1
The Historical Roots of Popperās Theory of the Searchlight
A Tribute to Otto Selz
As a student at the Pedagogic Institute of Vienna between 1925 and 1927, the young Popper took courses from Karl Bühler at the Psychological Institute. By then Bühler was a Viennese guru who attracted a large number of students and followers, among them the sociologist Paul Lazarsfeld and the Gestalt psychologist and philosopher of science Egon Brunswik. Even Konrad Lorenz took courses under Bühler. By his own account, Popper learnt much from Bühler, and he always credits him for his important theory of language. Bühler was a generalist, writing on psycholinguistics, perception, and child development, yet it is his Habilitation (Bühler 1907) on the psychology of thinking, supervised by the founder of the so-called Würzburg School of Denkpsychologie, Oswald Külpe, for which he is primarily remembered in the history of psychology. This work is also the starting point for Popperās dissertation under Bühler in 1928, Zur Methodenfrage der Denkpsychologie. Popper never published his dissertation, and in Unended Quest he recalls that it was āa kind of hasty last minute affairā (Popper 1974/1976, § 15).
Another psychologist Popper mentions in his autobiography, Otto Selz, was closely related to the Würzburg School. Selzās name occurs much less frequently in histories of psychology, but it becomes increasingly clear, mainly owing to A.D.de Grootās (1965) application of his ideas to the study of chess thinking, that he was by far the most original and important precomputational cognitive psychologist. The young Popper must have learnt from Selz via Bühler. Referring to his own experiments in the psychological laboratory, Popper, in Unended Quest, remarks about Selz: āFinding that some of my results had been anticipated, especially by Otto Selz, was, I suspect, one of the minor motives of my move away from psychology.ā (Popper, 1974/1976, § 15) These results are that humans ādo not think in images but in terms of problems and their tentative solutionsā (ibidem).
In their pioneering work Learning from Error (1984), William Berkson and John Wettersten were the first to argue in detail that Popperās later philosophy of science rests on an equally interesting psychological theory. In their little book, they compare Popperās psychological theory with the theories of Selz, Gestalt psychology, and Piaget, without however drawing on historical sources such as Popperās unpublished writings on psychology or specific authors read by him. When I began to study Popperās unpublished dissertation in the early 1990s, I discovered, contrary to Berkson and Wettersten, that his psychological and epistemological theories were far from original and enormously indebted to German Denkpsychologie, in particular to Otto Selz. Unfortunately, my article on Popper and Selz (ter Hark 1993) was not much read. Recently (ter Hark 2002) I studied Popperās psychological work preceding his dissertation, āGewohnheitā und āGesetzerlebnisā in der Erziehung. Popper submitted this work as a protothesis in the summer of 1927 to the Pedagogic Institute of Vienna. In fact it is not just a protothesis. As the subtitle, A structural-psychological monographā, indicates, his ambition was to write a full monograph, which, however, remained incomplete. In his Conjectures and Refutations: The Growth of Scientific Knowledge (1963, Chapter 1), Popper translates the title of his thesis as On Habit and Belief in Laws, and maintains that it contains his logical criticism of Humeās psychological theory of the genesis of belief, his bucket theory of mind and knowledge. To my surprise the thesis turned out to be in no way involved with the psychological problem of induction. Moreover, his reliance on the sensualistic framework provided by Richard Avenarius only further strengthened my idea that Popper, in 1927, was still far removed from recognizing the defects of a bucket theory. Indeed, prior to his reading of Selz, Popper simply had no deductivist or quasi-deductivist psychology of problem solving. His reading of the work of Selz therefore marks the real watershed in his intellectual life, since it ultimately led him to abandon the inductive (sensualistic) psychology of knowledge. My aim in this paper is to argue in detail how the young Popper gradually transformed Selzās theory of problem solving into his well-known deductivist theory of knowledge, his theory of the searchlight. In particular, I argue that this transformation spans a period of three years, beginning with his dissertation in 1928, taking more definite shape in a short article on mnemonic exercise in 1931, and culminating in Die beiden Grundprobleme der Erkenntnistheorie, written in 1930ā1933 (but not published until 1979).
1 Otto Selz and the Würzburg School
The historical background and genesis of Otto Selzās theory of reproductive and productive thinking still has to be written, but what is clear so far is that his theory both elaborates upon and criticizes the programme of the Würzburg School. This programme, headed by Külpe, was already quite revolutionary. The method of the Würzburg School is called experimental introspection because it combines introspection and a method of testing. Subjects were asked not only to respond to a question but also to describe their state of mind during the test. For several reasons applying the experimental method to higher forms of cognition was a daring enterprise in those days. Sensations, emotions, reflexes, memory, and association were the proper subject of experimentation, and it seemed far from likely that it could contribute to the study of thought. Moreover, judgement and reasoning were before long considered properly a part of logic and epistemology, normative rather than empirical studies. Perhaps the insurmountable obstacle for a psychology of thinking was the refusal to recognize any other form of experience than sensations, images, and fragments of images common to the sensualism of Berkeley, Hume, Condillac, and Ernst Mach.
