Skip to main content Accessibility help

The Limits of Behaviorism: A Review Essay on B. F. Skinner's Social and Political Thought

  • Peter G. Stillman (a1)


After a brief review of the debate that surrounds B. F. Skinner's social and political thought, this essay argues that, even if Skinner's behaviorism is accepted as valid, nonetheless his social and political thought is inadequate and self-contradictory. Skinner's treatment of social organization (or ’sociological problems”) contains two serious, perhaps insoluble, technical difficulties. His treatment of values (or “philosophical problems”) is fraught with contradiction, because he cannot meld what is valued with what leads to survival, because his planners can easily misperceive the objective world, and because he cannot justify cultural survival as the ultimate value. On “political problems,” Skinner is vague and incomplete; he proposes no effective means of “countercontrol”; and his behaviorism does not contain within itself the imperative to any particular type of political system. Because Skinner's social and political thought fails at crucial points, all the problems that Skinner tried to close must be kept open and—whatever may be the technical successes and failures of behaviorism—the answers to some fundamental issues of human life must be sought beyond behaviorism.



Hide All

This review essay owes much to others, especially faculty and students at Vassar College, where many of the points were first presented in a class. An earlier version was delivered as a lecture at Philosophers' Holiday, and copies of the lecture were freely distributed. As a result of the many comments and criticisms that ensued, arguments were modified and errors removed. I would like to thank the many at Vassar who were helpful, and especially Jesse Kalin, Michael Murray, Suzanne Vromen, Nancy Downing, Ed Jacobs, and Kim Landsman. In addition, I would like to thank Emily Gill and P. Gordon B. Stillman, who read and commented on this essay at various stages. None of them, of course, is responsible for the problems and infelicities that remain.



Hide All

1 B. F. Skinner's social and political thought is presented primarily in his Walden Two (1948; paperback ed., New York: Macmillan, 1960), Science and Human Behavior (1953; paperback ed., New York: Free Press, 1965), and Beyond Freedom and Dignity (New York: Knopf, 1971); it is defended in About Behaviorism (New York: Knopf, 1974). All citations to these books are in the text; the titles are abbreviated to W2, SHB, BFD, and AB, respectively. (Only with BFD is there a disparity in pagination between the paper and cloth editions; a rough way to reconcile this disparity is to subtract 5 per cent from the cloth edition's page number in order to obtain the paperback's page number.) These sources can be supplemented by numerous articles (e.g., Skinner, B. F., “Freedom and the Control of Men,” The American Scholar, 25 [Winter 19551956], 4765) and by many published interviews: Evans, Richard I., B. F. Skinner: The Man and His Ideas (New York: Dutton, 1968); An Interview with B. F. Skinner,” The Center Magazine, 5 (Mar./Apr. 1972), 6365; Hall, Elizabeth, “Will Success Spoil B. F. Skinner?Psychology Today, 6 (Nov. 1972), 65–72 and 130. In addition to About Behaviorism, Skinner “responds” to his critics both in parts of the interviews and in a chapter, Answers for My Critics,” in Beyond the Punitive Society, ed. Wheeler, Harvey (San Francisco: W. H. Freeman, 1973). Skinner's, important technical works include Verbal Behavior (New York: Appleton-Century-Crofts, 1957); Contingencies of Reinforcement (New York: Appleton-Century-Crofts, 1969); and Cumulative Record (Third edition; New York: Meredith, 1972). (Since Chomsky's review of Skinner's Verbal Behavior is so famous [Chomsky, Noam, review, Language, 35 (March 1959), 2658], a good defense of Verbal Behavior might be noted: MacCorquodale, Kenneth, “On Chomsky's Review of Skinner's Verbal Behavior,” Journal of the Experimental Analysis of Behavior, 13 [January 1970], 8399.) A complete (through 1969) bibliography of Skinner's publications can be found in Festschrift for B. F. Skinner, ed. Dews, P. B. (New York: Meredith, 1970), pp. 2327. For summaries of behaviorism in general, see especially Mikulas, William L., Behavior Modification: An Overview (New York: Harper & Row, 1972) and Bandura, Albert, Principles of Behavior Modification (New York: Holt, Rinehart and Winston, 1969).

