Praxeology: A Literalistic System

Claudio Gutierrez

It is doubtful whether any language–except possibly the arbitrary symbols of mathematics and symbolic logic–is entirely literal, just as it seems that reality cannot be thought of in purely objective terms. . . .
Frank H. Knight

After our presentation of the paradigmatic interpretation of theory it should be clear in the mind of the reader that a scientific system is not simply a formal framework where formal models find their right place. Basic or fundamental assumptions, indispensable for the existence of a theoretical system, are in themselves of an informal character. The residual categories that represent them within the formal system are open rather than closed concepts. The shade of formalism that accrues to the assumptions comes only from the fact that they govern formal models and must figure as "token" data for the logical closure of the system. No language is entirely literal since it interacts with the truths it tries to express. Not even logic or mathematics is completely literal. If their axioms are sophisticated enough then we never know exactly what they mean. "If we knew, we could avoid the possibility of asserting in one axiom what another axiom denies"–and we have no warranty that this is possible (POLANYI 64, p. 259). No language is watertight, and every line of reasoning must draw from an inarticulate background of non formal and ultimately unformalizable knowledge. The non realization of this fundamental truth is precisely the concurrent mistake of apriorism and positivism. Our language is neither literal nor unidimensional, or capable of a complete and exhaustive formal development.

One can take two authors as good representatives of these positions: M. Friedman and L. Von Mises. The former sees economic theory as only a conventional filing system for the ordering of empirical material (FRIEDMAN 53, p. 7). The latter sees it as a priori knowledge, a branch of praxeology or the theory of human action, which happens to be applicable or relevant in relation to some historical instances. Says he:

Praxeology is a priori. It starts from the a priori category of action and develops out of it all that it contains. For practical reasons praxeology does not as a rule pay much attention to those problems that are of no use for the study of the reality of man's action, but restricts its work to those problems that are necessary for the elucidation of what is going on in reality. . . . This does not alter the purely aprioristic character of praxeology. It merely circumscribes the field that the individual praxeologists customarily choose for their work. . . . (MISES 62, p. 41)

I think that it is possible to argue against each of these authors reasoning from his own point of view. It is possible to show that their positions are capable of being "stretched-out" to cover much ground of the opponent's conception. Thus, one could be "more empiricist than Friedman" and maintain that our dependence on experience is so great that we would be without a language (without a filing system) but for the creative intervention of experience: in ultimate analysis only experience can originate a language. One could also be "more theoretical than Mises": one could maintain that our dependence on theory is such that we would be unable even to identify the relevant facts for the application of praxeology were it not for the circumstance that theory itself (language) already gives those facts to us. The proof of our first assertion was already given in our discussion of Verstehen. Substantiation of the second one may require some more words.

The argument for the "stretching-out" of praxeology is connected with the problem of the compatibility of the a priori conception with the possibility of application of a praxeological theorem. I put forward the dilemma that if the theorem is a priori in the unidimensional sense in which praxeology seems to be intended, then the theory as represented in the theorem is inapplicable. If it is not a priori in that sense then praxeology is already defeated. The difficulty I see here has to do with the description of the (empirical) conditions that must form part of the theorem in order to be applicable. Even if the theorem is a priori it has to mention the factual situation under which one is saying that the theorem is valid. But this mention has to be made in a language and the language one has to use must not be a purely formal NOTE 1 one. If it were, empirical conditions would be inexpressible in the language and the theorem, without the mention of those conditions, would be inapplicable. Again, an empirical language capable of mentioning the conditions of application of the theorem would have to have been learned in close intercourse with experience. Furthermore, that experience should have occurred in precisely the area where praxeology is claiming to have something to say. Therefore, we are led to the conclusion that the application of a praxeological theorem supposes already the (empirically acquired) economic language and, by implication, (empirical) economic knowledge. The historical origin and development of praxeology (or marginal-utility theory) lend much strength to the belief that this is precisely the case. It has appeared at the very end of a long and complex evolution of concepts and theory and not at the beginning of it. And one should have expected the latter were the system of praxeology really an unidimensional, literal theoretical construction. One could here reply that historical order needs not to coincide with the logical order; it is true, and I do not consider the argument as a principal or substantial one. But it is difficult to deny that it lends force to the main reasoning, that is, to the dilemma of inapplicability of a pure praxeological theorem.