The researches of Külpeās graduates were given direction and point precisely by his determination to show that it is impossible to analyse thought into sensory elements, thereby discrediting Mach, and to contribute to the experimental analysis of thought, thereby departing from Wundt. As Külpe summarizes the first achievements of the School, these were largely negative: the traditional contents of consciousness, sensation, feeling, and images ā the very substance of Wundtian psychology ā proved inadequate to account for the intellectual processes of thoughtful association and judgement. Yet subjects frequently reported that they experienced certain conscious processes which they could describe neither as definite images nor as acts of will. To these impalpable experiences, which could not be classified under any of the standard categories, various names were given, but the name best remembered ā imageless thought ā came from a critical reviewer of the programme, Edward Titchener (1909). The significance of these Würzburg experiments is not so much the disproof of images, for that was not attempted; images do exist and are part of thinking. Their real significance is that they demonstrated that thought consists not solely of images and, more importantly, that there are (unconscious) regulatory and selective mechanisms in thinking. Thus Heinrich Watt (1905) conducted experiments in which he demonstrated the role of the Aufgabe, or task, as a directive influence in the problem he set for naming super-ordinates for sub-ordinates and parts for wholes. At the same time Narziss Ach (1905) dubbed the directive influence of the task on the outcome of thinking ādetermining tendenciesā. According to Ach, determining tendencies explain the ordered and purposeful character of thought processes; they rule out irrelevances and prevent chance stimuli from distracting the course of thought processes. They accomplish this by favouring those associations that are in line with the purpose of the subject. For instance, if the instruction is given to add two numbers, the subjectās representation of this goal (Zielvorstellung) will influence the particular stimulus presented. Thus, given 6; 2, the answers 8, 4, or 3 will result according to the goal-representations (corresponding to the instructions) of respectively, adding, subtracting, and dividing. But although critical of association psychology, the regulatory systems postulated by Watt and Ach were still associative in nature.
Although his experimental set-up was in the tradition of the School, Selzās theoretical work deviates from the School in two respects. The first difference is that Selz does not adopt even a modified version of associationism but rejects it completely. The second difference is that in Selzās work the explanandum of psychology is shifted from the content of thinking to the process of thinking. While the elder members of the Würzburg School set themselves the phenomenological task of describing and analysing thought experiences as a mental category sui generis, Selz, without denying the importance of imageless thought, believed that the essence of thinking is to be found in a series of āoperationsā. Selz, de Groot summarizes, ā⦠read protocols in a different way: he searched for the procedures (methods) by which the subject made progress ā¦ā (de Groot 1965, p. 51).
After his PhD, Otto Selz (1881ā1943) went to Bonn where he participated in the seminars of Külpe and Karl Bühler. Above all he was engaged in experimental investigations in the laboratory of Külpe. These investigations resulted in his first major work, his Habilitationsschrift, Ćber die Gesetze des geordneten Denkverlaufs. Eine experimentelle Untersuchung (1913). Taking his cue from Bühlerās theory of imageless thought, Selz, according to Külpe, made a significant step forward in the psychology of thought.1 In fact Selzās drift away from the programme of the Würzburg School was more radical than Külpe acknowledged. Already perceptible in 1910, the incipient rift among Selz and the Würzburg School became more obvious in the wake of a devastating review of Achās book on willing (Selz 1910).
With his second major work in the psychology of thought, Zur Psychologie des produktiven Denkens und des Irrtums (1922a), whose publication was postponed owing to the First World War, Selzās intellectual prestige was incontestably on the increase, and in 1923 he was called to the chair of Philosophy, Psychology, and Pedagogy at the Handelshochschule in Mannheim (Baden). From this period too stem two of his short philosophical essays, Oswald Spengler und die intuitive Methode in der Geschichtsforschung (1922b) and Kants Stellung in der Geisteswissenschaft (1924a), in which he attempts to bridge the gap between the natural sciences and the Geisteswissenschaften by means of his naturalistic and evolutionary epistemology.
In his scientific work Selz was increasingly marginalized because of his unremitting criticism of colleagues but also because of his formidably complex style of writing. A biologist disguised as a psychologist, Selz came into conflict with proponents of the Geisteswissenschaften, who blamed him for endorsing a mechanist view of man. Seeking to reconstruct psychological wholes on the basis of their elements, Gestalt psychologists considered him an atomist, whereas to others he was a one-sided rationalist. Closely allied to the Würzburg School, he did not shrink back from launching frontal at...