The purpose of this paper is to try to set the historical record straight while the history in question is still in the making. lt seeks to clarify the relations between gestalt therapy and Gestalt psychology, from which the therapy claims to derive. In considering gestalt therapy, I will confine myself to the work of Fritz Perls, the finder, as he calls himself, of this therapy (Perls 1969/1971:16), with emphasis on his later books. Perls himself writes, in his introduction to the 1969 reprint of Ego, Hunger and Aggression, that much of the material in it is obsolete. About this first book of his he remarks in another place that he wrote it because he wanted to learn typewriting and was bored with exercises (1969/1972:39). About the next book, Gestalt Therapy, by Perls, Ralph E. Hefferline, and Paul Goodman (1951/n.d.), his editor states that Perls regarded it, too, as outdated (Perls 1973:ix). Perls' own comment is in reply to a student who finds its language too technical: 'When did I write that book? In 1951. No, I am much more in favor now of making films and so on to bring this across, and I believe I have found a more simple language' (Perls 1969/1971:233). (In light of this statement, no objection can reasonably be made to the use of transcripts of films and of therapy sessions for an analysis of Perls' work.) My major sources will therefore be Gestalt Therapy Verbatim, In and Out the Garbage Pail , and The Gestalt Approach and Eye Witness to Therapy. In and Out the Garbage Pail might seem somewhat frivolous to the scholar, but Perls, in a conversation with himself, describes it as a serious scientific book (1969/1972:172), which means at the very least, I think, that he would not object to its use as a source in an analysis of his work. (It should be added that one side of the author questions the seriousness of his book.)

Now one more point about the limits of my topic. I will not be concerned with the merits of gestalt therapy as practice, but only with what Perls has written. And I will be concerned with it only insofar as it relates to Gestalt psychology. I will omit discussion of its relations to psychoanalysis, to existentialism, and to other systems of thought, although there is much to say about these too.

lt seems fair at the outset to identify my own point of view, which is that of Gestalt psychology. I do not presume to represent my remarks as what Max Wertheimer, Wolfgang Köhler, or Kurt Koffka would have said about gestalt therapy. The only Gestalt psychologist who, to my knowledge, has written about this therapy is Rudolf Arnheim. His one-paragraph letter to Contemporary Psychology, of course, had no room for analysis (Arnheim 1974: 570). lf the others have maintained silence, why do I now break it? I do so because there are today psychologists and students of psychology - I suspect there are many of them - who believe that gestalt therapy is Gestalt psychology, or, more moderately, that it is an extension of Gestalt psychology. I hope to disabuse them of this belief.

I was astonished to read the statement of Perts' biographer, Martin Shepard (1975: 198) , that traditional Gestaltists claim him. Certainly Arnheim does not claim him when he writes, 'I can see Max Wertheimer fly into one of his magnificent rages, had he lived to see one of the more influential tracts of the therapeutic group in question dedicated to him as though he were the father of it all' (1974:570). Perls himself is at times clearer than his biographer about his relation to Gestalt psychology. 'The academic Gestaltists of course never accepted me,' he wrote. 'I certainly was not a pure Gestaltist' (1969/1972:62). He admits not having read any of their textbooks, only some papers of Kurt Lewin, Wertheimer, and Köhler (ibid). Nevertheless, he claims that his perspective comes 'from a science which is neatly tucked away in our colleges; it comes from an approach called Gestalt psychology' (1969/1972:61). He continues by saying that he admired a lot of the work of the Gestalt psychologists, 'especially the early work of Kurt Lewin' (1969/1972:62).

First may I state the hard facts about his relation to Gestalt psychology. Perls tells us that he was Kurt Goldstein's assistant in Frankfurt in 1926 (1969/1972:4); he apparently also heard lectures by Adhemar Gelb (1969/1972:62). In this connection it may be pointed out that, while Goldstein did not view most of his differences with Gestalt psychology as 'insurmountable discrepancies,' he did not regard himself as a Gestalt psychologist but, rather, a holistic or organismic psychologist.

