20th World Congress of Philosophy Logo

Theoretical Ethics

Instrumental Rationality and the Instrumental Doctrine

Alistair M. Macleod
Queen's University, Canada

bluered.gif (1041 bytes)

ABSTRACT: In opposition to the instrumental doctrine of rationality, I argue that the rationality of the end served by a strategy is a necessary condition of the rationality of the strategy itself: means to ends cannot be rational unless the ends are rational. First, I explore cases-involving ‘proximate’ ends (that is, ends whose achievement is instrumental to the pursuit of some more fundamental end) — where even instrumentalists must concede that the rationality of a strategy presupposes the rationality of the end it serves. Second, I draw attention to the counter-intuitive consequences — in cases involving ‘non-proximate’ ends — of substituting (allegedly more manageable) questions about de facto ends for questions about the rationality of ends. Third, I argue-against Nozick — that it is a mistake to suppose that the only question dividing instrumentalists from non-instrumentalists is whether the instrumental doctrine needs supplementation. Finally, I try to show that questions about the rationality of ends need not be viewed as impossibly daunting.

bluered.gif (1041 bytes)

According to the instrumental doctrine of rationality in the version relevant to the argument of this paper, an action (decision, policy, strategy, etc.) is rational provided it is an effective and economical means to the achievement of some de facto objective. If we formulate the instrumentalist position in terms of the familiar doctrine of the practical syllogism, the crucial thesis is that the action which forms the conclusion of the syllogism is rational provided (1) the major premise identifies a de facto objective of the agent's, and (2) the minor premise shows the action to be an effective and economical means to the achievement of that objective. The typical noninstrumentalist position, by contrast, would be that for the action in the conclusion to be one it is rational for the agent to perform, it must serve an objective it is rational for the agent to pursue: the major premise must identify a rational objective of some sort, not simply an objective the agent happens to have.

I. The Instrumental Doctrine and "Proximate" Ends

One way of denting the instrumentalist position is to explore cases where the action said to be rational is an effective and economical means of enabling the agent to achieve an end he or she is pursuing only because its achievement is (held to be) indispensable to effective pursuit of some more fundamental objective. These are cases where the agent is pursuing (what we might call) a "proximate" end, an end which is thought to be worth pursuing only because its achievement is a means to effective pursuit of a more basic end. "Proximate" ends are interesting because they seem not to be covered by the ban the instrumentalist otherwise imposes on the asking of questions about the rationality of ends. Even from an instrumentalist standpoint, it would seem, the question whether a "proximate" end is or is not "rational" can be coped with. It can be maintained, quite straightforwardly, that pursuit of a "proximate" end is rational provided its achievement would be instrumental to the achievement of the more basic end it presupposes.

The need for this emendation, it might be held, shows nothing more than that greater circumspectness must be exhibited in the formulation of the instrumental doctrine: it should not be thought to impose a comprehensive ban on the raising of questions about the rationality of ends. Otherwise the concession is innocuous, it might seem.

However, although the concession may appear innocuous, there is reason to think that it has potentially damaging implications for the instrumentalist position.

For one thing, it creates an opening for the question whether there is a connection between the rationality of a strategy and the rationality of the end it serves. Once this question is put, it is difficult for the instrumentalist to say, bluntly, that there is no connection whatsoever. It is very difficult, that is, to claim that the two rationality-judgments in question are logically independent of one another—that the question whether the strategy is rational has nothing whatsoever to do with the question whether pursuit of the presupposed end is rational. It would be paradoxical, for example, for a defender of the instrumental doctrine to have to say in these circumstances that the rationality of the means an agent adopts in pursuit of a given end is wholly consistent with its being quite irrational for the end to be so much as pursued in the first place. Yet, for the truth of what might be called the linkage thesis to be conceded in cases where an agent is pursuing a "proximate" end is tantamount to the concession that, at least in these cases, the rationality of an action (decision, strategy, etc.) is not a simple function of its effectiveness in bringing about an end: the rationality of the end itself is a further necessary condition.

