Extended Critical Review of J. R. Maze, The Meaning of Behaviour. Allen & Unwin,
London, 1983.
John J. Furedy and Diane M. Riley University of Toronto
"... but the basic conceptual parameters of these content areas [of psychology] still call for clarification, and pose many problems of a logical, not merely empirical nature. Unfortunately, professional psychologists, because of the strongly praxis-oriented nature of the discipline, are in large measure uninformed... in their approach to these matters, and bravely hurl themselves into venerable philosophic blunders..." (p.2).
The "professional [research] psychologists" at whom the messages of this book are directed are unlikely to be sympathetic to the philosophical medium employed. They are the heirs of the emancipation of psychology from philosophical bondage, one aspect of which was the assumption that empirical questions in psychology could be resolved solely through logical ("armchair") analysis without controlled observation. The reaction has popularized an equally extreme view, prevalent especially among North American research psychologists, that logical and philosophical considerations are totally irrelevant for "empirical" science, and that philosophical assumptions are like religious ones: personal, not open to critical examination, and matters solely of individual preference. So one finds in many leading empirical psychological journals experiments designed and performed with considerable methodological rigour, but interpreted with concepts that exhibit philosophical sloppiness.
The perceived irrelevance of philosophy for psychology is reinforced by the fact that much of the work in philosophical psychology has had little apparent logical relation to issues that concern research psychologists. One notable exception to this are the writings (mostly in the thirties) and oral communications of the realist philosopher, John Anderson, who held the Challis
1
Chair of Philosophy at the University of Sydney from 1927 to 1958. Anderson's international reputation among professional philosophers is slight, because he published almost solely in Australian journals, and produced no systematic exposition of his views in book form. However, this is essentially available in a collection of his papers (Anderson, 1962), along with a philosophical summary of his position by Passmore (1962). We mention Anderson here because his position has significant implications both for the content and methods of psychological science, and is the background against which John Maze's philosophical contribution to psychology is best understood.
Perhaps the most important characteristic of Andersonianism is its thoroughgoing anti-dualist stance. This is both ontological (in denying the possibility of separate realms of existence, e.g., mind and body) and epistemological (in denying the possibility of different sorts of knowledge or methods of enquiry, e.g., different logics of enquiry and methods of investigation in the physical and psychological sciences). A related tenet is the denial of the doctrine of intrinsic relations, which is the view that there can be a relation, aRb, where being in the aRb relation is intrinsic to, or a property of, R itself. The relevance to psychology of this philosophical issue becomes apparent when the relation is that of knowing. This was extensively discussed by Anderson. Again, the relevance is both to the content of psychology (e.g., modern so-called "cognitive" psychology) and to its form (i.e., whether what is known in psychology--its data--are separable from the knowers, or whether, as is held by most research psychologists, all data is are "theory bound", or conditioned by the psychologists themselves). More generally, on the question of the "cognitive status" of theories (Nagel, 1960), Andersonian realism stands on the extreme realist side of the issue along with philosophers of science like Popper (1959; 1983), and opposed to instrumentalism
as exemplified by Kuhn (1962) [whose doctrine has been enthusiastically espoused by most research psychologists (e.g., Segal & Lachman, 1972), introductory psychology textbooks and cognitive psychologists].
Although Anderson's international impact has remained negligible, his influence on psychologists educated at the University of Sydney was considerable. Of these psychologists, the one who most systematically and effectively argued for the Andersonian approach to psychological issues was John R. Maze. In the early fifties, two seminal papers were published in the Psychological Review, then and now the primary journal of theoretical psychology. One paper (Maze, 1953) attacked the teleological, indeterminist approach as applied to homeostasis, along the anti-dualist lines of argument sketched in by Anderson. The other paper (Maze, 1954) dealt with the famous intervening variable/hypothetical construct distinction that had been recently introduced by MacCorquodale and Meehl (1948). Employing the arguments against constitutive or intrinsic relations, Maze contended that an intervening variable could not intervene, or do anything else, because, by definition, it had no properties. In particular, being an intervening relation does not constitute a property, so that, if we wish to infer the properties of unobservables, these must be specified as hypothetical constructs. Hence the move from hypothetical constructs to intervening variables, which MacCorquodale and Meehl regarded as a sign of explanatory progress, was, in fact, a regress to explanations that were empty.
As former Sydney undergraduates and students of Maze, it is perhaps not surprising that we share his view that the application to psychology of Andersonian realism is both sound and of great importance (Furedy, 1975, 1978; Furedy & Furedy, 1982; Furedy, Riley, & Furedy, 1981). In our view, Maze's two papers referred to above are classics in the field. Unfortunately, as with Anderson's papers, Maze's articles constituted reactions to other thinkers
3
rather than a systematic statement of a position. Furthermore, the two papers are not particularly well known. The paper on homeostasis (Maze, 1953), beyond drawing a somewhat tetchy reply from the person criticised (Stagner, 1954), appears to have had little impact. The paper on intervening variables (Maze, 1954) is one of many hundreds that were written on this topic, following the MacCorquodale-Meehl article, and, as is common in such cases, only the original MacCorquodale-Meehl paper is really well known.