And now the issues. Gestalt psychology arose in Germany around 1910 out of what was called the Crisis of Science. Not only science, but academic knowledge in general, was losing the confidence of more and more people, intellectuals included, because it could not deal with major human concerns, for example such problems as value or meaning, and, indeed, seemed uninterested in them. In psychology, in opposition to the traditional experimental psychology, there arose a speculative psychology whose goal was to understand rather than to explain. Let the experimental psychologists find causal laws in their narrow domain, so the argument went. The really central human issues must be dealt with outside the natural science tradition, in the tradition called Geisteswissenschaft - a word for which we have no contemporary English counterpart, although it is itself a translation of John Stuart Mill's expression, the mental and moral sciences.

Gestalt psychologists did not accept this split within their discipline. They believed that the shortcomings of the traditional psychology arose, not because it was scientific, but because it misconceived science. Scientific analysis, it was simply taken for granted at the time, was atomistic. The model of the traditional psychology was an atomistic, mechanistic conception of the physical sciences. Gestalt psychologists held that scientific analysis need not be atomistic. Using physical field theory as their model, they worked to develop a non-atomistic psychology within the tradition of natural science.

Here is a first issue: natural science vs. Geisteswissenschaft, explaining vs. understanding. Gestalt psychology is clearly an explanatory natural science. What about gestalt therapy?

Perls equally clearly supports an understanding psychology. Here are a few quotations:

• In scientific explanation, you usually go around and around and never touch the heart of the matter. (1969/1971:16)

• Aboutism is science, description, gossiping, avoidance of involvement, round and round the mulberry bush. (1969/1972:210)

• If we explain, interpret, this might be a very interesting intellectual game, but it's a dummy activity, and a dummy activity is worse than doing nothing. If you do nothing, at least you know you do nothing. (1969/1971:70)

• I reject any explanatoriness as being a means of intellectualizing and preventing understanding. (1969/1972:169)

This theme appears again and again in Perls' books.

It might be supposed that he is talking here about technique, about avoiding interpretations in therapy. He is, of course, also talking about technique, but some of these quotations go much farther. There are other indications of Perls' rejection of scientific psychology. He regards his approach as existential and asserts: 'Existentialism wants to do away with concepts, and to work on the awareness principle, on phenomenology' (1969/1971:16). Again, his approach is described as 'an ontic orientation where Dasein - the fact and means of our existence - manifests itself, understandable without explanatoriness; a way to see the world not through the bias of any concept' (1969/1972:61).

Science, of course, is conceptual.

In other connections, too, we see that Perls is operating outside the sphere of natural science. The structure of our lifescript, he says, is often called karma or fate (1973:120), by no means a scientific concept. Nor is satori (1970/1973.-13), nor 'mini-satori' (1973:131). Hints of vitalism appear in his writing. For example, Perls identifies his 'excitement' with Henri Bergson's élan vital (1970/1973:38). Again, he describes a tree whose roots grow in the direction of fertilizer and shift if the fertilizer is shifted; he comments: 'We cannot possibly explain / By calling this 'mechanics'' (1969/ 1972:28). In this connection, it is interesting to recall a remark by Koffka, 'I believe that the mechanist has no better friend than the vitalist' (1938:226). Perls, unable to account mechanistically for the phenomena of growth and regulation, resorts to vitalism. But science, as the Gestalt psychologists in particular have pointed out, need not be mechanistic; thus the failure of mechanism does not exclude a scientific approach.

In short, we find that Gestalt psychology is a natural science, while Perls - whether he knows it or not - stands in the Tradition of Geisteswissenschaft. It would be interesting to know what science he has in mind when he modestly acknowledges, 'The crazy Fritz Perls is becoming one of the heroes in the history of science, as someone called me at the convention, and it is happening in my lifetime' (1969/1972:265). Gestalt psychology is an explanatory science, while Perls chooses understanding psychology. The difference is so crucial that I could conclude at this point that there is no substantive relation between Gestalt Psychology ind gestalt therapy. Other important issues remain, however.

A related point is the anti-intellectualism that pervades gestalt therapy. 'Intellect,' says Perls, 'is the whore of intelligence - the computer, the fitting game' (1969/1971:24). 'It might sound a bit peculiar,' he concedes, 'that I disesteem thinking, making it just a part of role-playing' (1969/1971:37). 'The intellect . . . [is] a drag on your life' 1969/1971:76). 'Each time you use the question why, you diminish in stature. You bother yourself with false, unnecessary information' (ibid). I could multiply quotations. Gestalt psychologists, on the contrary, have the highest respect for disciplined thinking, one of whose finest achievements is science.