Once it is admitted that there is something to the linkage thesis in at least some cases, it is difficult to dismiss the question why it should be thought to have no force at all in other cases. If it cuts no ice in cases where the end is a merely "proximate" end for the agent simply to affirm his or her commitment to

achievement of the end—in the hope that the affirmation will render it unnecessary for questions to be faced about whether it makes any sense for the end to be pursued in the first place—why should it be supposed that such a simple affirmation suffices in cases where the end happens not to be a merely "proximate" end? In the absence of a satisfying explanation of the difference between the ("proximate"-end) cases to which the linkage thesis is conceded to apply and the cases in which it allegedly doesn't hold, the presumption must surely be that it applies in all cases.

There is, in any case, a very general argument that supports this presumption, an argument calls attention to a feature practical syllogisms have in common with ordinary ("three-proposition") syllogisms. In an ordinary syllogism, there is a rational basis for affirming the proposition in the conclusion of the syllogism only if (among other things) there is a rational basis for affirming both the proposition in the Major Premise and the proposition in the Minor Premise. The constituents of practical syllogisms may not all be propositions—e.g.,because (on some accounts) the conclusion may be held to be the agent's action or decision, or because the Major Premise may be some such thing as the agent's commitment to a goal or principle or ideal. Nevertheless, the argumentative force of such syllogisms can only be preserved—i.e., that the conclusion can only be represented as rationally endorsable—if both the Major Premise and the Minor Premise can be represented as rationally endorsable. Just as it would be suspicious if the sponsor of an ordinary syllogism were to make light of the fact that he or she had no basis for supposing the proposition in the Major Premise to be rationally justifiable despite his or her determination to represent the conclusion as embodying a rationally justifiable proposition, so too we ought to be suspicious of instrumentalists who think they can affirm the rationality of the action specified in the conclusion of a practical syllogism despite their acknowledged inability to represent the commitments or goals specified in the Major Premise as rational commitments or rational goals. And just as it doesn't help, in the case of ordinary syllogisms, for a ringing declaration to be issued that the person endeavoring to draw a rationally justified conclusion actually believes the proposition in the Major Premise (while lacking any good reason for regarding it as a rationally justified proposition), so too it doesn't help for the instrumentalist to try to substitute a claim to the effect that the person whose conduct is being guided by a practical syllogism actually has the commitment or goal specified in the Major Premise for a claim about the rational defensibility of the commitment or goal.

II. Counter-intuitive Implications of the Instrumental Doctrine

There are at least three ways in which some of the markedly counter-intuitive implications of the instrumental doctrine can be brought out.

The first is to ransack the historical record for offbeat, not to say pathological, examples of the kinds of ends or goals individual agents have in fact tried to achieve, and then ask whether it is really credible that in these cases, no less than in cases where much more mundane (and benign) ends or goals have actually been pursued, the strategies which have been adopted for their achievement must all be denominated "rational" simply on the footing that they were effective and economical strategies. For example, should we really be prepared to say—as the instrumentalist, presumably, is obliged in the end to say—that the pain-inflicting strategies of the sadistic dentist who uses his professional expertise to torture his victims must be held to qualify as "rational" provided they are an efficient means of achieving his de facto ends? Or that the scheming behavior over time of the serial killer is "rational" provided it enables him to kill people without being caught?

A second way of showing how counter-intuitive the instrumental doctrine is to take notice of the fact that it requires us to characterize as rational not only what we do when we adopt efficient strategies for the achievement of our ends or goals but also what we do when we hit upon efficient means of applying the rules, principles, ideals, etc. to which we happen to be committed. In face of the great variety we find in the rules, principles and ideals to which people are ascertainably committed—and the commitments have their source, of course, in a wide range of beliefs (sometimes defensible, sometimes indefensible) about law, morality, politics, religion, etc.—we must ask, once again, whether it is really credible that a strategy which successfully implements any of these rules, principles, or ideals, should be denominated "rational," no matter what its content happens to be. For example, should we really be prepared to say, with the instrumentalist, that the power-preserving strategies adopted by successful dictators are, in point of rationality, strictly on a par with strategies which effectively implement less arbitrary political agendas?

