-scientific” induction?

To give just one example: The Selzian concept of trying-out behaviour seems to me to
have important parallels in objective scientific research. Science tries out its methods, its
“models” (as Bühler puts it), and in such a way as to correspond completely with the
Selzian scheme. As is well known the actual ways of scientific research in no way cor-
respond with the logical principles of the representation; as little as the operations de-
scribed by Selz correspond with the objective logical operations. Despite this science is
in the end clearly driven by tasks, the determining tendencies come clearly to the fore.
(Popper, 1928, pp. 69–70)

The earliest sign of his concern with the nature of scientific research, this passage, with

Selz calling the tune, unmistakably shows Popper’s ideas on the logic of scientific discovery
emerging in the context of the psychology of scientific discovery. Although he rejected the
relevance of the psychology of discovery for the philosophy of science in The Logic of
Scientific Discovery (1934/1959), in the post-war years, he unwaveringly adhered to his pro-
posal to compare individual and scientific cognition and use the Selzian method of trial and
error as a measure. Indeed, he gave the Selzian method of trial and error the highest general
sense possible, incorporating not only individual psychology but all the sciences, including
the Geisteswissenschaften and evolution. The rejection of any Geisteswissenschaft, and the at-
tendant proposal of the methodological unity of the sciences, which was also part of Selz’s
ambitious plans of the “biology of the inner,” however, had to wait another 16 years until
Popper wrote The Poverty of Historicism and The Open Society and Its Enemies.

At this juncture, Popper seemed not that far along. Another indication of the still rudi-

mentary grasp of the depth and implications of Selz’s work at this juncture was his repeated
use of the term induction, from which we may definitively conclude that, in contrast to what
the autobiographical sketches in Unended Quest and Conjectures and Refutations urge us to
believe, Popper, as late as 1928, still endorsed an inductive methodology. Selz’s detailed and
frontal assault on association psychology, and his defense of a theory of schematic anticipa-
tions, in fact boiled down to a view of the animal or human organism as an active cognitive
subject constantly putting forward tentative proposals or hypotheses rather than as a passive
recipient, patiently waiting for the accumulation of information to be inductively safe.

How then is it to be explained that the inductive paradigm had such a strong hold on

Popper? Indeed, if, as he later confesses, deductive lines of thinking are primarily to be found
among biologically oriented psychologists, why did he stick to inductivism while discussing
Selz? An explanation is perhaps forthcoming if one considers other influences on the young
Popper, notably the biologists Morgan and Jennings. In his later work on evolutionary theory,
Popper pointed to the influence of the philosophy of “emergent evolution,” defended by
Morgan (Popper, 1999, p. 12). Not mentioned by him, but arguably even more important, was
the view of Jennings. In an article in Science (1927), a copy of which is in Popper’s papers,
Jennings eloquently and passionately argued for the doctrine of emergent evolution both as a
program in biology and as a philosophy of science and life. He opposed the then prevailing
mechanistic view of evolution as “the working of a great machine that never alters its mode of
action nor the nature of its product” (Jennings, 1927, p. 20). The method of science based on
this mechanistic and deterministic view of evolution, Jennings explained, is mainly rationalis-
tic and to but a minimal extent empirical. Against this view, the doctrine of emergent evolution
holds that new things, not thus computable, emerge as evolution progresses: “It holds that with
these emerge new methods of action, following new laws; methods not before exemplified;
methods that falsify the results of computations based on former methods of action” (Jennings,

386

MICHEL TER HARK

1927, p. 21). It is this emergence of new things that distinguishes the living from the nonliv-
ing, Jennings maintained. From this standpoint of emergent evolution, Jennings pleaded for a
radical experimentalism in science. Since there is always the possibility that new things or
properties have emerged, he averred, they cannot be discovered by ratiocination but only by ob-
servations and experiment. Wholeheartedly embracing John Hunter’s maxim, “Don’t think;
try!” Jennings concluded that thinking is a (fallible) instrument only for helping to decide what
to try, “but the last word must be try” (Jennings, 1927, p. 21). Thus, radical experimentalism
as a philosophy of science and emergent evolutionism as a philosophy of biology are mutually
supportive. The (inductive) method of trial and error is not only the method of science but also
of evolution itself. Indeed, the only possible method for progress in emergent evolution is by
trial and error. Jennings went on: “In such progress by trial and error will indeed be found free
play for the utmost sharpness of vision as to what it is best to try . . . but in the end a trial it
must be, with no antecedent certainty as to results” (Jennings, 1927, p. 22).