Let us now consider the mind-body problem. Gestalt psychology has formulated the hypothesis of psychophysical isomorphism, both as a position on the mind-body question and as a heuristic. Isomorphism starts from the prima facie dualism of mind and matter but hypothesizes that molar events in experience are structurally identical to the corresponding molar physiological events in the brain. This is a kind of parallelism, but more specific than mere parallelism; it is this specificity that has made isomorphism a powerful heuristic. Parallelism of any kind is, of course, a dualistic hypothesis.

How does Perls stand on this issue? He dismisses the mind-body dichotomy as a superstition (1969/1972:8) and comes out for monism: we do not have a body, he maintains, 'we are a body, we are somebody' (1969/1971:6). 'Thoughts and actions are made of the same stuff' (1973:14). Again, 'If mental and physical activity are of the same order, we can observe both as manifestations of the same thing: man's being' (1973:15). On the whole, he seems to adopt a double aspect theory, though at times his formulation sounds idealistic:

Reality is nothing but

The sum of all awareness

As you experience here and now. (1969/1972:30)

'Philosophizing is a drag,' Perls asserts (ibid). Of course it is if you do it so badly. But the present point is that, with regard to their positions on the relation of the mind and body, Gestalt psychology and gestalt therapy have nothing in common.

'Figure/ground, unfinished situation and Gestalt are the terms which we have borrowed from Gestalt psychology,' say Perls, Hefferline, and Goodman (1951/n.d.:ix-x). lt is time to examine the meanings of these terms in the two contexts.

For the meaning of Gestalt, I quote Köhler:

In the German language ... the noun 'Gestalt' has two meanings: besides the connotation of shape or form as an attribtite of things, it has the meaning of a concrete entity per se, which has, or may have, a shape as one of its characteristics. Since Ehrenfels' time the emphasis has shifted from the Ehrenfels qualities to the facts of organization, and thus to the problem of specific entities in sensory fields. ( Gestalt Psychology, 1947:177-178)

Perls' use of the term Gestalt is much vaguer. His attitude toward it he describes as an article of faith (1969/1972- 35). A gestalt is an essence, he says (1969/1972-63). Again, he describes it as 'the irreducible phenomenon of all awareness' (1969/1972:30). Perls recognizes that a gestalt is a unit of experience, that 'as soon as you break up a gestalt, it is not a gestalt any more' (1969/1971:16). But he does not go any farther into the description of its properties. Neither Perls' Gestalt Manifesto (1969/1972:213) nor his old Gestalt Prayer has any relation to any known meaning of the word Gestalt.

A segregated entity possesses figural characteristics: shape and the substantiality of a thing by contrast with its background, which usually has no shape and is less compact. It owes its shape to the one-sided function of the contour, which ordinarily belongs to the figure, but not to the ground. There are other functional differences, too, between figure and ground. Although perceptual figures may be reversible under certain circumstances, this is not the rule.

Edgar Rubin's terms "figure" and "ground" were eagerly adopted by Perls. For example, "The dominant need of the organism, at any time, becomes the foreground figure, and the other needs recede, at least temporarily, into the background" (1973-8). lt may be that needs possess the characteristics of shaped figures, but if so, this must be shown, not simply assumed. (More likely, it is the need-object organization that should be subjected to such analysis; the goal, as end, is comparable to the edge of a closed figure, as Köhler [1939:79] has pointed out.) Without any analysis, Perls seems simply to be using the distinction between figure and ground as equivalent to that between important and unimportant. While the figure is important in the perceptual field, it has its own specific properties that are lost in the equation. And why do you need figure-ground terminology to say that something is important?

To change a habit involves pulling that habit out of the background again and investing energy . . . to disintegrate or to reorganize the habit' (1969/1972:66). This time Perls apparently means - focus attention on the activity usually performed automatically. I have no doubt that it is possible to conceptualize an activity sequence in Gestalt terms, but Perls has not done it - he has merely used the words. lf his expression is equivalent to Rubin's distinction, this remains to be shown.