2 About eighty per cent were critical, according to Skinner, (Center Magazine, Mar.]Apr. 1972, p. 63). The generally favorable reviews stress the validity of Skinner's evidence, the cogency of his reasoning, or the desirability of his goals. Four review symposia contain favorable reviews: Review Symposium,” American Journal of Sociology, 78 (Nov. 1972), 696708; Skinner: Pro and Con,” Contemporary Sociology, 1 (Jan. 1972), 1929; B. F. Skinner's Beyond Freedom and Dignity,” Soundings, 55 (Fall 1972), 335368; and Self-Determination: Fact or Artifact?Philosophical Studies [Ireland], 21, 3183. Other sympathetic or enthusiastic discussions can be found in Day, Willard F., review, Contemporary Psychology, 17 (Sept. 1972), 465467, and in some chapters of Beyond the Punitive Society, ed. Wheeler, esp. John R. Platt, “The Skinnerian Revolution” and Dennis C. Pirages, “Behavioral Technology and Institutional Transformation.”

3 Critical reviews of Skinner's work abound. The most cogent recent ones, which develop many of the points mentioned in this paragraph as well as others of importance, are: Champlin, John R., “Behavior (-ism & -alism) & Theory (Political),” Polity, 5 (Winter 1972), 243249; Kateb, George, “Toward a Wordless World,” The Atlantic, October 1971, pp. 122125; Margolis, Joseph, “Freedom as Determinism,” Society [Transaction], 9 (Sept./Oct. 1972), 8090; Spitz, David, “The Higher Reaches of the Lower Orders,” Dissent (Spring 1973), pp. 243269; and Watts, Meredith, “B. F. Skinner and the Technological Control of Social Behavior,” American Political Science Review, 69 (March 1975), 214227. Two highly critical reviews in psychology journals are Colman, Andrew M., review, British Journal of Psychology, 64, (Feb. 1973), 150151; and McCall, Raymond J., “Beyond Reason and Evidence: the Metapsychology of Professor B. F. Skinner,” Journal of Clinical Psychology, 28 (Apr. 1972), 125139. The review symposia cited in note 2 contain critical articles which raise numerous issues. Beyond the Punitive Society, ed. Wheeler, also contains many critical reviews, of widely varying cogency; see especially the chapters by Nathan Rotenstreich, “Skinner and ‘Freedom and Dignity’,” and by Max Black, “Some Aversive Responses to a Would-Be Reinforcer.” Unfortunately, the two best-known reviews are flawed. Chomsky, Noam (“The Case Against B. F. Skinnner,” New York Review of Books, 30 December 1971, pp. 1824) misunderstands or misrepresents Skinner at various points, as does Sennett, Richard (review, The New York Times Book Review, 24 October 1971, pp. 1 and 1218); their mistreatment of Skinner's sentence on p. 155 of Beyond Freedom and Dignity is but one minor example (see Skinner, B. F., “Letter to the Editor,” New York Times Book Review, 21 November 1971, p. 50, and Sennett's continuing blind response). Less recent but still acute analyses of Skinner's thought include Hacker, Andrew, “Dostoevsky's Disciples: Man and Sheep in Political Theory,” Journal of Politics, 17 (Nov. 1955), 590613, and Kariel, Henry S., “The Political Relevances of Behavioral and Existential Psychology,” American Political Science Review, 61 (June 1967), 334342, Continuing examination of behaviorism can be found in the journal Behaviorism, 1- (1972-).