The primary importance of the present book is that it represents the first systematic exposition of Andersonian realism, as it relates to psychology. It is a very cogently and powerfully argued work, with clear, if uncomfortable, implications for many current areas of psychology. Of these the cognitive and information-processing area is perhaps the most influential at present, and Maze provides a good account of some fundamental difficulties (see also Heil, 1981). Social psychology also comes in for attack, especially the recent theoretical accounts that have proposed "agency" as a causal explanatory concept (Gauld & Shotter, 1977; Harre, 1979). It is, indeed, fair to say that Maze takes up most of the fundamental philosophical problems that most psychologists have shelved, and his targets are drawn from the best current research in psychology. For example, Maze's discussion of the conceptual problems inherent in the theorising of eminent current cognitive researchers like Shiffrin and Neisser (pp. 74-5) clearly illustrates the principle that one does not have to have done empirical research in an area (or, indeed, in any area) to be able to offer cogent criticisms, especially when that research rests on assumptions that are philosophically naive.
The emancipation of research psychology from philosophy has resulted in background assumptions being insufficiently examined. The most important function of Maze's book is to provide an examination of many of these
fundamental assumptions. That examination leads him to conclude, with some plausibility, that the assumptions are inconsistent either with other assumptions or with the evidence. For this reason alone, the book is necessary reading for any psychologist who recognizes the need for reflecting about the subject, and who does not view the production of empirical research reports as sufficient for scientific progress.
Our overall assessment of Maze's position is that it is correct in its general Andersonian realist approach, but that it suffers from flaws that appear to stem from not taking certain psychological facts into account. A central tenet of his position is that mental events are relational. This allows a coherent solution to the mind-body problem, and avoids both the dualist and reductionist traps. Also evident is a unified approach to all sciences, as illustrated by a consistent determinist line, and the clear exposition, as well as avoidance, of teleological traps in the explanation of behavior. The two more specific, and--in our view--weak, points in Maze's position are his conceiving of mental events as purely cognitive, i.e., propositional (p. 83, 94), and his almost exclusive reliance on the Freudian account of mental activity.
The extreme cognitivist view of mental events, that all mentation is propositional (e.g., p. 83, 94), is inconsistent with the evidence of non-propositional mentation (e.g., Dreyfus, 1972; Kolers & Smythe, 1979; Searle, 1980). One important set of examples of non-propositional mental processes is formed by S-R or habit-learning processes which are psychological, even if the dependent variables (e.g., the GSR) measured in experiments are often physiological. Maze's implicit exclusion of these S-R processes from psychology is parallel to the attempts of radical behaviorists like Watson to place cognitive processes outside psychology, and is just as "imperialist" (Furedy, 1973) and inconsistent with the facts.
5
Another interesting subclass of non-propositional processes which Maze ignores is the experience of after-images, which--in the sophisticated observer (who does not believe, propositionally, that the after-image is on the surface on which it has been "projected")--is a (non-propositional) reaction by the organism, rather than a cognitive (propositional) process. There are, moreover, other classes of non-propositional processes: affect and motivation. Maze's neglect of these aspects of psychological activity is particularly strange in view of his Freudian emphasis, and he nowhere makes clear why he "leaves aside vegetative processes and reflexes" (p. 83) in his treatment of mental life. At least on the face of it, it appears to be these very processes that are important in accounting for drive-based behaviour. In treating affect and motivation as totally propositional, Maze omits the important reactive or responsive components of feeling and drive, and thereby provides an account of mental activity which is not only not fully detailed (as he admits), but must also remain significantly incomplete.
The fact that any account of mental activity must take motivational factors into account does not mean that the account provided needs be exclusively Freudian. Maze's almost total reliance on Freudian theorising is particularly unfortunate, because there is every indication from his book that he is more than competent to formulate his own theory of human psychological life. The reliance on, and almost unquestioning acceptance of, Freudian theory undermines the cogency of much of his argument on the nature of human mentation. This limitation is particularly apparent in his assertion (p. 163) that everything the organism does is a consumatory act. This assertion is too extreme to be plausible. Maze spends the last part of the book defending it, but to little avail. Perhaps the least convincing part of this line of thought emerges with Maze's defence of the Freudian view that all culture is based on basic
consumatory acts (hunger, thirst, copulation, pain avoidance, and temperature regulation), so that in the activities of art, philosophy, worship, etc., "what we are really doing is one or more of the basic consumatory acts" (p. 165).
There are, of course, many instances in the history of science of positions that appeared to be quite implausible, but that turned out to be correct, while the more plausible alternatives were found to be false. Moreover, the fact that a doctrine may be offensive is no reason to reject it, so the view that painters are just symbolically smearing their own feces around deserves consideration. However, a scientific position needs to be testable, and it is also necessary to provide consistent logical and empirical support. Testability is not a strong feature of Freudian theory (as illustrated, for example, by the existence of the contrary, alternative possibility in this sort of theorising--that when babies smear their feces around, they are, in fact, performing works of art), and Maze also fails to detail the evidence in favor of his radical, Freudian views.