Perls asserts that ritual 'makes the gestalt clearer, makes, the figure stand out more sharply' (1973:29). The meaning is apparently once more that the special importance of something is being emphasized. I need not repeat my remarks about importance. But what is the figure that is made to stand out by a handshake or a toast? Perhaps the handshake emphasizes the beginning or the end of an encounter, but what is the structure of the encounter? The use of figure-ground terminology is no substitute for specifying the characteristics of a social event.

At one point Perls tells us that he is bogged down in his writing and remarks, 'I would not be a Gestaltist if I could not enter the experience of being bogged down with confidence that some figure will emerge from the chaotic background' (1969/1972:37-38). What he means, it would seem, is that he is sure he will find something to say. Again, what is gained by speaking of figure? What is lost, I repeat, is the specific meaning of figure and ground. Incidentally, a chaotic background is hardly conducive to the segregation of a figure.

Perls finds it important that figure and background be easily interchangeable. 'Otherwise we get a disturbance in the attention system-confusion, loss of being in touch, inability to concentrate and to get involved' (1969/1972:93). lt has been pointed out earlier that in perception reversible figures are the exception. From the context it appears the Perls means that, for optimal functioning, there must be an alternation between what he calls coping and withdrawal, there must be flexibility of the personality, and the like; but what these have in common with figure and ground in the sense of Rubin and the Gestalt psychologists is never made clear.

In all these examples, and many others that might be discussed, it seems to me that the figure-ground terminology is used so loosely by Perls that it conceals problems rather than clarifies them.

Since Gestalt psychologists emphasize organization, let us turn to that problem. As Köhler puts it, organization 'refers to the fact that sensory fields have in a way their own social psychology' (1947:120). That is, certain units or groups exist which are relatively segregated from their environment: certain parts of, say, the visual field belong together and are segregated from others. Wertheimer investigated the factors that govern perceptual organization: similarity, proximity, good continuation, closure, etc.

Of Wertheimers factors of organization, the only one in which Perls shows any interest is closure and lack of closure. The latter term he uses interchangeably with 'unfinished situation' - a technique, not a concept, derived from Lewin. Let us consider some examples of unclosed gestalts as they are used in gestalt therapy.

'Our life is basically practically nothing but an infinite number of unfinished situations-incomplete gestalts.' writes Perls. 'No sooner have we finished one situation than another comes up' (1969/1971:15). The neurotic 'individual somehow interrupts the ongoing processes of life and saddles himself with so many unfinished situations that he cannot satisfactorily get on with the process of living' (1973:23). These unfinished situations from the past compel him to repeat them in everyday life (1973:91). (Incidentally, Freud's repetition compulsion is here made a matter of unclosed gestalts without, so far as I can see, shedding any light on it.) If we find a certain plausibility, along with a disdain for specific analysis, in the treatment of unsatisfied needs as unclosed gestalts, this plausibility is lost in further examples. In the case of one patient, Perls remarks, that he was unable in one session to 'achieve full closure, milk the symptom dry' (1969/1972: 139). War, with its frustrations, is apparently an incomplete gestalt; at any rate, peace is the possible closure (1969/1972:87).

Here is a final example of the many Perls provides: 'We . . . have to fill in the holes in the personality to make the person whole and complete again' (1969/1971:2). I happen to believe that the phenomenal personality, like other percepts, can he conceptualized as an organized whole, though the theoretical problems involved are extraordinarily difficult and only the most primitive beginnings have been made - not, by the way, by gestalt therapists. Until we can say something specific about this organization, it does not add to our knowledge to say that 'the neurotic man of our time' is an 'incomplete, insipid personality with holes' (Perls 1969/1972:294). As I have indicated, in some of these instances there is a certain vague plausibility about Perls' use of complete and incomplete situations, closed and unclosed gestalts. But vague plausibility is not enough for a theory of neurosis or therapy or personality - or of anything. lt is necessary to be clear about the specific characteristics of the structure we are calling neurosis or personality, about the nature of the processes involved, and the nature of the closure demanded by that structure. Such questions are never found in the material I am considering, and we are left with a terminology so vague as to defy any specific use. A concept loosely applied to a perceived figure, to a neurotic personality, and to war does not shed any specific light on any of these phenomena. For a theory, we must also be able to say in what ways the perceived figure, the personality, and the war are different, not merely stretch the same term to include them all.