I will present in the rest of this chapter a concrete analysis of a considerable part of a praxeological system (ROTHBARD 62). I will try to further my case as to the invalidity of the belief that, as logical inference goes, the system is a literal, strictly lineal, or waterproof one. NOTE 2

Before entering that analysis, I want to call the attention of the reader to an attack on contemporary logic which is very interesting. It is found in the introduction to Rothbard's book together with a defense of the verbal (less formal) logic of the past. It is interesting because one could think that with the help of a less rigorous logic it might result possible to "formally deduce" praxeological theorems, using nothing but the praxeological axioms or premises. Although the discussion which is to follow will be conducted "with the weapons chosen by the adversary" and I will refrain from using any esoteric formula of mathematical logic, I think it necessary not to let these concepts pass without at least a perfunctory refutation. Says Rothbard:

It is the great quality of verbal propositions that each one is meaningful. On the other hand, algebraic or logical symbols, as used in logistics, are not in themselves meaningful. Praxeology asserts the action axiom as true, and from this (together with a few empirical axioms such as the existence of a variety of resources and individuals) are deduced, by the rules of logical inference, all the propositions of economics, each one of which is verbal and meaningful. . . . Logistics, therefore, is far more suited to the physical sciences, where, in contrast to the science of human action, the conclusions rather than the axioms are known. . . . (ROTHBARD 62, p. 65).

It is just not true that the propositions of traditional logic are, all of them, directly or semantically meaningful. What about "S-P"? Moreover, if the logical aspect of a proposition is what is being considered, even in traditional logic, the proposition is not taken as semantically meaningful but rather as a token of some formal structure of general validity. It is clear that when one studies "verbal logic" one is not interested in the mortality of Socrates, even if constantly using the phrase "Socrates is mortal". On the other hand, "logistic" propositions are always meaningful in the sense of requiring the full cooperation of non formal rules. The rules are explicit rather than implicit, and this seems to be a big difference between the two logics, to the clear advantage of symbolic logic. But logical propositions of the new garb are also meaningful in the sense of the professional identification referred to before. The trained logician knows at each step what he is doing, even if the scholarly presentation of the matter, for example in a textbook, requires that he pretends to ignore it. It is also inexact that the physical and the social sciences must differ from each other in a parallel fashion, one being deduction from not-known-to-be-true hypotheses and the other from evident propositions. The demonstration of this could, in a way, follow the same line as the previous argument. More radically, that demonstration is connected with all that has been said in this study; very especially with what we are about to say in criticism of the praxeological system.

The following is our selection of statements from Rothbard's reasoning: NOTE 3

1") Human action is defined simply as purposeful behavior. . . . (ROTHBARD 62, p. 1)

This, I consider, is the basic axiom or fundamental praxeological postulate. If now we ask what its logical status is, one could try for one of these: either a nominal definition, or a useful device, or a real definition, or an empirical truth, or a category (if one accepts that there is a difference between "category" and "real definition"). We should not be concerned here with that status since (1") is the axiom of the system and presently we are not quarreling with the status of the axiom but only with the character of the deduction. However, it is very important that we establish in a precise way what the informative content of the axiom is, so as to make certain that some statement are, or are not, implied by it.

Now, the question arises whether we are being too demanding in precision for the basic postulate. Did we not say that ambiguity is ineradicable from all knowledge bearing on reality? Yes, but Rothbard has not said so, since literalism consists in the negation of that ever present ambiguity. To plead for recognition of ambiguity at the very start would be to concede the whole argument. Another important question is this: we are now concerned only with the criticism of the chain of reasoning, and not directly with the primitive proposition and its truth or logical status. Are we going to return to the latter problem once we are through with the former? The answer to this question is no. The logical status of the primitive axiom will be of much less interest once it will have been proved that one cannot derive from it alone the other propositions of the system. The non recognition of this is still a case of the literalistic fallacy itself. There are some other ways of justifying theory which have nothing to do with the pretended apriorism of certain axioms or principles, as it is hoped this essay itself has been able to show.