The above treatments generally criticize Skinner from outside his analytic framework and behaviorism, and are thus rejected by Skinner and many behaviorists regardless of the validity of their criticisms. Only one article has attempted the sort of internal refutation that this review essay attempts: Scribner, Phillip H., “Escape from Freedom and Dignity,” Ethics, 83 (Oct. 1972), 1336, which brilliantly proves that Skinner's goal—the survival of the culture—requires for its realization not Skinner's automatically good man but rather a man who has been conditioned to be a rational decision maker (and who can therefore take into acount new phenomena, for instance) in a society of free institutions (so that he can act on his rational decisions). While this review essay is consistent with Scribner's argument, it does not rely on nor use it in any way.

4 Skinner, , “Answers for My Critics,” pp. 256257; see also AB, p. 7.

5 Conversely, of course, the refutation of Skinner's social and political ideas is not a refutation of his psychology.

6 This statement leaves open the applicability of behavior modification in institutions like schools, prisons, etc., where, however, even Skinner sees problems (BFD, pp. 171–172; AB, p. 191).

7 The problems here may be insurmountable. Skinner's discussion of “Social Behavior” in Science and Human Behavior is, according to a sympathetic reader, “very thin, indeed,” even though it only deals with two-person situations: “Two-person behavior is indeed complicated, and it presents difficult problems for operant conditioning. The same behavior … may be negatively reinforcing to one [person] and aversive to the other. Adding a third person, or an nth’ person, enormously complicates the situation, so that it virtually defies analysis.” (Wheeler, Harvey, “Introduction: A Nonpunitive World?” in Beyond the Punitive Society, ed. Wheeler, , p. 12). Another way of phrasing the problem is to analyze the Skinnerian “environment,” which Skinner never closely examines. Skinner assumes that environments are readily modifiable for the purpose of shaping behavior; but, since much of the human environment is the behavior of others, to shape the behavior of one person, Skinner must modify that person's environment, i.e., the behavior of others—and so on, in infinite regress, because to shape the behavior of the others involves modifying their environments, i.e., the behaviors of yet others.

In a different way, groups may pose another “sociological” problem for Skinner. As he notes, he is a “methodological individualist” (AB, p. 241); further, as he has Frazier say, “We know almost nothing about the special capacities of the group” (W2, p. 293). Yet Skinner seems not to realize that the group, with its group structure and behavior, may have characteristics and behaviors (including “special capacities”) that are not merely the sum-total of the characteristics and behaviors of its individual members and that are not readily—if at all—susceptible to analysis and control by a methodological individualist, who “reduces social processes to the behavior of individuals” (AB, p. 241). In slighting the group, with its special capacities, structure, and behavior, Skinner seems to be slighting a pervasive form of social organization and ignoring an important aspect of and influence on human behavior. (This point is more fully developed by Watts, “B. F. Skinner and the Technological Control of Social Behavior.”)

8 As the example of ethical training in Walden Two makes clear, Skinner's sincerity in advocating “planned diversity” is in doubt. Diversity drastically increases behaviorism's technical problems. On the other hand, if Skinner really opposes diversity, then the survival of a Skinnerian society is rendered very problematic. In short, Skinner is on the horns of a dilemma: diversity is needed for survival but technically almost impossible; uniformity is technically desirable but makes survival improbable.

9 Skinner generally overlooks costs and the problem of cost constraints; he writes as though he and his engineers had unlimited resources. (It should be noted that, in this instance, he could minimize the Problem of costs if behaviorism could be successfully introduced through the establishment and spreading of communities, like Walden Two.)

10 “Subjective” and “perceptions,” among other terms, are strictly speaking not terms in the behaviorist's vocabulary. Those who are not behaviorists can easily understand the terms' meanings; and behaviorists can easily “translate” them into technical language (AB, pp. 19–20).

11 Skinner might try to point to “deferred aversive consequences” as grounds for not using heroin. But then he both must prove that the deferred consequences are indeed aversive and must convince the heroin users that the consequences will be, on balance, aversive. Even were this possible, the aversive consequences would be only weak current reinforcers, since the aversive consequences are in the fairly distant future (more distant than the next, positively reinforcing “fix”), i.e., the aversive consequences are “remote” (BFD, p. 134). Further, even “unremote” aversive consequences are not effective against some reinforcement schedules (BFD, pp. 34–35; AB, p. 71).