The following is a passage from Köhler on the extension of the concept of Gestalt:

The concept 'Gestalt' may be applied far beyond the limits of sensory experience. According to the most general functional definition of the term, the processes of learning, of recall, of striving, of emotional attitude, of thinking, acting, and so forth, may have to be included.... By no means is it believed, however, that any of those larger problems can actually be solved by the application merely of general principles. On the contrary, whenever the principles seem to apply, the concrete task of research is only beginning; because we want to know in precisely what manner processes distribute and regulate themselves in all specific instances. ( Gestalt Psychology, 1947:178-179).

lt is this crucial step - the working out of the Gestalt concept in connection with specific problems - that Perls has omitted. He does have some things to say - at times, it seems, almost inadvertently - about how organization occurs, and it is interesting to compare these remarks, with the formulations of the Gestalt psychologists. The conditions of organization suggest to the Gestalt psychologist what processes must be responsible for them. In accordance with the principle of isomorphism, the demonstrated relational properties of perception (and of other psychological phenomena which I will not discuss here) suggest corresponding physical interactions in the nervous system, particularly in the cerebral cortex. These interactions depend on the properties of the cortical events in relation to each other (Köhler 1940:55); and these properties, in turn, are ultimately largely a consequence of the nature of the stimulation that starts the chain of events leading to perception.

For Perls, interest, cathexis, motivation, or attention produces organization. This view appears in his first book (1947/1969:53) and is more explicit in Gestalt Therapy. We read, 'The figure/ ground contrast . . . is . . . the work of spontaneous attention and mounting excitement' (Perls, Hefferline, and Goodman 1951/n.d.:73). Again,''Objects' of sight and hearing exist by interest, confrontation, discrimination, practical concern' (1951/n.d.: 372n). What would seem to be a motor theory of perception is, at times, assumed: 'The eyes and fingers cooperate in drawing outlines, so that the animal learns to see more shapes and to differentiate objects in his field. By outlining one differentiates experience into objects' (1951/n.d.:312). In another place Perls suggests that 'we start with the impossible assumption that whatever we believe we see in another person or in the world is nothing but a projection. Might be far out, but it's just unbelievable how much we project and how blind and deaf we are to what is really going on' (1969/1971:72). Although he does not hold with it completely, Perls seems to be saying that this assumption has something to it. The statement is less radical, but the meaning essentially unchanged, when he tells us that cathected objects become figure (1973:19). Once more, it is asserted that things-by which I assume he means phenomenal things - 'come about, more or less, by man's need for security' (1970/1973:20).

It is difficult to discuss Perls' theory because we are not told on what the interest, attention, and cathexis are acting to produce percepts. lt is certainly not on organized entities, since they do the organizing. Presumably, therefore, they are acting on sensory data. If this is the case, Perls' (partially implicit) theory is not only not Gestalt psychology; it is formally similar to the theories that Gestalt psychologists have criticized again and again, ever since Köhler's paper of 1913, 'On Unnoticed Sensations and Errors of Judgment' (1913/1971). Indeed, Perls' theory, if it were spelled out, would seem to be very similar to those put forth by G. E. Müller and Eugenio Rignano in the 1920s, both of which were criticized by Köhler. About such theories it may be said that neither attention nor interest creates form; rather, a form must be perceived before it can be attended to or cathected. In both cases, the directional process presupposes the organization; the argument is thus circular. A similar problem arises if a motor theory is really meant: if visual organization comes from kinaesthesis, then that kinaesthetic organization remains to be explained. All the theory has succeeded in doing has been to push the problem into another sensory modality.

lt is not necessary, so far as I can see, that a theory of therapy include a theory of perception. But if the author insists on such a theory, there are certain known pitfalls he would do well to avoid. If he believes that his theory is a Gestalt theory, he would be well advised to look into what the Gestalt psychologists have to say.

Gestalt psychology is most developed in perception and cognition, while gestalt therapy is concerned with personality, psychopathology, and psychotherapy. Comparison of approaches to such different areas is often difficult. Nevertheless, in the present case, additional issues invite comparison. As it happens, none of them is trivial.

