, ">

“On Practice”

It is not unreasonable to assume that a theory and, in general, any cognitive undertaking does not shape up and develop entirely by itself, in complete independence of strictly non-cognitive contexts, such as systems of beliefs, ideologies, world views, and even the conditions—historical, social, economic, national, etc.—out of which the contexts arise. In this sense, theories are heteronomous.[1] Yet they do not seem to be completely heteronomous in so far as a certain degree of what can be called "internal evolution" can be detected in the formation and development of theories. The degree of this internal evolution, which makes a theory at least relatively autonomous, varies according to the type of theory considered. Thus, in the formal sciences—probably because they are not stricto sensu cognitive—the degree of autonomy is maximum, and, as some would say, complete, although contexts may occur, as with the natural sciences, that often mold the ways in which the formal sciences develop. There is less autonomy in the natural sciences than in the formal sciences. In the social sciences and, for somewhat different reasons, in philosophy, the degree of autonomy is considerably reduced, and, according to some authors, practically nonexistent. In any case, however, one can admit that there is some autonomy, and in particular, what we may call "formative autonomy. "

It may, and often has been argued, that heteronomy only concerns the formation and development of theories, namely, the concrete ways in which knowledge is produced; heteronomy therefore, pertains to what is usually called "the context of discovery." Such a context can oscilate between a "minimal context"—conceptual frameworks, assumptions not far removed from, or too alien to, the theories—and a "maximum context." The latter does not necessarily include "everything," for otherwise interminable descriptions of any "process of discovery" would ensue. It may include, however, such contexts as ideologies, world views, etc., on the one hand, and political and social institutions on the other hand. It is then contended that no matter how much these contexts may clarify the discovery of a theory, they will say little, or nothing, about whether or not the theory—or the thesis, or the hypothesis—at stake is acceptable, or valid, etc. The latter point seems to be within the jurisdiction of the so-called "context of justification" or, as I will also call it, "context of validation," where factors contributing to the formation and development of a theory or, in general, of knowledge are not normally taken into account unless they are a part of the methodological procedures. Accordingly, it is affirmed that even if knowledge were formatively heteronomous, it would still be autonomous "validatively."

Neat distinctions are like well focused pictures: their degree of "definition" is high. Hence they are eminently persuasive, for we can always appeal to them when the images begin to get blurred. The question is whether the price paid for a neat picture is too high.

At the present stage, it does not seem to be too high. The distinction between the context of discovery and the context of justification (or validation) serves to warn us against perfectly unnecessary confusions, worth no price at all. Furthermore, as Wesley O. Salmon has emphasized, it is possible to maintain the distinction between the two without assuming that they exclude each other. The same "factor" can thus be pertinent both psychologically—and, I assume, also sociologically—and logically to a given theory or hypothesis. A factor common to both contexts does not obliterate the distinction proposed since it functions within each context in its own way. If such is the case, the price to be paid for the distinction between the context of discovery and the context of justification is quite a bargain.

Furthermore, if this were not enough, the distinction in question avoids the pitfalls of piling everything in a kind of conceptual heap. Such heaps emerge as soon as we try to relate a theory to rather "distant" contexts, as is the case with social and political ideologies, not to mention the concrete conditions from which these ideologies are supposed to emanate, whether reflectively or reactively. Let us call any theory 'T', and any context of the kind alluded to 'C'. It is only reasonable to contend that even if C were the cause of T, C could not explain T other than genetically, and C would not logically imply T. T is not derivable, in any sense of the word, from C, and in this respect T is autonomous with respect to C. If T could be derived from C, then it could be shown that it is so derivable. If, for example, it is held that a political ideology, I, forms the context for the justification, or validation, of T, or of the statements in T, then these statements should be derived from I. If they are so derived, then the justification or validation, and not only the real production, of T by I will be in perfect order.

To the best of my knowledge, at present, those derivations which have been made were either unsatisfactory, or they have failed to distinguish between the contents of T and the contents of I—sometimes I is even a rationalization or a specification of I. Even in such a case, however, it is doubtful that non-genetical derivations—to say nothing of logical ones—have been performed. We may trace in T a number of evaluations, prescriptions or orientations which originated in I, and we still wonder to what extent any of them helps to validate T.