12 The implications of Skinner's methodology are far-reaching. He really proposes an end to knowledge or, rather, that there is no knowledge—for scientists pursue not science, knowledge, and facts, but instead that which reinforces them, just as Skinner's engineers pursue not the objective interest of the community but instead that which reinforces them. Among other things, there is something strange about a “scientific” methodology that ends up denying the existence of scientific knowledge. At most, the validity of such a methodology might be thereby thrown open to doubt.

13 It is hard to know exactly what would constitute “proof” here. Skinner admits that he does not yet have enough evidence (BFD, p. 160; AB, p. 189). Both this essay and Scribner's “Escape from Freedom and Dignity” try to demonstrate logically that Skinner's system will not work as he projects.

14 Nuclear weapons and ecological problems may seem to be partial exceptions, where the decentralized has little or no advantage over the centralized. To a large extent, however, the current nuclear and ecological problems have been brought about by centralized planning and by concentration (i.e., centralizing tendencies) in industry.

15 As Skinner himself sees, some cultures which he does not like have survived a long time: “If you control through superstition and ignorance, as has been the case in India, the control may be profound (a maharaja and his descendants may prevail for centuries)” (Evans, , B. F. Skinner, p. 49). If survival is the criterion, surely some Indian cultures must be rated highly, as must be the hunter-gatherer culture of primitive man. Skinner does add that “in the long run you have to consider the survival of the group” (loc. cit.); but surely “centuries” (in India) and millennia (for hunter-gatherers) are an adequately “long run.”

16 Rogers, Carl R. and Skinner, B. F., “Some Issues Concerning the Control of Human Behavior: A Symposium,” Science, 124 (30 Nov. 1956), 1065. Obviously, survival was not an unconditioned action for Socrates or Thomas More.

17 As he himself admits. In answer to the question, “Why should I be concerned about the survival of my culture?” Skinner writes, “The only honest answer to that kind of question seems to be this: “There is no good reason why you should be concerned, but if your culture has not convinced you that there is, so much the worse for your culture'” (BFD, p. 137). Skinner, thus, is in the strange position of using science to create and propound myth. See also note 12, above.

18 Skinner could actually come up with a less weak defense of survival as a value; following Hobbes, he could argue that survival is an instrumental value, necessary because it is a means to all other values. But then Skinner would have to start from and emphasize individual, not cultural, survival. And survival is not as logically sound an instrumental goal as, for instance, Hobbes's peace, as Skinner implicitly realizes (BFD, p. 210); it might also commit Skinner to favoring a hunter-gatherer culture (see note 15, above). Even such a stronger argument, however, would not avail completely. How would Skinner explain those who would “rather be dead than Red,” who would rather smoke cigarettes and risk cancer than give them up, or who gamble or shoot heroin to their own destruction? In short, individuals and cultures do differ in what reinforces them, including whether or not survival is reinforcing. In any case, for a Skinnerian society to have a single goal (instrumental or final) poses the problem raised by the next paragraph in the text.

19 In About Behaviorism, Skinner says that “we must look to the culture,” not to individuals (AB, p. 206). But looking to the culture is inadequate, on Skinner's own terms, because, in the absence of wise behavioral design, “the results are by no means well-designed social environments” (AB, p. 191). In order to get a “well-designed social environment,” “explicit design” is necessary (AB, p. 205), and it requires human action and intervention: “In the behavioristic view, man can now control his own destiny because he knows what must be done and how to do it” (AB, p. 251). In short, despite Skinner's appeal to culture (AB, p. 206), the question remains: “Who will implement Skinnerian ideals?”