P

SYCHOLOGY AND

E

PISTEMOLOGY

“Zur Methodenfrage der Denkpsychologie” was Popper’s second piece of writing in psy-

chology, yet it was seen to contain no thematic overlap whatsoever with the manuscript on dog-
matic thinking one year earlier. Indeed, the analysis of the Gesetzerlebnis, carefully laid out in
1927, is conspicuous by its absence in the dissertation, which, as I have attempted to show, con-
tained the roots of Popper’s method of trial and error. Yet the method of trial and error is a the-
ory of the interplay between dogmatic thinking and critical error elimination. So the question
is how and when Popper came to transform his earlier theory of dogmatic thinking into his evo-
lutionary epistemology, conceived both as a theory of individual and scientific development.
To appreciate the enormous difficulty of integrating the early view on dogmatic thinking and
the later theory of trial and error, it is important to note that their respective basic assumptions
were fraught with considerable tension. The early theory of dogmatic thinking was primarily a
theory of a stage of human character, which, because of its prejudiced nature, its inclination to
hold on passively to what is familiar, and its fear of the unknown, has to be overcome and re-
placed by the active engagement through self-activity typical of critical thinking. In the later
theory of trial and error, by contrast, dogmatic thinking, in every phase of the growth of (sci-
entific) knowledge, is indispensable to objective knowledge. It is not just critical thinking that
is active, but dogmatic thinking also is now conceived as an active process of putting forward
trials or hypotheses, thereby changing from an inductive and associative process to a form of
noninductive or deductive learning. How did this reversal come about?

At this point, Selz’s theory of schematic anticipations seems to come in, for in his sketch

of a deductive psychology of knowledge in Die beiden Grundprobleme, Popper, seeking to an-
swer the question of how (intellectual) reactions, drawn from the mind rather than from the
world, nonetheless prove themselves adaptive in objective circumstances, invoked anticipatory
mechanisms. Yet Selz is mentioned only in passing. From a little article, “Die Gedächtnispflege
unter dem Gesichtspunkt der Selbsttätigkeit” (Popper, 1931), published in the monthly journal
of pedagogical reform, Quelle, however, it appears that Popper had finally come to appropri-
ate completely Selz’s theory of anticipations in an attempt, just like his thesis on
Gesetzerlebnis, to solve a pedagogical debate between, on the one hand, the Lernschule, and,
on the other, the Arbeitsschule concerning the role of memorization in education.

The labor schools attempted to steer education away from a drill school approach, typi-

cal of the Lernschule, toward seeking children’s active engagement through self-discovery.
Having a huge amount of knowledge at one’s disposal was the ruling principle of the

THE PSYCHOLOGY OF THINKING

387

Lernschule. This principle demanded a lot of memorization. Mnemonic exercise was
achieved, according to the school, by accumulation of knowledge and frequent repetition of
this material. The ensuing description of the psychology underlying the pedagogical program
of the Lernschule shows Popper using for the first time a metaphor that would figure promi-
nently in his later writings on epistemology: “To the Lernschule memory is nothing but a con-
tainer of material, a sort of bucket of knowledge” (Popper, 1931, p. 610). The essence of
memory, on this view, is to let in and store knowledge. The consequences for pedagogy,
Popper concluded, were that mnemonic exercise could only be achieved by repeating the
process of storing and retrieving information as much as possible, and by an accumulation of
dictated knowledge, which will enlarge memory space.

