Site hosted by Build your free website today!

Rational Action and Education

** NOTE: The plain-text version of this article, below, has been corrected by the author to remove typographical errors generated by an OCR (optical character reader). Page breaks have been indicated in [square brackets]. Footnotes originally indicated by half-sized superscripts have been entered in [square brackets]. Also, near the top of page 166 there are formulas where the letters X,Y,Z have subscripts of either 1 or n, and those subscripts have been indicated with [square brackets].


Kenneth R. Conklin, "'Rational Action' and Education," paper read to annual national convention of the Philosophy of Education Society, Boston, April, 1974. Published in PHILOSOPHY OF EDUCATION 1974: PROCEEDINGS OF THE THIRTIETH ANNUAL MEETING OF THE PHILOSOPHY OF EDUCATION SOCIETY, ed. Michael Parsons. Edwardsville, Ill.: Studies in Philosophy and Education, 1974, pp. 165-174.



(Second Concurrent Session)

Boston University

An action cannot be rational unless it is done deliberately to achieve some purpose. Once a purpose is adopted, an action is not rational unless it is likely to help achieve the purpose under the prevailing circumstances as perceived by the agent. An agent may have faulty perceptions of circumstances. He may have foolish or evil purposes. But even so we customarily acknowledge that his action is rational if it helps him achieve his purposes under the circumstances as he perceives them.

Philosophers became interested in "rational action" over a decade ago. when they were preoccupied with linguistic analysis and foundations of scientific method. Therefore the clarification of the concept has emphasized the quantification of publicly observable actions and their probabilities rather than the rightness of private purposes. The result was the adoption of the ideal model: decision-making under risk as used in economics and general game theory. This model also seemed appropriate to the growing interest in applying the "systems approach" to the study of education. We shall see, however, that rationality depends upon the existence and quality of private purposes. A common-sense distinction among three types of rationality is much more important for studying education than the analysis of risks in the systems approach.

The Hempel Model and Its Inappropriateness

Carl Hempel worked out the decision-making-under-risk model of rational action.[1] Suppose there are a number of possible actions one might take, and that in any case not more than three outcomes are possible. Let us label the outcomes X, Y, and Z. Suppose that in view of our purposes we place value A upon X, value B upon Y, and value C upon Z. These values are numbers indicating the worth of the outcomes, and are positive if the outcome would be considered good or negative if bad. Suppose we

[End page 165 / Begin page 166]

believe that if we perform action number 1, the probability of outcome X is X[1], the probability of Y is Y[1], and of Z is Z[1]. Suppose that, in general, we believe that if we perform action n the respective probabilities are X[n], Y[n], and Z[n]. Then for any particular action n the value we can expect for the outcome of the action will be AX[n]+BY[n]+CZ[n]. The action most likely to achieve our purposes will therefore be the action n that maximizes the mathematical number represented by AX[n]+BY[n]+CZ[n]; hence, this is the action that would be called "rational."

One problem with this definition of "rational action" is that it does not take account of how conservative or how speculative we may be. For each possible action we may compute not only its expected value but also its greatest possible gain if luck is with us and its worst possible loss if luck is against us. A conservative criterion for rational decision might be to choose the action which has the smallest size of loss in the event of bad luck. A more daring criterion might be to choose the action with greatest minimum gain in the event of good luck. An exceedingly speculative criterion would be to choose the action which maximizes the greatest possible gain, all thought of losses aside. The Hempel model does not tell us whether it is rational to be speculative, or what goals are rational. But once we choose our goals and our degree of risk, then the Hempel model tells us which action would be most rational in that context.

Given various proposals for educational action, and the possible outcomes for each action, the most immediate practical difficulty with the Hempel model is to establish the probabilities for the outcomes. In education we are dealing with individual children who are all different, so the probabilities for the various outcomes of a given action will be different from child to child. Still, this problem may not be insurmountable, and is actually present in all scientific fields since no individual item is completely standard. A more serious difficulty is the effect of observation and experiment upon the subject's responses (e.g., Hawthorn effect); yet, a similar difficulty is present in quantum mechanics. Another difficulty is determining what all the possible outcomes of an educational action are; yet, the problem of "concomitant learnings" in education may be only somewhat more complicated than the problem of "unintended consequences" or "side effects" in the sciences generally. Another problem is to quantify the values of possible outcomes, but other sciences have this problem too. All the problems of the Hempel

[End page 166 / Begin page 167]

model in the sciences are more severe in education, but so far it appears the problems are merely different in degree but not in kind.

Yet there is one way the Hempel model fails in education that is fundamentally different from its difficulties in all the sciences. Educational purposes are more or less general in scope, and purposes at one level may not be comparable with purposes at another level.[2] A specific observable outcome may be construed as contributing to various purposes at various levels of generality: e.g., course outcomes, school outcomes, life outcomes, each regarded with respect to personal life-style of a student, the general welfare of society, and the degree to which one experience facilitates openness to another. There is simply no way to quantify and reconcile all these incommensurable values. Even if some mathematical technique could be developed to do this, it would be too complex to be used in practice.

The Hempel model is useful because it calls attention to the fact that both fact and value are involved in rationality: and action is called "rational" only if it is justifiable on the basis of the factual assumptions and values of the agent. But the model is too narrow in requiring an agent to choose an action which maximizes the mathematically expected value of the outcome -- surely an agent could make mathematical or logical errors, or be unclear on the precise values he attaches to outcomes, and still act rationally. The Hempel model is also too broad, since it allows agent to hold any values and any factual assumptions whatever, including hallucinations and superstitions. Indeed, in criticizing the model we feel torn between insisting that agent should have "good" reasons, "good" logic, etc., vs. insisting only that he must have some kind of factual assumptions, logic, and values which seem reasonable to him and which are more or less consciously held at the time of action.

Subjective, Conventional, and Real Rationality

Perhaps the most elementary kind of rationality is simply having reasons which seem correct and seem to justify action. This would be subjective rationality: action is rational when one believes it to be consistent with one's own assumptions about fact and value. Even with such a permissive definition we would still find some actions labelled irrational; e.g., cases in which the agent

[End page 167 / Begin page 168]

says truthfully of his own action that he did it blindly, without motive, or in a fit of passion. But subjective rationality is too permissive, since it allows actions to be called rational merely by adducing hallucinations or superstitions or schizophrenic logic in their support.

A more sophisticated kind of rationality is conventional rationality: agent has reasons (values and factual beliefs) which are generally acknowledged to be correct and generally acknowledged to justify the action. It is the general social acceptance of agent's factual assumptions, values, and mode of reasoning that distinguishes conventional rationality from subjective rationality. Philosophers might well argue about the relative importances of the three elements of fact, value, and logic: some might grant the status of conventional rationality to action which possesses general acceptance of any one of the three elements; other philosophers might require a particular one of the three, two of the three, or even all three. There could be many different interpretations of conventional rationality.

The most stringent kind of rationality would be real rationality: agent justifies his action on the basis of his knowledge of true factual conditions, true values, and really correct logic. Most philosophers would deny that there is such a thing as real rationality, or would claim that only an omniscient God could know what is really rational. Nevertheless, people do occasionally claim that their subjective rationality is superior to conventional rationality because it accords more nearly to real rationality, and we shall see how this claim is important in education.

The Hempel model could be used as a specialized, narrow version of any of the three types of rationality identified here. Hempel himself was so concerned with the quantification of rational decision that he explicitly required only what we are calling subjective rationality; however, conventional rationality or real rationality could also be adopted for use in the Hempel model. The point is that Hempel dealt with quantification of choice, while the three types of rationality distinguished here are concerned with the orgin and status of the facts, values, and logic used as reasons to justify action.

There are some philosophical issues cutting across everything said so far. We have been assuming that an agent must have reasons consciously in mind at the time of action if action is to be called "rational." But some philosophers might want to avoid reference to mental states and define rationality

[End page 168 / Begin page 169]

strictly in terms of observable characteristics of action. Indeed, Hempel's model was an effort to do this. Although I have been referring all along to reasons in the mind of an agent which motivate him to act as he does, Hempel avoids reference to agent's mental state.

Hempel says an action is rational if post-facto analysis convinces us that the action was likely to maximize expected gain on the basis of information we believe was available to agent at the time of action. But Hempel seemingly does not require agent actually to be aware of the information, nor to make any risk analysis before acting, nor even to have any thoughts at all. Hempel is really talking about the criteria whereby observers judge whether to call another person's action rational. Hempel is trying to define the conditions which make conventionally rational our action of calling another person's action rational. Viewed in this way, Hempel's model has nothing to do with education at all, nor with any first-order action: the model only helps us engage more rationally in the meta-action of talking about education or other first-order action.

Of course Hempel's model could be made more useful by construing it as I have done in this paper. Now we talk about agent's action, and we say it is (at least subjectively) rational if he had values, perceptions of the factual situation, and reasoning processes which would justify the action. Our problem is that we may make errors in ascribing values, factual perceptions, or reasoning processes to agent. But this does not affect whether agent's action is rational -- it merely affects whether we judge it correctly. The problem of distinguishing between rational action and identical but unthoughtful or hazily-thought-out action is precisely comparable to distinguishing between knowledge and right opinion. We feel that the essence of knowledge is having conscious awareness of the rightness of opinion, yet we do not know how to specify the publicly observable characteristics that guarantee the existence of "conscious awareness of the rightness of opinion." Similarly, the essence of rational action is having reasons in mind at the time of action which seem to justify the action. The fact that we cannot specify publicly observable characteristics which unerringly inform us about agent's mental state (except according to the subjective rationality of the analytic or positivist viewpoint).

In order for mistaken action to be rational, must we show that it was reasonable for agent to make his mistakes? That would

[End page 169 / Begin page 170]

lead to an infinite regress. We avoid the regress by our distinction among three types of rationality. Action is at least subjectively rational if it was purposeful and if agent's use of his own version of Hempel's model would provide a justification for agent's action in terms of agent's values, perceptions of fact, and logic. Action is conventionally rational if it is subjectively rational and in addition the factual perceptions, values, and logic are characteristic of agent's peer group. Action is really rational if it is subjectively rational and in addition the factual perceptions, values, and logic are really right.

Rationality in Education

As used in ordinary parlance, "rational" is often a synonym for "acceptable," "nice," "proper," "prudent," or "careful." It is an antonym for "reckless" or "absurd." To say that an action is rational is often to say merely that we approve it. When we approve an action we are inclined to label it "rational" without further inquiry into whether the agent has reasons in mind. That question typically arises only when we disapprove his action and want to know whether he was behaving "rationally" despite his "bad" deed. When we call a "bad" action "rational" it is our approval of the reasons or of the logic of justification that is conveyed by the label "rational." The statement "People should behave rationally" therefore translates into the statement "People should behave in ways that I approve," or, "People should function in accord with my version of subjective rationality." To say that an action is rational is to make what C. L. Stevenson would call a "persuasive definition." [3]

This characterization of "rational action" is quite prevalent in our schools. So long as a student does not offend teacher's sensibilties, student's behavior will probably be deemed rational without further inquiry. This becomes dangerous in the case of students who are quiet or timid or even psychotically withdrawn, since the absence of offensive behavior may be construed as prima facie evidence of rationality and the teacher may not detect a child's problems. When students do violate a teacher's sensibilities, their actions will be scrutinized to determine motives. Minor aberrations will be tolerated so long as the student's underlying set of reasons and principles of reasoning obtain teacher's approval. But strongly conflicting reasons and logic

[End page 170 / Begin page 171]

when manifested persistently will cause a student to be labelled "incorrigible" or "improperly socialized," and he will be subjected to powerful "disciplinary" techniques (e.g., behavior modification) designed to bring his fundamental assumptions and modes of reasoning to the point where the teacher will approve or at least tolerate them. And by his own standards virtually everything the teacher does to the student is rational, since the teacher will find reasons meeting his approval justifying what he does.

The concept of subjective rationality seems so permissive as to be useless; yet, education is precisely the formation of subjective rationality. A child's self-concept is the set of beliefs, values, and patterns of reasoning which serve to justify actions leading to consequences pleasant to him. Festinger's theory of cognitive dissonance shows that people feel motivated to adopt reasons and actions which are mutually consistent. Education can make use of the motivating power of cognitive dissonance by setting up situations that will promote the development of a child's self-concept in such a way that his subjective rationality will produce actions deemed good by those in authority.

To say that education transmits the culture is to say that education socializes the child -- it bends the child's subjective rationality into conformity with conventional rationality. The child comes to adopt as his own the reasons and modes of reasoning considered correct by society. Even a student studying for a doctorate in some highly specialized field of knowledge is primarily seeking to master the beliefs and folkways of reasoning that are commonly accepted by the cognoscenti in that field.[4]

There is a great difference between saying that the task of education should be to make a child conventionally rational and saying that the task should be to teach a child the skill of behaving in accord with conventional rationality. In the former case we actually bend the child's subjective rationality until it conforms with conventional rationality, so that the child spontaneously behaves in socially accepted ways. In the latter case we do not tamper with his subjective rationality; rather, we teach him what conventional rationality demands and how to behave in ways that conform to it, and then we leave it to the child's subjective rationality to choose whether and when to behave with conventional rationality for his own private reasons.

It seems that education almost always tries to make a student conventionally rational, although this does violence to his

[End page 171 / Begin page 172]

right as an individual to possess a unique subjective rationality. If education would try only to teach students the knowledge and skill necessary for them to give the appearance of conventional rationality, there would be no infringement of personal rights. For example, telling students that Standard American English is the only correct usage, and then trying to make them adopt it in private as well as public activity, is an infringement. But teaching students how to use Standard American English as a tool for getting jobs and communicating in middle class society is not an infringement so long as we make it clear that students are free to decide for their own private reasons whether and when to use it.

To the extent we believe it important to protect the right to a subjective rationality which may differ significantly from conventional rationality, to that extent we evidently believe there is a higher kind of "real rationality." If there is such a thing as "real rationality," and if conventional rationality falls short of it, then any person whose subjective rationality conforms on some occasion with real rationality would be doing the right thing despite its being judged as conventionally irrational. Civil disobedience is a case in point. Someone may deliberately violate statutory law, claiming that such a violation is necessary on account of a higher-order (e.g., "natural" or "God-given") law. He claims that his subjective perception of his act is superior to the conventional perception of it because he is in tune with a more real perception of truth than most other members of society.

Education for allegedly real rationality is the goal of every group that seeks to evangelize or alter the consciousness of outsiders. Religious groups sometimes claim to possess metaphysical truths that support conventionally irrational action or oppose conventionally rational action. By teaching religious doctrine they hope to enlist new adherents who will foresake the evils of conventional rationality and adopt righteous behaviors which, although conventionally irrational, are justified by revealed metaphysical truth.

But the attempt to evangelize or alter people's fundamental assumptions is not limited to religious education. Science education has the same mission. Religion deals with metaphysical truth and science deals with empirical truth, but both are engaged in educational movements seeking to get people to become aware of non-conventional truths and to adopt allegedly "real rationality" based on these truths. The whole point of science education in the general curriculum for the average citizen is to

[End page 172 / Begin page 173]

make the citizen aware of particular truths, and procedures for arriving at truth, that will cause the citizen to change his perceptions of rational action. The hoped-for social result of science education for the masses is to alter conventional rationality to make it conform more nearly with what scientists perceive as real rationality. Because of their training, specialists in science do unusual actions which they regard as really rational but which violate convetnional rationality; e.g., cutting up dead bodies, working in the presence of smelly or dangerous substances. In this regard the scientist is like a priest, who also does conventionally irrational things that his training convinces him are really rational; e.g., praying, or lifelong abstention from sex.

Sometimes education for real rationality is based on the attempt to make students recognize and act on pre-selected truth and procedure, as in the cases of religious education and science education. But sometimes education for real rationality is an attempt to bring individuals or even a society to the point where they will discover for themselves new truths which the planners of the education did not directly anticipate. Here we have education for self-realization, and also education for democratic social reconstruction. For example, Maslow identifies kinds of peak experiences while leaving unspecified the exact contents of those rxperiences; and many educators have expressed interest in helping students achieve such peak experiences. Piaget identifies stages of cognitive ability while leaving unspecified the exact contents of the understandings that will be achieved by the application of such ability. Raup, Axtelle, Benne, and Smith in The Improvement of Practical Intelligence,[5] and William O. Stanley in Education and Social Integration,[6] identify one task of education as the promotion of social reconstruction by democratic means, leaving unspecified exactly what substantive laws, attitudes, and beliefs will be adopted as conventional rationality in the newly emerging society.


There are many interpretations of the concept "rational action." All of them have some application to education, although the most technical interpretations (e.g., the Hempel model) are least applicable. Of course we should try to behave rationally. Of course schools should operate rationally and should teach children

[End page 173 / Begin page 174]

how to behave rationally. When we say such things in general conversation we probably mean to endorse simultaneously all the behavior acknowledged as rational according to all the interpretations of the term; on special occasions, however, when confronting a problematic situation, we interpret the term in a particular way in view of the demands of the situation.

Education should help a student clarify his own values, factual beliefs, and thinking processes so that his action will be at least subjectively rational. This conclusion is in keeping with the notion that to be human is to have a personal identity and to live with integrity. It is also in keeping with the slogan, "The unexamined life is not worth living." Education should help people understand the values, facts, and modes of reasoning generally accepted by society, together with the skills needed to behave artfully in accord with them. This conclusion is in keeping with the notion that education transmits the culture. But we must not force students to adopt conventional rationality in lieu of subjective rationality. Rather, our task is to teach students what conventional rationality expects, so that their freedom to make intelligent personal choices with knowledge of social consequences is enhanced. "Forewarned is forearmed." Finally, education should give people the faith that real rationality exists, and the zeal to pursue it. The search for correct facts, righteous values, and really valid modes of reasoning is the search for transcendent self-realization. Faith in the existence of real rationality makes it at least conventionally rational to challenge conventional rationality.


[1] Carl G. Hempel, "Rational Action," Proceedings and Addresses of the American Philosophical Association, vol. 35 (Yellow Springs, Ohio: Antioch Press, 1962), pp. 5-23.

[2] Harry S. Broudy, "The Philosophical Foundations of Educational Objectives," Educational Theory 20, 1 (1970): 3-21. See also, Kenneth R. Conklin, "The Concomitance of Fact-Value, Emotion-Cognition, and Goal-Action," in Philosophy of Education 1971: Proceedings of the Twenty-Seventh Annual Meeting of the Philosophy of Education Society, ed. Robert D. Heslep (Edwardsville, 111.: Studies in Philosophy and Education, 1971), pp. 254-263.

[3] Charles L. Stevenson, Ethics and Language (New Haven: Yale University Press, 1944).

[4] Kenneth R. Conklin, "Due Process in Grading: Bias and Authority," School Review 81, 1 (1972): 85-95.

[5] R. Bruce Raup, George Axtelle, Kenneth Benne, B. Othanel Smith, The Improvement of Practical Intelligence (New York: Harper and Brothers, 1950).

[6] William 0. Stanley, Education and Social Integration (New York: Bureau of Publications, Teachers College, Columbia University, 1953).

[end of article]


Send comments or questions to:

You may now look for another scholarly article published by Ken Conklin before he came permanently to Hawai'i



(c) Copyright 2012 for this website, by Kenneth R. Conklin, Ph.D. All rights reserved