A third way of drawing out the counter-intuitive consequences of the instrumental doctrine is to exploit the fact that the ends or goals we want to achieve—and the rules, principles and ideals we seek to apply—are hardly ever, if ever, adopted by us for no reason whatsoever. On the contrary, we adopt them for reasons of a wide variety of kinds. In face of the diversity of these reasons and in face of the fact that it is an essential feature of these reasons that they purport to be good reasons, we should ask whether it is possible to take seriously the instrumentalist view that it is simply irrelevant to the rationality of what we do when we pursue our ends or goals (and when we apply the rules, principles or ideals to which we are committed) whether the reasons for which we adopt the ends or goals (and the reasons for which we commit ourselves to the rules, principles or ideals) are good, bad or indifferent. Successful strategies for the structuring of institutions in the light of irrational religious beliefs, for example, are surely not on a par—from the standpoint of rationality—with strategies for the effective implementation of socio-political ideals for which good reasons can be given.

III. The Instrumental Doctrine: Shared Ground between Instrumentalists and Non-instrumentalists?

According to Robert Nozick, (1) instrumentalists and non-instrumentalists have something in common: all endorse the instrumental doctrine of rationality. They disagree only about whether there's more to rationality than instrumental rationality.

But what if, from the standpoint of the non-instrumentalist, what is added to the instrumental doctrine alters its force? What if the account of what makes goals rational—the account setting out the "substantive" doctrine of rationality such a non-instrumentalist might be taken by Nozick to be "adding" to the instrumental doctrine—makes it necessary for the account of what it means for an action to be instrumentally rational to be revised in a radical way instead of simply being "taken over" from the instrumentalist?

Thus, suppose our non-instrumentalist comes up with an account of what makes a goal G rational. Let's call this part of the proposed doctrine of practical reason theory ALPHA. What is the relationship between theory ALPHA and the instrumental conception of rationality which—in common with all non-instrumentalists, according to Nozick—our non-instrumentalist also endorses? Could it be that these two parts of the comprehensive view on offer purport to give accounts of unrelated phenomena—ALPHA an account of what makes G rational and the instrumental doctrine an account of what makes effective strategies for attaining G rational?

Take the case where a given goal G is—by theory ALPHA—rational. If a proposed strategy for its achievement (S, say) is also rational—this time in light of the instrumental doctrine (i.e. because S happens to be an effective and economical means of achieving G)—should we not say that, for our non-instrumentalist, there is a connection between the rationality of S and the rationality of G? Should we not say, that is, that part of what makes S a rational strategy is that the goal it serves, viz. G, is a rational goal? Prima facie, this is much the most straightforward way of integrating the two parts of the non-instrumentalist position.

On this account of the structure of the non-instrumentalist position, theory ALPHA has a role to play not only in the answering of the question whether G is a rational goal but also in the answering of the question whether S is a rational strategy for its attainment. In order to have something to say about what makes strategies (instrumentally) rational, the non-instrumentalist will not be in a position simply to incorporate the instrumental doctrine into the new (more comprehensive) account of practical reason: the doctrine cannot be appropriated without also being transformed in the process. There will consequently be two differences between the non-instrumentalist who subscribes to theory ALPHA and the instrumentalist. (2) First, the one has an account on offer about what makes goals rational, while the other has no such account. Second, the one maintains that the rationality of the goal served by a strategy is part of what makes the strategy rational, while the other associates its rationality solely with the efficiency with which it contributes to the attainment of the decision-maker's de facto goal.

IV. Towards a Non-Instrumental Doctrine

A standing obstacle to the rejection of instrumentalism has been the suspicion that it is the only theory of practical reason which is unencumbered by an extravagantly ambitious account of the ground of the rationality of the (non-"proximate") ends we adopt or of the rules, principles or ideals to which we subscribe. To this suspicion there are two responses.

The first is that if wide-ranging skepticism about our ability to sponsor a credible account of the rationality of our ends and commitments proved to be justified, the conclusion we should have to draw—out of deference to the linkage thesis—is that we are simply in no position to pronounce on the rationality of any of the decisions we make. It should be noticed, however, that drawing this conclusion presents no barrier to continued recognition of the role reason plays in the vetting of the factual premises of practical syllogisms. With Hume, (3) for example, we could continue to maintain that it is one of the tasks of reason to ascertain whether the causal connections between an agent's decision and some desired state of affairs are such as to make the decision an effective means of bringing about the state of affairs. However, if we wanted to follow Hume in refusing to allow that the terms "rational" and "irrational" have any strictly correct application to such things as actions or decisions, (4) our theory would not deserve to be regarded as a theory of practical reason at all.