In his later work, Popper replaced the bucket theory with his theory of the searchlight, but

in 1931, this revolutionary role was accorded to Selz’s theory of anticipations. As Popper now re-
minded his readers, the decisive turn away from association psychology “was initiated by Kant
and carried through, according to strict experimental methods, by the school of Külpe, especially
Bühler and Selz” (Popper, 1931, p. 610). The fundamental mistake of association psychology,
Popper argued, was its attempt to derive the whole of human memory, even the whole of intellec-
tual capacities, from a single and simple form of associative memory (what he calls the bucket).
Popper’s alternative account of the genesis of the different functions of memory followed Selz’s
Denkpsychologie in detail. The role of associative memory, he observed, is restricted to the pro-
cessing of nonsense syllables in the laboratory, but even in such rather artificial situations, sub-
jects often establish meaningful connections between stimuli. With this understanding of mean-
ing, Popper concluded, thinking enters memory and “the laws of the mechanisms of association
are replaced by the ‘laws of ordered thinking’ (Selz).” And a few lines further: “Selz has coined
the name ‘intellectual operations’ for the functions of thinking” (Popper, 1931, p. 613). That
Popper’s alternative account of memory and memorization wholly depended for its conception
on ideas he took over from Selz is corroborated by a further passage in which the latter’s theory
of schematic anticipation is put forward as providing the Arbeitsschule with the required notion
of psychological activity underlying even rote memory:

Selz has shown that “reproductive thinking” is an extremely active process, a production
process (Arbeitsvorgang). The important method, the important tool of this production
process, is the scheme of thought (Denkschema). In this scheme an unoccupied space
(Leerstelle) takes the place of lacking thoughts (or pieces of thought), thoughts that have
to be reproduced. The systematic completion of these unoccupied spaces of the scheme
(the “determined complex-completion”) leads to reproduction. (Popper, 1931, p. 616)

Rather than being a passive and mechanical process, Selz taught, human memory turns

out to be a systematic reconstructing of schematic anticipations and their gaps.
Denkpsychologie clearly demonstrated, Popper explained, that forgotten information has not
simply been overlooked but has left “an unoccupied space in memory, analogous to the un-
known x in a mathematical equation, which prompts the urge to fill it in (complex comple-
tion)” (Popper, 1927, p. 616). Retrieving information from memory thus becomes a process
of “methodically reconstructing schemes of thought” (Popper, 1927, p. 616). In this way, in-
culcating information can be changed from a dull mechanical process into a conscious and
methodical procedure evoking the child’s interest and giving it pleasure. “Inculcating, then,
becomes a process of thinking (Denkarbeit)” (Popper, 1927, p. 616). This psychology of
memory, Popper believed, can help steer education away from the Lernschule, in which chil-
dren are treated as empty buckets to be filled by the accumulation of knowledge, toward seek-
ing children’s active engagement through thinking, without neglecting the role of memoriza-

388

MICHEL TER HARK

tion in favor of the exercise of purely intellectual functions. It is also this Selzian theory which
seems to have been crucial in bringing together the early theory of dogmatic thinking and the
method of trial and error, conceived both as a theory of learning and an epistemological the-
ory (of “conjectures and refutations”).

Indeed, still subscribing to the (associational) theory of Groos, Popper, in his monograph