If now we concentrate on the issue of establishing a precise sense for the axiom, we will find the task very difficult. The issue is fogged with ambiguity in the praxeological literature. Take for example the efforts in clarification that one finds in the work of Israel M. Kirzner. For him praxeological rationality consists in the "consistent pursuit of one's own purposes " (KIRZNER 60, p. 32). The use of the word "consistent" introduces a complication of terminology since clearly it is not only logical consistency that is implied. One should like to say that it is rather the persistence of a purpose as such, as a purpose, that is intended, the invariability of the ends and the respective line of action during a definite span of time. Still, there is the additional declaration that "in the praxeological view, action is rational by definition" (KIRZNER 60, p. 167). To that view, even "a man who is swayed from the pursuit of his own best interests by falling prey to a fleeting temptation is yet acting 'rationally' in the praxeological sense. In the praxeological view, the man has simply substituted a new set of ends. . . ." (KIRZNER 60, p. 168-9). This forces us into concluding that praxeological judgments are intended as true only in relation to assumed (fixed) programs. They depend completely on a tendency of human beings which demands that given programs be respected. It is said that "the selection of an end can never, as such, be judged in regard to its rationality. . . ." (KIRZNER 60, p. 169). One cannot avoid the implication that praxeology as an a priori injunction must be somehow equivalent to a plea that ends remain invariable: Consilia sunt servanda! Under this light, I think, we can paraphrase (1") as

1') There must be in the world such a thing as persistent conscious motion toward a fixed goal.

Now, if, according to all that has been said, we are going to disregard in the axiom all that is not information; if, in other words, we are going to disregard its existential aspects to concentrate only in its conceptual content, then (1') can be simplified further to read simply as a nominal definition:

1) Human action is persistence of ends.

It is there where one should be able to find the whole content of praxeological knowledge in an embryonic fashion; i.e., the knowledge that we are supposed to encounter afterwards in expanded guise in the rest of the system. Hence, (1) will be our starting point or primitive statement for the critical demonstration game.

2) We could not conceive of human beings who do not act purposefully . . . . (ROTHBARD 62, p. 1)

3) The first truth to be discovered about human action is that it can be undertaken only by individual actors . . . . (ROTHBARD 62, p. 2)

Proposition (2) is reserved for later analysis. Proposition (3) seems not to be a nominal definition simply equating "actor" with "individual". Rather, it seems to be saying that there are no collective actors. It is a real rather than a nominal definition. I contend that it is not implied by (1) unless one makes (1) to imply it. I can conceivably take (1) as allowing for collectives "consciously moving toward a goal". If I rather prefer not so to take it, then I am making a dialectical decision, drawing an addition or correction to the original picture or primitive interpretation of the content of (1). The addition would not be arbitrary, however. It would be based on the tacit knowledge I have about how people act, on an elicitation of what we have come to see as the normal use of the words "human action". This being so, one begins to wonder whether the original interpretation exists as a separate axiom; or rather the proposition "fully, clearly and necessarily present in every human mind" (MISES 62) is nothing short of the whole of ordinary language continually reinterpreted by the very use we make of it. More simply, one could accept, at least for the sake of the argument, that there is some original, non-rich, interpretation of (1), and that as one learns about human action, one gradually enriches it with new content. Let us give a name to this operation of altering the interpretation of a given primitive statement in order to convey more information. Let us call it "dialectical redefinition", or perhaps better, "retro-definition". We will have occasion for repeated use of this newly-coined term in the course of this analysis.

4) Action requires an image of a desired end and "technological ideas" or plan on how to arrive at this end. (ROTHBARD 62, p. 2)

I think this is another clear case of retro-definition. I can conceivably take (1) to allow for "magical ideas" being used in our pursuing of ends. Why not? But we tacitly know, independently of (1), that magic just would not work–although "we" in this context should not be identified with the whole human race. So, we prefer to take "purposeful behavior" as implying "technological ideas."