An alternative response to the suspicion is to say that it arises out of an exaggerated fear that it is beyond our cognitive capacity to come up with answers to questions about the rationality of goals and commitments. The fear seems to be an exaggerated one because a fairly modest account of what is involved in the vetting of the ends of action (or in the critical assessment of rules, principles or ideals) may be all that is needed when it comes to giving content to a non-instrumental doctrine of practical reason. If we set aside the rather formidable language of "rationality" and take our cue from the close connection there is between claims about what it is rational to believe (or do, or pursue, or commit ourselves to) and claims about what there are good reasons to believe (or do, or pursue, or commit ourselves to), a non-instrumental doctrine of practical reason is much more likely to be seen as indispensable than as unattainable.

To say that a belief is rational is at no great distance, it seems plausible to hold, from saying that there are good reasons for the belief. Similarly, to say that an action is rational comes to much the same thing as saying that there are good reasons for performing it. Again, to raise questions about the rationality of an agent's ends is tantamount to raising questions about how good the agent's reasons are for supposing the ends to be worth pursuing, and to ask whether the commitment to given rules, principles or ideals is rational is more or less equivalent to asking whether good reasons can be given for endorsing (or rejecting) the rules, principles or ideals in question. Difficult though some of these questions about "rationality" may prove to be even when they are recast as questions about "good reasons", it is not plausible to dismiss them out of hand as plainly unanswerable. Indeed, they seem to be questions we should be able to handle without the positing of extravagant doctrines about Reason with a capital R.

bluered.gif (1041 bytes)


(1) Robert Nozick, The Nature of Rationality, Harvard University Press, 1993, Chapter 5 ("Instrumental Rationality and its Limits")

(2) Despite what Nozick says about there being only one difference between instrumentalists and non-instrumentalists—viz. a difference about what sort of supplementation, if any, is needed for the instrumental doctrine within a comprehensive theory of practical reason—he seems to feel the pull of the alternative account I have offered above. After saying that the questions he has been raising about instrumental rationality are not designed to challenge its "legitimacy," he adds, within brackets: "I have no doubt about that, at least in the case of instrumental rationality directed toward rational goals." (op cit, p. 134. Emphasis supplied.) The clear implication seems to be that there is at least some basis for "doubt" about the adequacy of the instrumental doctrine if, according to the doctrine, it is irrelevant to the rationality of a strategy whether the goal it serves is or is not "rational."

(3) David Hume, Treatise of Human Nature, Book III, Part I, Section I

(4) Although Hume is often assumed to be a champion of the instrumental doctrine, he is probably better interpreted as rejecting it. He denies that the rationality of an agent's decision can be inferred from the fact that the "means-end" judgment on which it rests is true. The inference is blocked, not because it violates what I have called the linkage thesis (though it is a nice question whether Hume's objection to the derivation of an "ought" from an "is" should be treated as evidence that he would have been sympathetic to this thesis), but because it is incompatible with the very narrow (and wholly "theoretical") doctrine of reason he sponsors. He conceives of reason as a faculty of judgment—as capable of pronouncing only on the truth or falsity of judgments. Since "actions" are not judgments, they can be neither true nor false, so they simply aren't the sorts of things which can either conform to or violate the canons of rationality. ("Reason is the discovery of truth or falshood. Truth or falshood consists in an agreement or disagreement either to the real relations of ideas, or to real existence and matter of fact. Whatever, therefore, is not susceptible of this agreement or disagreement, is incapable of being true or false, and can never be an object of our reason. Now 'tis evident our passions, volitions, and actions, are not susceptible of any such agreement or disagreement. . . . 'Tis impossible, therefore, they can be pronounced either true or false, and be either contrary or conformable to reason.")

bluered.gif (1041 bytes)


Back to the Top

20th World Congress of Philosophy Logo

Paideia logo design by Janet L. Olson.
All Rights Reserved


Back to the WCP Homepage