on the Gesetzerlebnis, had not yet reached that part of his alternative theory of habit that is in
fact the core of his deductive epistemology (namely, anticipations preceding repetitions), but
now having changed Groos for Selz, there were passages in the pedagogical article that were a
sign of a burgeoning deductive psychology of knowledge and mind. Recalling that the bottom-
up attempt of compiling the higher forms of thinking from the mechanisms of association, so
characteristic of association psychology and the related constellation theory, had completely
failed, Popper was anxious to point out that memorization guided by “the laws of ordered
thinking,” although equally mechanical, was yet completely different from associative mem-
ory; he dubbed it “automatized insightful memory.” Automatized insightful memory was de-
fined by opposition with the (failed) bottom-up approach of association psychology, its gene-
sis proceeding the other way around. The key distinction was that in a theory of ordered
thinking, the process of mechanization sets in later than in an associative theory. Only after the
pupil has familiarized himself with the relevant piece of memorial knowledge by means of the
intellectual operations as described by Selz can mechanization be initiated. The result in both
cases seems to be the same, yet the difference is as big as that between a skilled “piano player
and a gramophone record” (Popper, 1931, p. 613). As this analogy indicates, insightful mem-
ory becoming automatic is a process on a par with the development of skills, of know-how. It
was here that the relevance of the thesis of 1927 came in. Seeking to revolutionize the tradi-
tional concept of habit, Popper, in 1927, appealed to Groos’s law of mechanization, which was,
in fact, an explanation of the genesis of know-how. In Groos, however, mechanization was seen
to be a purely associative process. Unsurprisingly, therefore, Popper no longer relied on Groos
in his article of 1931, instead drawing on Selz.

11

An interesting consequence of this Selzian ap-

proach to memory was that “where these reactions, these processes of reconstruction, are not
shaped yet, there is nothing which can be abbreviated” (Popper, 1931, p. 618). In the context
of the article, this was a pedagogical warning not to let the process of mechanization begin too
soon, but at the same time it can be seen as conveying the noninductive epistemological mes-
sage that, since learning (mechanization of skills) can take place only on the basis of already
shaped intellectual operations, these intellectual operations have to precede empirical knowl-
edge acquisition.

This contention would be elaborated further by Popper in his sketch of a deductive psy-

chology of knowledge in his first philosophical work, Die beiden Grundprobleme der
Erkenntnistheorie (1933/1979). In fact, this sketch continued the discussion of the behavioral
aspect in the PhD thesis, but now, having incorporated Selz’s theory of specific reactions,
overriding priority was given to the reactive system. This, however, raised the epistemologi-
cal question of how (intellectual) reactions, drawn from the mind rather than from the world,
nonetheless prove themselves adaptive in objective circumstances (Popper, 1933/1979, p. 24).
In answering this question, he dropped the name of Selz. Subjectively preformed reactions,

THE PSYCHOLOGY OF THINKING

389

11. Selz explained the genesis of know-how as a form of unmediated problem solving (Selz, 1913, chapter 1). He
criticized members of the Würzburg School, notably Narziss Ach and H. Watt, for giving an associative explanation
and argued that from the fact that subjects were not aware of intermediate (cognitive) processes occurring between
problem and solution, it in no way followed that the mechanism was associative. On the contrary, he concluded, un-
mediated problem solving is a process of skillfully actualizing dispositional knowledge.

Popper argued, “could adapt themselves through ‘trying-out behaviour’ (probierendes
Verhalten) (Selz), through failure and finally fitness, to the objective situation” (Popper
1933/1979, p. 25).

12

It is only in outlining how this adaptation comes about that full justice is

done to the deductive features of this process of knowledge acquisition. But in that phase of
the sketch, Selz was no longer mentioned, and it was clearly the author of the sketch who
wished to be credited for elaborating the theory. Yet here it is also manifest that he was in ef-
fect advancing Selz’s position. Thus, his remark that the structure of intellectual reactions, like
the belief in causality, can be modeled on the structure of physiological reactions echoed
Selz’s integrative theory of cognitive development (Popper, 1927, p. 25). Likewise, his use of
the centrally important concept of anticipation exactly paralleled Selz’s contention that coor-
dinations differ from associations in that the coordination “is not established by the experi-
enced succession of stimulus and operation, but can precede it” (Selz, 1922, p. 570). Indeed,
the following passage was no more than a reworking of Selz:

If the co-ordination (Zuordnung) between intellectual reactions and objective situations is
established via trying-out behaviour, then the co-ordinations always precede their adequacy
(Bewährung) in time. The co-ordinations therefore are as regards their adequacy anticipa-
tory (as long as the reaction has not proved itself adequate, it can be called an “unfounded
prejudice”). The fulfilment will also often fail to occur: the anticipatory co-ordination be-
tween reaction and stimulus is tentative. Therefore I dub the subjectively pre-formed intel-
lectual reactions shortly “Anticipations.” (Popper, 1933/1979, pp. 25–26)

This was almost an attempt to freeze Selz’s theory of schematic anticipations into obliv-

ion by eschewing all mention of him; indeed, Popper even takes credit for having invented the
concept of anticipation.