5) All action aims at rendering conditions at one time in the future more satisfactory for the actor than they would have been without the intervention of the action. (ROTHBARD 62, p. 3)

As with (3), here one has to choose between interpreting the statement in a nominal way or in a real way. The difference is in this case that one could not say that (5) is a true real definition, for the simple reason that it will not be a true statement at all! In fact, I know that I sometimes act not for altering the future but merely for enjoying the present, i.e., the action itself, as in play, and artistic or religious contemplation. To maintain the contrary would be equivalent to saying that one never does anything except for the future. The unpalatable result will be that, practically speaking, I have no present at all. But this is clear nonsense. I enjoy, very often, the action of not-being-concerned. This is what we commonly call "relaxation". Since the real interpretation is false, one has to take the other horn of the alternative: the definition is a nominal one. More properly speaking, this being an alleged inference from (1), it is a (nominal) retro-definition. But I have defined "retro-definition" as a device that tends to make the original definition convey more information; and it is certainly the case in this occasion that the nominal definition tends rather to make it convey less information; hence I am led to propose for it the rather awkward name of "inverse retro-definition". Its function is to "save" the original definition or some of its desired consequences from the assault of adverse experience. It is clear that the intention of (5) is the postulation of Homo aeconomicus. Its retrodefinitional form makes the postulate appear as a deductive inference from the "basic axiom". NOTE 4

6) Action takes place by choosing which ends shall be satisfied by the employment of means. . . .

7) When we must use a means so that some ends remain unsatisfied, the necessity for a choice among ends arises. . . . (ROTHBARD 62, p. 4)

These statements simply do not follow from (1); neither do they follow from any of the propositions (l)-(5). We suddenly begin to read about several ends (for a single actor, as understood) when in (1) there is no mention of any kind of multiplicity. One can, of course, reply that this is an auxiliary (and empirical) hypothesis. But this just would not do, because (1) does not talk about multiplicity of ends and hence it is inapplicable to the case envisioned by the hypothesis. We have to reinterpret (again!) the original statement so that it may allow for multiplicity of (simultaneous) ends. But neither would this do because ends may be either compatible or incompatible. If they are the former, they are one (bigger) end, not really several (conjunction is a very simple logical operation). If they are the latter, then they are no end at all (the actor does not know what he wants). Reinterpretation of (1) in terms of compatible ends is superfluous; in terms of incompatible ends, impossible. Is there a way out? There is, but it implies a full analysis of the crucial notion of "substitution," not present at all in the basic axiom short of reinterpretation beyond possible recognition.

8) All means are scarce. . . . (ROTHBARD 62, p. 4)

The analysis could run parallel to the analysis of (5). The postulation being made here is the dogma of the applicability or marginal analysis to the real world. This is also an inverse retro-definition, therefore it is of nominal type. It is not clear, however, as in the case of (5) that the real interpretation is false. What is claimed, beyond a restatement of the essential content of (5), is that there will never be abundance in the world. The persistence of insatisfaction is asserted, a new flat addition to the original assertion of persistence of ends. Prima facie, that addition seems to be true, although recent technological, medical, and social developments make less improbable that a state of practical non scarcity could be some time attained by human society (AYRES 61, pp. 230 ff.). Therefore, the real definition is not clearly false, although it is not, to my mind, clearly true either. Because of this qualification, then, we can say that (8) could alternatively be interpreted also as a straight retro-definition (adding informative content to the original axiom).

9) The actor may be interpreted as ranking his alternative ends. . . . (ROTHBARD 62, p. 4)

Is this to say that there must always be only one end? For "ranking" in this context means to assign an ordinal number to every partial end so as to make compatible otherwise conflicting ends. A scale of preferences must be built and the scale itself is to be, from now on, the speaker for the formerly conflicting ends. Its voice is to be now the single end. One could say that while this ranking is being done the actor is not economizing but, perhaps, "philosophizing," since the selection of ends is not the business of praxeology. After the ranking is done, the actor is not economizing either but "mathematizing," since purely tautological operations are not the business of praxeology either. This may be called the dilemma of ranking. Its solution implies, again, an analysis of the concept of substitution and of the specific role of subjectivity in economic theory. But apart from that, it is also true that (9) is not a deduction from (1). The supposition that ranking must precede action is a clear addition to the original interpretation (1). One can easily imagine a situation in which action is performed with the inquisitive intention of finding out what the relative weights of one's wishes are. "Ends are more or less defined in the process of realization. . . . " (KNIGHT 56, p. 172).