Now, the distinction between the context of discovery and the context of justification works adequately only within the rather restricted conditions for which it was devised, namely, conditions in which derivability, as a tool for logical and methodological reconstruction, is of paramount importance. This, however, presupposes a certain notion of "theory," which is by no means mandatory. Other theories concerning "theories," and, in general, concerning "knowledge" may enter this competition in the two areas which I am now hurriedly exploring: the philosophy of science (as well as the philosophy of philosophy) and the sociology of knowledge. One of these "non-orthodox" theories I am quite willing to defend.

My defense ought not to be understood as a suggestion that one go to the other extreme, like those who have rejected the distinction between the context of discovery and the context of validation. Such a rejection was possible only because one of the contexts was eliminated and the latter, in turn, happened when attention paid to the relations between the propositional structure of a science and its extrascientific contexts led to the conclusion that this structure is theoretically justifiable—and not only genetically describable or explainable—by dint of certain empirical conditions, particularly psychological, social and "institutional." Thus justificationist absolutism has been replaced by psychologistic or sociologistic absolutism, and I sometimes wonder which one is the better, or the worse, of the two.

My plans are easier to announce than to fulfil, but I will try against all odds.

Let us consider, to begin with, what type of relations other than derivability or causal production, can be established between a theory, T, which tries to be "objective," and thus "autonomously" justifiable, and a context, C. Since not all theories, and certainly not all contexts are of the same kind, it is likely that even if a new given relation is discovered, it will function quite differently according to the kinds of T's and C's considered. I shall leave these annoying details to non-philosophers.

It is sometimes held that C is a part of the validation, or justification, of T in so far as it is a necessary, but not a sufficient, condition of T. This view seems acceptable to me, but not trivial. If, for example, a social ideology is the necessary, but not the sufficient, condition for a scientific theory, then the ideology will be part of the discovery, formation and development of the theory, but it will neither help nor hinder the validation of the theory.

My plan is a strategy, and it is carried on in two tactical steps, which could be called "maneuvers." The first maneuver consists in starting with what I have called "theory," in order to try to relate it, intelligibly, or rationally, to as many contexts as possible. The second maneuver starts with some context in order to see how theories can be brought to bear on it.

First maneuver. A viewpoint on "theory" somewhat different from, although not necessarily incompatible with, the one ordinarily associated with the distinction between the context of discovery and the context of justification (or, again, validation) can try to trace in the theory some elements which, to begin with, may be not too far removed from the theoretical structure. Among these elements are expectations of results which are not logical derivations of the theory—as, for example, the capacity to adapt the theory to situations which had not been previously considered; the possibility of modifying the theory changing (generally, enlarging, but sometimes restricting) the meanings of some of its terms, or the scope of some of its concepts; the introduction, or substantial modification of an observational language, or parts thereof. These "elements" are a function of contexts which operate as theoretical assumptions and which do not need to be removed from the theory as if they were "only" empirical conditions of its production. The distinction (between the context of discovery and the context of justification) thesis does not eliminate altogether the aforementioned elements, but tends to mistrust them. To be sure, it tends to consider an observational language as extra-theoretical. I am aware of the fact that it is not always easy to pin down the "elements" or "factors" just alluded to, and that when they can be pinned down, it still remains difficult to ascertain how they are related to the theory. The innumerable debates on the nature and function of the so-called "assumptions" or "pre-conceptions" show that this matter is far from being clear in the mind of everyone (they are not clear in my mind). Nevertheless, if there are some such assumptions or pre-conceptions, they function theoretically or, in any case, cognitively. When the latter are considered, a theory becomes richer, namely, it increases its cognitive content. Consider in this respect two quite different theories: the theory (principle, or law) of inertia, in its Galilean version, and the nominalistic theory (or doctrine) of Occam. The principle, or law, of inertia, as it is ordinarily formulated in basic Physics textbooks, does not seem to contain assumptions or pre-conceptions of any kind, unless we count among them stipulations of the "in ideal conditions of absence of friction . . . " type. It has been shown, however, (by Hanson and Koyré among others), that the conceptual content of the principle is richer than its textbook formulation. Similarly, Occam's nominalistic doctrine has theological pre-conceptions and yet this doctrine is not derived from them, since the same pre-conceptions could serve to outline a non-nominalistic theory. These pre-conceptions, on the other hand, are not indifferent to the arguments pro and con Occam's views. In point of fact, without these pre-conceptions, Occam's doctrine would cease to be what it is, and the arguments adduced in its defense would probably change drastically.