13

Having earlier, in 1931, while explicitly referring to Selz’s theory

of memorial knowledge, discarded the idea of the mind being nothing but the conduit for
sense-impressions, he now had finally drawn the epistemological implication that our knowl-
edge of the external world is drawn from our mind, from “trying-out anticipations, which are
co-ordinated tentatively to the ‘material’ of receptions” (Popper, 1933/1979, p. 26). It was this
theory of anticipations that finally enabled him to achieve his “revolutionary revision” of the
concept of habit announced in his thesis of 1927, for by equating still unfounded anticipations
with unfounded prejudices (or dogmas), the early theory of dogmatic thinking was incorpo-
rated into the new evolutionarily oriented theory of problem solving, thereby transforming the
Gesetzerlebnis from an essentially passive and conservative phenomenon, fearfully resisting
what is new, into the initial and indispensable phase of an ultimately critical learning process
continually exploring new territory.

C

ONCLUSION

Contrary to what his autobiography tells us, Popper did not reject induction as a philo-

sophical procedure in the 1920s. The rejection of induction, and the attendant deductive crite-
rion of demarcation between science and pseudoscience, had to await Popper’s development in
psychology from an essentially associationistic and inductive stance toward a noninductive the-
ory of memory, cognition, and learning. Aided by Selz’s theory of schematic anticipations, and

390

MICHEL TER HARK

12. In a footnote, he referred to Selz’s three books on Denkpsychologie.

13. Given Popper’s mention of Selz’s method of trial and error in his dissertation, his reference to the literary au-
thor Bernard Shaw, in particular to his Back to Methuselah: A Metabiological Pentateuch (1921/1998), as a source
of his theory of learning seems to have been more an attempt to suppress his own psychological past than to be his-
torically accurate.

its correlate at the level of learning, the method of trying-out behavior, Popper finally came to
defend the view that before repetition even can set in, the organism must have already actively
put forward dogmatic trials, which, like schematic anticipations, operate in a limited search do-
main and are ultimately defined by their serviceability to life. Drawing on the insights of con-
temporary Denkpsychologie, Popper concluded accordingly that psychologically induction
does not exist. The (psychological) problem of induction therefore was a genuine problem to
the young Popper, prior to and independent of the “two fundamental problems of epistemol-
ogy,” (logical) induction and demarcation. Failing to recognize the autonomy of this psychol-
ogy of knowledge almost automatically leads one to overlook its role in the genesis of Popper’s
philosophical rejection of induction and, hence, in the denial of an inductive criterion of de-
marcation. For rather than having laid the blueprint for an as yet nonexistent deductive psychol-
ogy of knowledge, Popper’s philosophical theory of knowledge clearly emerged in the process
of appropriating and integrating Selz’s theory of anticipations into the deductive operations of
philosophical reason. (For a more detailed explanation, see ter Hark, 2004, chapter 5.) Indeed,
in shaking off his earlier sensualistic psychology of knowledge and replacing it with a theory
of anticipations, Popper also abandoned his inductive stance in methodology and developed his
well-known deductive theory. Accordingly, it was not so much the psychology of anticipations
that was new in Popper as it was the specific linkage he established between, on the one hand,
the classical epistemological notions of synthetic a priori knowledge and, on the other, the
Selzian notion of anticipation. Anticipations, he claimed, are forms of synthetic a priori knowl-
edge, but a priori in a genetic sense only; anticipations can be refuted a posteriori (Popper,
1933/1979, pp. 30–32). It was precisely this (psychological) notion of a priori knowledge, of
anticipations, that Popper appealed to in his attempt, in the philosophical section in Die beiden
Grundprobleme preceding the sketch of a deductive psychology, to force a breakthrough in the
deadlock between classical rationalism and classical empiricism. This made room for one of
his most characteristic (and valuable) ideas, namely, the hypothetical and fallible nature of all
human knowledge. As he put this idea there, the most general axioms of natural science are
formulated without logical or empirical justification, but “in contrast to rationalism they are
not accepted as a priori true (in virtue of their evidence), but as merely problematic, unfounded
anticipations or tentative hypotheses. Their verification or refutation proceeds, strictly empiri-
cal, only on the basis of experience: by deducing propositions (predictions) which can imme-
diately be checked empirically” (Popper, 1933/1979, p. 106).