It is to be hoped that Maze's next work will pursue issues of testability and evidence in a more rigorous way. He is aware of this shortcoming, as indicated by his statement that "there is hardly scope here to develop this argument more fully" (p. 173), but it is just that sort of development that is needed, if the strong assertions made are to be adequately supported. The problem is that the link between the "higher" mental activities and instinct is not adequately specified, and his final assertion (p. 176) that everything in social life comes from genetic dispositions because there is no other source is plainly false, because it totally neglects the enormous contribution of learning to human behaviour. Perhaps the implicit defence of this last assertion is that learning cannot be a source, because it in turn depends on genetic factors, but this sort of defence would appear to make the original assertion tautologous. In any case, the main point is that more explicit elaboration is needed for such strong assertions.
7
Another strong, but inadequately supported, claim is that the drives which govern all behaviour are all physiological in nature. This claim is extreme because of the context, wherein Maze focusses almost solely on the basic instinctual drives of hunger, thirst, sex, pain avoidance, and temperature control (p. 165), and neglects the physiological complexity of the human brain, (a complexity which at least allows the possibility of other drives like curiosity). Maze briefly considers this possibility, but dismisses it without adequate examination of the evidence that has accumulated following Berlyne's (1960) initial review. Similarly, the possibilities that "affection" and "social feelings" may have the kind of physiological basis which would qualify them as drives are not given adequate treatment (p. 172).
There are three sets of other problems. The first of these is that a number of the positions could have been worked through most thoroughly. This is especially evident in the last chapter, and particularly as regards the motivational (determinist) basis of disinterested enquiry. We should add that this issue is a particularly knotty one, because it involves the determinist examining his or her own motives.
Another set of problems relates to some of the expressions used. Those of "affection" and "social feelings" (p. 172) are not clearly formulated, which diminishes the usefulness of the consequent Freudian account in which these expressions play a central role. Similarly, the assertions concerning disinterested enquiry lack rigour, as when he writes that "... it may be that the interest in enquiry is especially an expression of the demands of the sexual instinct; but there can also be a general endeavour to overcome self-alienation through achieving a wide-ranging realism". Such statements may be comforting, but they are vague, and suffer from the lack of testability that characterizes much of Freudian theorising.
8
Maze sometimes uses expressions that are unnecessarily prejorative, and would probably be difficult to clearly justify on the basis of evidence. So when he describes Woodfield's position as "quirkily and idiosyncratically formulated" (p. 23), asserts Rosenblueth et al. to show an "uncritical fuzziness" (p.43), and accuses Puccetti and Dykes of having produced a "recrudescence" of a "needless debate" (p. 93), one feels that the objective standards of the realist have slipped somewhat.
More importantly, problems arise from Maze's failure to discuss the question of the relation between his and Anderson's views. Unlike Maze's earlier papers of the fifties, the book makes no reference at all to Anderson's papers. In our view, this is not just a matter of scholarly form. Maze's book is an important contribution in the application of philosophical realism to research psychology, and one wishes to understand in what respects his views are elaborations of, and departures from, the ideas of earlier realists. This is especially important in the case of Anderson, since his views were so sketchily and unsystematically presented. As regards departures, our own guess is that Anderson would not have adopted the sort of cognitive imperialism that Maze seems to espouse.
Another reason why it is unfortunate that Maze has not explicitly commented on the similarities and differences between his and Anderson's positions in this highly systematic book, is that there are few psychologists who are as familiar as Maze is with Anderson's orally expounded ideas. Here, Maze could have made a unique contribution to our understanding of the relation of philosophy to psychology.
Nevertheless, the weaknesses stand in contrast to the many strengths of the book. Maze is particularly skilled in his use of examples to illustrate his points, and employs the argument from counter-example in an intellectually compelling way. To take just two illustrations, consider first his critical
9
analysis of the concept of "goal directedness":
"...Such self-activation and self-direction would certainly distinguish purposive behaviour from mechanical-causal events, but to list them as evidence for purposiveness is a petitio principii: they are not evidence for it but the whole substance of it, as I shall try to show in more detail later. Certainly, self-activation in the full sense of an event that has not external cause but is 'self-generated1 cannot be listed as an objectively observable aspect of behaviour. We can see, say, a sprinter leap from the starting-blocks without being pushed or pulled by anything else, and we can see a rocket hurl itself from its launching pad also without being externally pushed or pulled, but we cannot claim to observe that the former event was uncaused any more than we can claim it about the latter. The causal processes may well be concealed." (pp.13-14).
Our second illustration is taken from Maze's critique of the dualist, purposive views of social psychologists such as Gauld and Shotter (1977) and Harre and Secord (1972). In referring to the work of the latter, Maze first summarises, and then criticises their views as follows:
"...the analysis of social episodes and the explanation of their genesis involves reference to principles of order for the sequences of happening different from those which are used to order items causally which derive from the immediate reference to biological mechanisms for their explanation (Harre & Secord 1972, p. 162).