Gestalt psychology has, from its inception, been interested in value. Challenging the widely held view of ethical relativism, the view that what is right and wrong changes with time and place, it has tried to understand values in terms of relations within happenings themselves. The value of an action is seen as depending on its appropriateness to the demands of the given situation. Thus, Gestalt psychologists have held that values are not arbitrarily attached to objects or actions, depending on subjective evaluation or on the individual's history of rewards and punishments. An analogy of Wertheimer's will perhaps be helpful:

Someone in adding makes seven plus seven equal fifteen. ... And he says, I call it good because I love the number fifteen.... The determination of the fifteen is ... in violation of that which is demanded by the structure of the objective situation. If I prefer the fifteen in this case ... this is irrelevant to the fact that the fifteen is wrong. 1935:360-361)

What about Perls? In Ego, Hunqer and Aggression, ethical relativism is simply taken for granted, and good and bad are derived from feelings of comfort and discomfort (1947/1969:59). The next book, Gestalt Therapy, describes two ingredients of moral evaluations: ' (a) On the one hand, they are simply technical skills that one has learned, guesses as to what leads to success' and '(b) On the other hand, they are group-loyalties . . . : one acts in a certain way because it is the social expectation, including the expectation of one's formed personality' (Perls, Hefferline, and Goodman 1951/n.d.:424). Here values are obviously regarded as external to the events in question. They might just as well be reversed if the individual's personal history had been different or if he belonged to a different group.

The same relativism, more baldly and more cynically expressed, is to be found in Gestalt Therapy Verbatim: 'The whole idea of good and bad, right and wrong, is always a matter of boundary, of which side of the fence I am on' (1969/1971:9). Perls distinguishes three kinds of philosophy: in addition to existentialism, which includes gestalt therapy, we have already encountered aboutism, encompassing science, gossiping and other futile activities, and then there is shouldism or moralism, in which we find topdog and underdog engaged in self-torture games (1969/1971:16-18). Shoulds are internalized external controls, and they interfere with the healthy functioning of the organism (1969/1971:20).

It would be difficult to find a view of values farther from that of the Gestalt psychologists than Perl's view. The Gestalt psychologists have shown that '"value-situations fall under the category of gestalt' (Köhler 1938:86). Perls has treated them without regard for this category, indeed without regard for values.

A word about truth. Apart from calling it one of the fitting games, Perls says that 'by 'truth' I mean nothing but the assertion that a statement we make fits the observable reality' (1970/ 1973:13). This conception is precisely the one that Wertheimer has shown to be inadequate. For the same statement may, in one context, be true, in another false, in a third unintelligible. Nor does Wertheimer regard truth as a game: 'Science is rooted in the will to truth. With the will to truth it stands or falls. Lower the standard even slightly and science becomes diseased at the core. Not only science, but man' (1934:135).

I have already mentioned the relation between mechanism and vitalism. Gestalt psychology has consistently rejected both. Machine theories of the nervous system have been its particular target: Gestalt psychology has emphasized free dynamics within the limits imposed by anatomical constraints. Perls, quite the contrary, refers to the organism as a machine (1969/1971:15), and to the 'thinking system,' as he calls it, as a computer (1970/1973:28-29).

I would now like to say a word about phenomenology as it figures in Gestalt psychology and in gestalt therapy. (I am using the term 'phenomenology' as psychologists generally do, to refer to the unbiased description of the phenomenal world, not to refer to Edmund Husserl's theory of intentionality.) For Gestalt psychology, phenomenology is a first step, a propaedeutic to experimental research and to a science of functional relations that transcends phenomenology. Perls calls himself a phenomenologist (1969/1972:37)-, for him this method plays a different role than in Gestalt psychology. Phenomenology, he says, 'is the primary and indispensable step towards knowing all there is to know' (1969/1972:69).

I have by no means exhausted my material. For example, Perls' misuse of the equilibrium concept might be discussed. His understanding of heredity and of evolution might be culled from his writings and contrasted with that of Gestalt psychology. His view of person perception, like that of object perception, could be shown to differ from that of the Gestalt psychologists. His mostly implicit conception of the thinking process might be examined, and so on.

From the material already discussed, it is not difficult to reach a conclusion. What Perls has done has been to take a few terms from Gestalt psychology, stretch their meaning beyond recognition, mix them with notions - often unclear and often incompatible - from the depth psychologies, existentialism, and common sense, and he has called the whole mixture gestalt therapy. His work has no substantive relation to scientific Gestalt psychology. To use his own language, Fritz Perls has done 'his thing'; whatever it is, it is not Gestalt psychology.