20 In response to some of the criticism of Beyond Freedom and Dignity, Skinner admitted that “I do not think that I quite understood the limitations of control and countercontrol.” Center Magazine (Mar./Apr. 1972, p. 65). Perhaps as a result, his formulation of control and countercontrol in his recent work stresses heavily the “social environment” (i.e., the culture) as the controlling factor (AB, esp. p. 206). But Skinner's reliance on the culture is inadequate again (see note 19, above). If the culture controls completely, then individuals have no causal role, and Skinner's appeals to his readers (BFD, p. 215; AB, p. 251) are meaningless and superfluous. If, on the other hand, individuals do have some causal role and can change their culture, as Skinner suggests when he says that “man can now control his own destiny” (AB, p. 251), then Skinner must look beyond the changeable culture to other things if he is to find any effective countercontrol.

21 If “countercontrol” is seen in a different interpretation to avoid these problems, i.e., if countercontrol is merely the requirement that the masses be kept happy by the controllers, then Skinner has failed to solve the problem of the “happy slave,” which he taxes the “literature of freedom” for failing to solve (BFD, p. 40).

22 Skinner's unanalyzed use of the word “environment” raises problems (see note 7, above); similarly, his unanalyzed use of “control” can be seen as raising the problems here. If “control” means manipulation and total control, then the controllers totally shape the controllees, who can and do have no countercontrol (except, of course, for trivial limits, like their physique and genetic make-up), because any countercontrol would limit the controllers' ability to manipulate and control totally. On the other hand, if ’control” means merely influence, then the controllers can exercise some countercontrol, but then also the controllers can only partially shape the controllees. In brief, if control means manipulation, there is total shaping and no countercontrol; if control means influence, there is only limited shaping and effective countercontrol. If Skinner's controllers are to be able to bring about all the benefits he promises, however, they must have control in the sense of manipulation, and then there is no countercontrol.

23 Evans, , B. F. Skinner, p. 86, and also Skinner's remarks on pp. 45–46, 49, and 104; see also AB, pp. 176–180.

24 Psychology Today, Nov. 1972, p. 67.

25 Houriet, Robert, Getting Back Together (New York: Avon, 1972), pp. 320321. For a theoretical defense of decentralization and self-management within a Skinnerian world, see Hedman, Carl G., “An Anarchist Reply to Skinner on ‘Weak’ Methods of Control,” Inquiry, 17 (Spring 1974), 105112.

26 Possible Skinnerian objections to self-management as a program are invalid. One objection might be that not everyone has the technical expertise necessary (W2, p. 266); but the dissemination of such technical knowledge is probably easier, cheaper, and less risky than the alternative of establishing the few Skinnerian planners with full control. (For the risks of Skinnerian behaviorist planners, see W2, pp. 256–257; BFD, pp. 34–35; AB, p. 198; and Evans, , B. F. Skinner, pp. 5355, and esp. p. 54, where Skinner, talking of the misuse of behavioral techniques [some-times in primitive forms], says, “I think a science of behavior is just as dangerous as the atom bomb. It has the potential of being horribly misused.”) Another objection might concentrate on the results, and argue that if the society is “effective,” then self-management may well work; but if the society is “ineffective,” then self-control will produce little benefit (SHB, p. 439). The same generalizations, however, apply to the planners in the centralized planning model, since they too are subject to environmental influences.

27 Mill, John Stuart, On Liberty (New York: Library of Liberal Arts, 1956), Chap. 1, p. 3; Hobbes, Thomas, Leviathan, ed. Peters, Richard S. (New York: Collier Books, 1962), Chap. 21, pp. 159160; Locke, John, An Essay Concerning Human Understanding (New York: Dover, 1959), Book II, Chaps. 21 and 27.

This review essay owes much to others, especially faculty and students at Vassar College, where many of the points were first presented in a class. An earlier version was delivered as a lecture at Philosophers' Holiday, and copies of the lecture were freely distributed. As a result of the many comments and criticisms that ensued, arguments were modified and errors removed. I would like to thank the many at Vassar who were helpful, and especially Jesse Kalin, Michael Murray, Suzanne Vromen, Nancy Downing, Ed Jacobs, and Kim Landsman. In addition, I would like to thank Emily Gill and P. Gordon B. Stillman, who read and commented on this essay at various stages. None of them, of course, is responsible for the problems and infelicities that remain.


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed