Excerpted and rewritten from Chapter 18, Understanding Schools
Gary K. Clabaugh & Edward G. Rozycki, (New York: Harper & Rowe 1990)

Is Virtue a Learning Outcome? Can Virtue Be Taught?
©1999 Edward G. Rozycki

Can you tell me, Socrates, whether virtue is acquired by teaching or by practice, or if neither by teaching nor practice, then whether it comes to man by nature, or in what other way?  
-- Plato: Meno

edited 8/17/11

The debate as to whether morality can be taught is an ancient one. Today, educators are becoming more and more involved with explicit instruction in "values." By observing some important distinctions, we can shed some light on the question. More importantly, we will see that many kinds of learning are similar to what we might call moral learnings and can be dealt with in a similar fashion.

Let's begin by narrowing down the question, "Can morality be taught?" This is an ambiguous question. It could mean

a. Can a someone, e.g. a teacher or parent, be held responsible for the the formation of character in a child? or,

b. Can there be a scientific process (a summatively specified cause) which a teacher could set in motion which would result in a student's becoming honest?"

The answer to a. is a clear yes. We often hold people responsible for things which they have little causal control over, in the scientific sense. One informal meaning of cause is a legally or morally accountable person, as  when a policeman asks, "Who caused this accident?" knowing that the driver of the wrecked car had been distracted.

To the second question, we will argue no. The reasons for this will be given shortly by way of analogy rather than argument. The question of whether morality can be taught can be reformulated as "What (or how) does one teach to cause (i.e."efficiently cause" in some summative sense) a person to become an honest person?" This question is very much like the question, "How much power do you need to become an authority?" What such a question indicates is a failure to understand that the relationship of power to authority is not a causal one, just as the saying, "Might does not make Right" indicates.

It is important to carefully make out the distinction between power and authority. Someone's power depends upon his or her control of resources; authority depends upon some other people's consenting, for good reasons, to recognize it as such. Power is a summative concept: conditions can be specified that eventually reach sufficiency.  But authority is a matter of consent. If a person has enough power, we may feel coerced to concede to his or her demands. But, as we have seen, this power of compulsion is not authority. Only when we defer freely, for the proper reasons, to the judgment of another is this the recognition of authority.

In a parallel fashion we might argue that certain kinds of training and behavior give us reason to bestow moral titles, as it were, "honest". "trustworthy", etc. It may be reasonable or traditional or even moral to show people such respect. But it is not necessary. For example, Racism denies titles of respect to people on the basis of their skin color, even though they may show all the qualities that other people have that Racists respect. Racists are making a moral error, however; not a logical one.

Character Development vs.  Skill Aquisition

Let's contrast learning as character development with learning skills. We normally evaluate skills incrementally; character tends to be an all or nothing thing. We can make judgments of more or less skill, say, at basketball, by seeing how accurately a player shoots, whether he or she defends well or works well in the offense. But judgments of honesty are a different thing. A person who steals regularly once a month does not merely have less character than someone who doesn't steal at all, or more than someone who steals twice a month. He or she is dishonest, of "no character" or "bad character." Positive evaluations of character, once disconfirmed, tend to be lost -- unless "forgiven" as "out of character." A clever thief may do acts characteristic of honest people to gain our confidence. These do not suffice to make him honest. Rehabilitation is not merely a matter of doing good and refraining from evil. There is another aspect to be considered.

Once lost, character can only be reestablished within certain traditions of "redemption," (or, as we might put it more comfortably to our secularized sensibilities, "rehabilitation.")  One purpose of traditional religious rituals of atonement is to reestablish an individual's status as trustworthy. (To digress: we might wonder how a society can handle problems of character when there is no longer any consensus on how individuals atone for their misdeeds.) This is the point of the saying, "Forgive and forget!" Moral status is granted as a whole piece, so to speak, rather than incrementally. No specific tests of honesty can establish a person's honesty. That person, may, after all, be trying to deceive us. But specific individual acts of dishonesty may, but need not -- if done for an understandable reason -- invalidate that ascription.

This is an important point. Certain characterizations of learning are not so much summative descriptions of a learner as ascriptions, acts of acknowledgment on our part that a learner is entitled to be recognized in a certain way. These ascriptions are conceded to someone for having participated in a socially valued (or despised) process, or because a recognized authority has bestowed them. (For more on this see the Technical Appendix to Chapter 18, Item A.)

What then can "moral education" amount to? Moral education depends upon a consensus that children (people) who have been trained in a particular way, or have been exposed to example and preachment of a certain kind are worthy of trust, i.e. they should be recognized and dealt with as morally upstanding persons. (Such recognition, it should be noted, is not the same as a prediction that morally upstanding persons will not commit offenses, although we normally only concede such recognition on the expectation that offenses will not, at least, be actively pursued.)

Summative and Ascriptive Learning -- expanding on the distinction

Ascriptions tend to be defeasible, that is, subject to "defeat" (withdrawal) if knowledge of certain invalidating conditions is found. (See H.L.A. Hart on "defeasible concepts" in "The Ascription of Responsibility and Rights.") It is important to realize that much that occurs in schooling is a matter of social ascription, rather than a summative description of achievement.

For example a student may receive a Ph.D. from a university. This degree is a recognition rather than an achievement, even though it may be based on achievements. The diploma might be rendered worthless or revoked if any of the following conditions are later discovered:

a) the student cheated on his comprehensives;

b) his major professors had falsified their credentials and were not eligible to examine him;

c) the college had lost its accreditation;

d) a mistake had been made in computing his credit hours.