It should be pointed out that the assumptions of a theory can become oppressive rather than liberating—or impoverishing rather than enriching. An increase in the cognitive content of a theory is not obtained by eliminating all assumptions, but rather by trying to find some others. It often happens that a change in the theory increases its cognitive content.

Other elements or factors which it is reasonable to assume are a part of the structure of a theory contribute to its validation. Such elements include: the community of researchers; the consensus by virtue of which a theory is accepted, or rejected; and the possibility of communicating the content (meaning) of the theory and of its terms. 

I do not understand by "the community of researchers" a given group of human beings doing research. Although there must be such a group, concretely operating within given social and cultural conditions, what matters now are rather the rules which make such research acceptable or subject to intelligible discussion by the members of the group. These rules are not only methodological rules, but also, and perhaps above all, rules which circumscribe the content and scope of the research. Fundamentally, they are the series of steps and "maneuvers," both positive and negative, which Lakatos has described as "scientific research programmes." To some extent the methodology of such progammes coincides with what I have previously called "assumptions" or "pre-conceptions," but the latter has a stronger pragmatic character than the former. This does not mean, however, that they must be rejected as not being a part of the ensuing theories. Pragmatic reasons include important elements concerning the context of discovery. Such reasons are coincident rather than incompatible with validation. They might also be the locus where the two contexts—discovery and validation—meet.

Something similar can also be said of the two other "elements" or "factors" already mentioned, particularly the element of "consensus," a rather ambiguous and even equivocal term since it includes the notion of disagreement. Unless we admit a quasi-Platonic criterion of ideal objectivity which allows us to decide whether or not a given theory is acceptable or the arguments adduced in its favor are plausible, or whether a given observational language is adequate or inadequate; or unless we subscribe to some notion of 'irresistible evidence,' we cannot help but conclude that what is called "objectivity" is, in the last resort, the same as what is inter-subjectively agreed to. This does not mean, of course, that a consensus is reached by vote counting, or by tyrannical rule, or by more or less "authoritarian" and "authoritative" influences— even if this is just what sometimes happens. No matter how much one argues, reasons, computes, or appeals to empirical data formulated in an observational language—which, by the way, may itself be the subject of some "consensus"—what in the last resort makes a proposition acceptable or not acceptable is the consensus, or agreement, that it is, indeed, acceptable or not. This consensus functions within the rules laid down, implicitly or explicitly, by the community of researchers by virtue of habits engendered by a multitude of common experiences. The consensus is tied up to the possibilities of communication and, therfore, to some "linguistic" context. The latter must be understood in a rather wide sense, as covering all the possible criteria of rationality. Thus, the "linguistic context" which is associated with "elements" still wider in scope than those mentioned up to the present, will stretch the context of validation to its limits—something I am sure some will consider frankly intolerable.

We must now consider what has been named "world views," an expression whose meaning is so nebulous as to suggest that we might be wise in dispensing with it. I believe, however, that we can utilize it by considering it as a basic and very large frame of beliefs and evaluations. On the one hand, a world view provides orientations for the understanding of, and action on the world (generally, selecting some assumedly fundamental features or elements), and, on the other hand, and concomitantly, it helps us to perform intellectual, or practical, decisions when diverse and perhaps conflicting interpretations or norms are presented.

Now, although both function in similar ways, a world view must not be confused with an ideology. The latter is usually confined to beliefs and evaluations concerning the political and social environment. For "preferential reasons" I will concentrate on the so-called "ideologies," regardless of whether they are of a "conformist" or of a "non-conformist" type; namely, whether they are, in Mannheim's vocabulary, "ideologies," reflecting a social system more or less faithfully, or "utopias," reacting to a social situation by proposing a new ideology. In both cases, ideologies are linked to social conditions, without which they would have little, or no sense. For this reason, it is sometimes said that if a theory is socially conditioned, and if this condition is a part of the context of validation, then the theory will have to be validated within the scope of such conditions.

Now, I believe that this has to be taken cautiously. Not even those who most staunchly defend theories as a function of social conditions would, in their sanest moments, dare to say that these very conditions validate, or fail to validate, the theory. This is true even if we take "validate" in an ample sense, rather than equating it with "logically derive." For example, an appeal to social conditions cannot validate or invalidate an optical theory even if the latter has emerged within given social conditions—for the latter could give origin to several optical theories. Thus the notion of "social condition" really expresses a condition rather than a cause. It is quite another question to ask whether an "ideology" can, or cannot play a part in the procedures which validate a theory. I am thereby raising the problem of whether there is any relation, not strictly genetical in character, between what we may call "ideological conceptualization" and "theoretical conceptualization" proper.