This “deductive empiricism” was clearly a synthesis of the Selzian emphasis on the ge-

netic priority of tentative and risky anticipations in problem solving and the epistemological
requirement that justifying proceeds on the basis of experience, and hence an unmistakable
sign of the interaction between Popper’s psychological and philosophical theory of knowledge
in the early 1930s (Popper, 1933/1979, p. 106). What all this strongly suggests, therefore, is
that Popper, rather than transferring his deductive empiricism to an unexplored field in the
psychology of knowledge, recognized the deep analogy between Selzian anticipations and
scientific hypotheses and, in particular, its enormous potential for criticizing alternative epis-
temological theories in philosophy.

R

EFERENCES

Bahle, J. (1930). Zur Psychologie des musikalischen Gestaltens, eine Untersuchung über das Komponieren auf ex-

perimenteller und historischer Grundlage. Leipzig: Akademische Verlagsgesellschaft.

Berkson, W., & Wettersten, J. (1984). Learning from error: Karl Popper’s psychology of learning. LaSalle, IL:

Open Court.

Boring, E. (1929). A history of experimental psychology. New York: Appleton.

THE PSYCHOLOGY OF THINKING

391

Bühler, K. (1927). Die Krise der Psychologie. Jena: Verlag von Gustav Fischer.
de Groot, A. (1946). Het Denken van den Schaker. Amsterdam: N.V. Noord-Hollandse Uitgevers Maatschappij.
Groos, K. (1901–1902). Experimentelle Beiträge zur Psychologie des Erkennens. Zeitschrift für Psychologie, 26,

145–167.

Groos, K. (1923). Das Seelenleben des Kindes. Berlin: Verlag von Reuther & Reichard.
Hacohen, M. (2000). Karl Popper, the formative years 1902–1945. Cambridge: Cambridge University Press.
Jennings, H. (1927). Diverse doctrines of evolution, their relation to the practice of science and of life. Science, LXV,

19–26.

Kindler, H. (1929). Über die bedingenden Faktoren und die Erziehbarkeit von Aufmerksamkeitsleistungen. Archiv

für die Gesamte Psychologie, 72, 179–302.

Köhler, E. (1926). Die Persönlichkeit des dreijährigen Kindes. Leipzig: Verlag von S. Hirzel.
Köhler, W. (1925). The mentality of apes (E. Winter, Trans.). New York: The Humanities Press. (Original work pub-

lished 1917)

Krüger, F. (1926). Neue Psychologische Studien, erster Band: Komplexqualitäten, Gestalten und Gefühle. München:

C. H. Beck’sche Verlagsbuchhandlung.

Mach, E. (1896). Die Principien der Wärmelehr. Leipzig: Verlag von Johann Ambrosius Barth.
Newell, A., Shaw, J. C., & Simon, H. A (1958). Elements of a theory of human problem solving. Psychological

Review, 65, 151–166.

Popper, K. (1927). ‘Gewohnheit’ und ‘Gesetzerlebnis’ in der Erziehung: Eine pädagogisch-strukturpsychologische

Monographie. Hoover Institute Archives. Karl Popper Papers, box 12, file 11.