These are conditions that are not normally reviewed for each and every student. That is a good reason for introducing the term defeasible to identify such conditions rather than merely generalizing that all descriptions are conditional upon satisfying certain criteria. Defeasible conditions are normally assumed rather than investigated and judgmentally critical. They are a matter of social convention rather than summative procedure. And they play a large part in evaluating schooling outcomes.

The naive notion that degrees are outcomes of learning parallels the naive notion that virtue can be an efficiently causal outcome of teaching. That degrees are not revoked as , through the years, their possessors lose the knowlege the acquired on the road to obtaining the degree, demonstrates that degrees, unlike CPR certificates, are not contingent on the continued possession of certain knowledge.

But even descriptive judgments require a decision to bestow them. And one must be recognized as having the authority to make such a bestowal. How many quadratic equations must someone do before the teacher will certify him or her as knowing how to do quadratics? Two? A thousand? (Cf Ludwig Wittgenstein, Philosophische Untersuchungen, (Frankfurt-am-Main: Suhrkamp, 1967) Section 151, 152)

Consider what might happen if we discovered that the person who judged him competent were a music teacher, or his brother. This single point of evidence might defeat the characterization of her or him as competent. If we were concerned we would have to correct and repeat the testing process. But why should a test be relevant?

For example, we might agree among ourselves to have a computer designate the student as "competent in quadratics" if the student interacted with the computer by solving problems of different sorts. We could decide in advance of all such testing that such outcomes indicated competence. But this is still our decision! And the computer has no authority over and above those who programmed it to render that decision in specific cases! These considerations indicate the considerable complexity underlying judgments of competence. Let's see if we can illustrate it simply.

Authority and Defeasibility

There are two critical factors to consider: authority and defeasibility. Let's first distinguish between authorities who use public criteria of judgment as opposed to private criteria. Public criteria are those which can be articulated and presented for anyone's inspection, whether or not they understand or are interested in the judgmental process to be undertaken. Private criteria are those a person conceded authority uses in making a judgment, even if he or she is unable to articulate them for public presentation.

A characterization is defeasible to the extent that very specific defeating conditions are normally assumed to be absent. The greater the number of such conditions, the higher the defeasibility. Consider, for example, the following judgments made after some reasonable procedures, e.g. measuring with a tape measure, giving a math test and inquiring of the neighbors:

a) John is five feet tall.

b) John can do quadratic equations.

c) John is honest.

What could defeat these judgments? A variety of things, but they would have to be more bizarre for a) than for b) and c) . For example, the tape measure was misprinted; it was Harry, mistaken for John; John was wearing lifts in his heels, etc. Much more pedestrian are the defeating conditions for b),e.g. John cheated; the problems were too easy; he has forgotten how; he took the test on drugs, etc. c) might be defeated if John stole something, told a lie, or a committed a hundred other pecadillos.

There are various conceptions of cause used to explain the connections people see between things. "Efficient cause" is -- which has become the central notion of cause used in the sciences -- is just one of these. It is disputable whether other traditional notions of cause, e. g. persons as causes, attitudes as causes, are illumined by it.

Now, people tend to believe that the relationship between teaching and learning is causal. Again, it is not clear that they believe that the relationship between teaching and learning is causal in the technical, "efficient cause" sense. This technical sense of cause is summative. The reason for identifying defeasible learnings is that they cannot be summative, because the conditions which bring us to concede or bestow the judgment of learning hold only so long as they are not overridden. Summative judgments cannot be overridden. What does this mean?

A startling conclusion we can draw is that some kinds of learning are not caused in any scientific sense of cause. Should we give up, then, on trying to develop a scientific pedagogy? No. What we might give up on is accepting traditional defeasible ascriptions of learning as anything more rough approximations of what we are looking for. If we really want to do a summative causal analysis, a new language of learning outcomes may have to be developed. This is a task which remains to be seriously undertaken.

We should note here that the less explicit the procedures for rendering the judgments of learning, the more defeasible such judgments likely are. Certainly, there will be more consensus on what it means to be five feet tall, than on what it means to be able to do quadratics or to be honest. What defeasibility seems to point out is that while we may not easily agree on what something is, we tend to agree on what it is not. Perhaps this is why criticism is more likely than praise in the political environment of the school.

Figure TA 18.1 schematizes a range of learning outcomes as they depend on authority and defeasibility.

Figure TA 18.1

In figure TA 18.1 both "John is five feet tall" and "John is honest" are in parentheses because it is controversial that they are learnings. We might include them as learnings if our conceptions of physical and developmental possibility were to change. They are included on the chart merely to indicate the endpoints of the range of items given.

Note an interesting point illustrated in figure TA 18.1. To claim that John speaks some Spanish is a different kind of claim than to say unconditionally he can speak Spanish. A few, finite tests can determine that he speaks some Spanish. What justifies the claim that he speaks Spanish, not just some Spanish? One way such decisions are made is to have authorities (experts) recognized as being able to speak Spanish make the decision. We can see that unless we are willing to acknowledge some people as authorities on Spanish, the argument leads infinitely backwards. (For example, "On what basis do we determine someone is an authority?" This is like asking how much power it takes to give authority.) The general point to be made of all of this is that important judgments about learning depend upon established social practices of recognizing some people to be authorities.

If the argument of this section is correct then rehabilitation is not merely a matter of learning to refrain from dishonesty, but of acceptance into a community willing to bestow trust.