It can be claimed that an answer to the above question depends upon what kind of theory we have in mind, since, as I already pointed out, it is customary, and indeed correct, to maintain that there is a closer relation between ideologies and social theories than there is between ideologies and theories in the natural sciences, where the relation becomes quite problematic. Now, unless we wish to admit that there is an essential rather than a gradual difference between various types of sciences, it will be mandatory to answer the question unequivocally, so that what is asserted applies to all kinds of theories, no matter how different they are among themselves.

My opinion is that the notion of justification, or validation, can be expanded from a strictly methodological and "reconstructionist" area to a much wider area, where validation does not follow the theory once it is constituted, but rather actual constitution of the theory and its validation overlap. Thus, classical preformist theories in genetics adduced facts which validated the theory, but at the same time the possibilities of validation—including, among other things, the selection of an observational language—were not alien to the ways in which the theory was formulated. In present day genetics a theory has been developed according to which the nucleus of each cell in an organism possesses all the necessary information to reproduce the entire organism. If many cells prove incapable of such reproduction, it is only because the nuclei are programmed for the performance of specialized functions. The theory alluded to is largely based upon research on the composition of chemical substances which transmit, and select, information though molecules of ribonucleic acid. Thus we can say that the results of this research not only serve to confirm the theory, but at the same time condition its structure as well as its development.

It may be claimed that the orthodox notion of validation, understood as the logical and methodological reconstruction of a theory, takes into account confirmability, as well as its degrees. But what it does not take into account is the fact that the very structure of the theory—not only its actual development as a process of "discovery"—is molded by the possibilities of validation. Above all, it does not take into account that validation can also include a research programme which serves as a framework for the formulation of given types of theories. The steps taken to constitute and develop the programme are parallel to the phases of theoretical construction, and since these steps include procedures of validation, without which they would be mere gropings in the dark, it is reasonable to conclude that at least part of the so-called "validation" is not alien to the steps taken to constitute and develop the programme. Therefore, validation cannot be viewed as a mere logical and methodological reconstruction.

Some readers may say that the notion of "validation" is so tenuous that it scarcely deserves the name any longer. I agree, but this was precisely one of the aims in my strategic-tactical maneuver: I am trying to suggest that such a notion can be expanded to the extent that when it reaches certain limits we are no longer certain of whether we are dealing with validation or with certain discovery. They may be two sides of the same coin. In order to lessen the intellectual shock which this invitation to conceptual indeterminacy might produce, semantic stipulation can be laid down which will appease hard-boiled methodologists. I suspect, however, that it would be best to consider particular cases and ascertain whether, and to what extent, the notion of validation can be expanded without losing its meaning.

Now, to what extent can an ideology be part of the validation of a theory? I will offer some suggestions which I hope are not entirely foolish.

An ideology can be considered as an expression of "human practice," and specifically of what Jürgen Habermas has called "interests," defined as basic orientations rooted in the conditions that make reproduction and self-constitution of the human species possible. In any case, practice provides certain orientations and establishes certain aims for the development of scientific research programmes and, in general, of knowledge. These orientations and aims contribute to the validation of the theories developed in the program in so far as they single out or select, the elements to which procedures of validation apply. In this sense—but in this sense only—it can be said that "practice" orients theory, and plays a part in establishing the criteria for the validation of the theory.

Once this is acknowledged, however, it must be pointed out that the procedures used in validating theories cannot be derived even indirectly, from "practice" and from ideologies. Firstly, as I surmised a while ago, parts of theories develop "autonomously," and the more they do so, the more they are validated "autonomously," that is to say, "theoretically," or "scientifically," and non-"ideologically." Secondly, even in areas where ideological orientation is strongest, as in social theories, and possibly in some philosophical theories, the criteria of validation, albeit more ample and flexible than those proposed by "traditional" and "orthodox" methodologists, cannot be extended indefinitely. Within the frame of an ideology a certain number of criteria, such as explanation, prediction, intelligibility, richness of content, practical applicability, etc.; can be laid down, and in this sense the ideology contributes to what we may understand as "validation." The ways in which these criteria function are "ideological," rather than "pragmatical," in so far as they are steps, no matter how tentative and rectifiable, which can always be justified within the framework of fulfilling the aims pursued. Although it might be claimed that the idea of a "pragmatic function" is, at bottom, "ideological," it is only to the extent that it becomes a criterion which may or may not work according to whether, and how, it is applied.