Popper, K. (1928). Zur Methodenfrage der Denkpsychologie. Unpublished doctoral dissertation, University of

Vienna, Austria.

Popper, K. (1931). Die Gedächtnispflege unter dem Gesichtspunkt der Selbsttätigkeit. Die Quelle, 81, 607–619.
Popper, K. (1935). Logik der Forschung: Zur Erkenntnistheorie der moderne Naturwissenschaft. Vienna: Julius

Springer.

Popper, K. (1945).The open society and its enemies (Vol. 2). London: Routledge.
Popper, K. (1963). Science: Conjectures and refutations. In Conjectures and Refutations. London: Routledge &

Kegan Paul.

Popper, K. (1972). Objective knowledge. Oxford: The Clarendon Press.
Popper, K. (1974). Unended quest. London: Routledge.
Popper, K. (1979). Die beiden Grundprobleme der Erkenntnistheorie. Troels Eggers Hansen (Ed.). Tübungen: J.C.B.

Mohr. (Original work published 1933)

Popper, K. (1999). All life is problem solving. London: Routledge.
Seebohm, H. (1970). Otto Selz. Ein Beitrag zur Geschichte der Psychologie. Unpublished doctoral dissertation,

Universität Heidelberg.

Selz, O. (1913). Über die Gesetze des geordneten Denkverlaufs. Stuttgart: Verlag von W. Spemann.
Selz, O. (1922). Zur Psychologie des produktiven Denkens und des Irrtums. Bonn: Verlag von Friedrich Cohen.
Selz, O. (1935). Versuche zur Hebung des Intelligenzniveaus. Ein Beitrag zur Theorie der Intelligenz und ihrer

Erzieherischen Beeinflussung. Zeitschrift für Psychologie, 134, 236–301.

Selz, O. (1964). Die Umgestaltung der Grundanschauungen vom intellektuellen Geschehen (“The Revision of the

Fundamental Conceptions of Intellectual Processes”) In J. M. Mandler & G. Mandler (Eds.), Thinking: From
association to gestalt. New York: John Wiley and Sons. (Original work published 1927)

Selz, O. (1980). Die Gesetze der produktiven und reproduktiven Geistestätigkeit. Kurzgefasste Darstellung (“The

Laws of Cognitive Activity, Productive and Reproductive: A Condensed Version”). In N. Fridja & A. de Groot
(Eds.), Otto Selz: His contribution to psychology. The Hague: Mouton. (Original work published 1924)

Selz, O. (1991). Die geistige Entwicklung und ihre erzieherische Beeinflussung. In A. Métraux & T. Herrmann

(Eds.), Wahrnehmungsaufbau und Denkprozess. Ausgewählte Schriften. Bern: Verlag Hans Huber.

Shaw, B. (1921). Back to Methuselah: A metabiological Pentateuch. New York: Penguin. (Original work published 1998)
ter Hark, M. (1993). Problems and psychologism: Popper as the heir to Otto Selz. Studies in History and Philosophy

of Science, 24, 585–609.

ter Hark, M. (2002). Between autobiography and reality: Popper’s inductive years. Studies in History and Philosophy

of Science, 33, 79–103.

ter Hark, M. (2004). Popper, Otto Selz and the rise of evolutionary epistemology. Cambridge: Cambridge

University Press.

Volkelt, H. (1914). Über die Vorstellungen der Tiere. Leipzig und Bern: Verlag von Wilhelm Engelmann.
Watson, R. (1977). Important psychologists, 1600–1967. In J. Brozek & R. B. Evands (Eds.), R. I. Watson’s selected

papers on the history of psychology. New Hampshire: University of New Hampshire.

Watson, R. (1978). The great psychologists. New York: Lippincott.
Wettersten, J. (1992). The roots of critical rationalism. Amsterdam: Rodopi.

392

MICHEL TER HARK