10) All human choices are continually changing. . . as a result of changing valuations and changing ideas about the most appropriate means of arriving at ends. . . . (ROTHBARD 62, p. 6)

This statement seems to be an empirical generalization about the evolution of mankind in the ethical and the technological fields. In that condition, it cannot be a consequence of (1). Nevertheless, it appears in a section entitled "First Implications of the Concept [of action]". My special interest lies only on part of (10), namely the part before the first ellipsis, and it is interesting because it appears to be in direct contradiction to (1). But nobody wants to assert clear contradictions like this. So we have to save the consistency of the system by understanding (10) as a qualification of (1), i.e., we want to read "(1) or (10)" rather than "both (1) and (10)". Now, "(1) or (10)" is not a contradiction, but unfortunately it is a tautology. It says that "either ends persist or ends are continually changing". Can anyone save the informative power of such a proposition? A way of doing that would be to interpret the connection between the two statements in a quantitative sense, that is, the disjunction of them as asserting a differential degree of frequency or probability. If (1) is the main statement and (10) the qualification, the content of the former will be a description of the most probable or common occurrence in the real world.

But this, incidentally, would be irrefutable only in the restricted sense in which one says that every probability statement is irrefutable, not in the sense of a priori praxeology. NOTE 5 Again, there is another presumable way out. The proposition that we had reserved for later discussion, that is (2), says that one possibility is excluded from the whole range of possibilities open by the tautology "(1) or (10)": the absolute extreme possibility of all human beings being such that they would be changing ends every discernible unit of time. This interpretation is not completely irrelevant. Some minds could find spiritual comfort in the fact that there is some, at least some–however fleeting–permanence in human endeavors.

11) Means. . . are called goods. . . . Such goods may all be classified in either of two categories: [consumers' goods or factors of production]. . . . (ROTHBARD 62, p. 6)

12) The factors of production may all be divided into two classes: Those that are themselves produced, and those that are found already available in nature. . . . (ROTHBARD 62, p. 6)

All these are logical divisions. Logical division is no implication from an original proposition that uses the undivided concept. It cannot be, since the original proposition does not distinguish between the parts that are the result of division, which the "derived" propositions speak of. All divisions are either nominal–and to that extent arbitrary–or real, and therefore dependent on experience. If they are real, they amount to empirical generalizations. They are no deduction from purely formal axioms. It is in the nature of logical division to have either one of these two purposes: to serve as a catalog (a filing system) or to serve as taxonomy (a system of concepts which somehow reflect the natural divisions of the world). It goes without saying that the divisions in (11) and (12) are not intended as a simple catalog. They must be intended as a taxonomy, and so they are. But to make this taxonomic division, one has to use fully the (empirical) economic knowledge that is at hand. One still would want to say that these divisions are also "auxiliary hypotheses" in order to "apply" or make "relevant" economic "pure" theory. So let them be! But then it becomes apparent that the term "relevant" is being used in an awkward way and that there is no difference between the sense of your word "relevance" and the sense of my word "truth".

13) If we wish to trace each stage of production far enough back to original sources, we must arrive at a point where only labor and nature existed and there were no capital goods. . . .

14) There is another unique type of factor of production that is indispensable in every stage of every production process. This is the "technological idea". . . . [Once learned] it becomes a general condition of human welfare in the same way as air. (ROTHBARD 62, p. 9)

It is inherent to any logical division that it facilitates some talk and hinders some other. In this sense we can say that some divisions are true and others are false. Some divisions are "truer" than others–they cannot be totally arbitrary. In the particular case of divisions (13) and (14) we can run into linguistic trouble. This is another strong reason against the praxeological approach. According to (13) one cannot talk about capital before the humanization of man. Nevertheless, some economists or philosophers might think it profitable so to talk. Knight, for one, likes to talk that way. He insists on the fact that capital is always produced by another capital, even man himself is capital (KNIGHT 51b). Another hindrance occurs, according to (14), with respect to knowledge, in many a respect the most productive and valuable capital of all. But we cannot call it capital, not even a factor of production, not even a good, not even a means, since it is unlimited! If one cannot treat knowledge as capital in the praxeological setup, then something is wrong with that setup, or perhaps some radical limitations in the nature of economic thinking are being uncovered in this connection.