Second maneuver. Let us now consider the question from the other side.

I think that we can start from the following situation: theories and, in general, knowledge—but also systems of norms—are products of activities performed by human beings who live in communities and societies, and are conditioned by physical and biological factors—including the activity of adapting to, transforming, or "respecting" the world—and by economic relations of production, social structures, political institutions, national, tribal and racial forms of life, techniques, rites, etc. All of these are "realities"—or "real situations"—as well as results of reactions to realities, and reactions to these reactions. Within these "realities" we find systems of norms and forms of life, ideologies, world views, etc., which, as pointed out earlier, reflect the aforementioned realities or express a state of rebellion against them.

It can easily be seen how difficult it is to distinguish between what "reality" is and what "activity" is; some of the factors, elements or conditions just mentioned fall under both categories. It is common to call "realities" a set of given situations, and to call "activities" what men do, or plan to do, in order to confront these situations, including the decision not to confront them at all. It is more confusing to make a distinction of this sort than not to make one at all; in some sense, everything I have been talking about are "activities," even if they are called, redundantly, "real activities." Among them are theories of all kinds, both scientific and philosophical, and, in general, cognitive acts, or "knowledge."

From this point of view, both knowledge and theory can be characterized as an "activity." If we summarize all the activities alluded to earlier under the ambiguous but handy name of "practice," we shall be able to assert that theory, as a theoretical activity, is practice. In any case, knowledge is not a reflection of a transcendent, nor even of a transcendental world; it is not what may ensue from a purely "contemplative" and supposedly disinterested life, nor is it alien to the concrete conflicts and problems which human beings confront throughout their existence in this one and only world. There is no need for asserting that there is an impassable gap between theory and practice; indeed, if attention is paid to what may loosely be called "propulsing force," the "primacy of practice," specifically of "social practice," seems almost inevitable. Thus we must acknowledge the paramount importance of ideologies in so far as they express, conservatively, revolutionarily, or sceptically, the conditions of practice.

Does all this mean that nothing can any longer be called "theory"? Such a claim would go too far, for we might then take refuge in the "night where all cats are black" fallacy. The idea of a close relation between theory and practice emphasizes only that there are not two completely separated human worlds: the "contemplative" and the "active." If we persist in using the expression "two-worlds", we had better extend the two worlds over some kind of continuum; one of whose extremes we call "theory"—or knowledge, or science, or whatever—and the other "practice." The suggested primacy within this continuum indicates only that it is the locus where cognitive activities (among others) take place. None of this means that in order to validate a theory we must directly appeal to practice, or to some ideology, whether conformist or revolutionary, developed within the context of practice. We should rather appeal to a series of techniques—logical, epistemological, and, in general, rational—which have probably originated in practice, or even within the context of ideologies emerging from practice, but which have developed throughout history in a process of trial and error. Some of these techniques may, indeed, be chosen "for ideological reasons," and in this sense their adoption may be considered as a step towards the validation of theories, but the adoption of a technique of validation is not to be confused with validation procedures.

From this viewpoint, therefore, the validation of theories is not completely independent from "practice," but this does not mean that practice, or the "ensuing" ideologies, mechanically produce any techniques of validation. Futhermore, once the practical or theoreretico-practical ends are established, the use of these techniques develop, or tend to develop, autonomously. Thus, one can choose a logical system, in order to perform inferences, but the inferences performed are logical, not "practical," and still less "ideological." Because "interests" emerge within the context of Practice and are often made explicit by ideology, one can, for example, populate the need to formulate predictive statements. Prediction expresses then an "ideological preference." Yet, the rules of prediction are not ideological—they are rationally established rules. If the need to formulate predictive statements is denied, then no set of predictive rules serves any purpose. Thus, one may conclude that validation is not equivalent to prediction. But once prediction is rejected, and some other way of validation is chosen—as happened, for instance, with "understanding," in the sense of Verstehen—it will be necessary to find out what kind of rules of "understanding," as Verstehen, function adequately. The rules themselves will not be ideological—no matter how "ideologically" one chooses a method of "understanding" as opposed to a method of "prediction."

Thanks to these admittedly subtle distinctions, it may then be possible—without subscribing to any set of eternal truths, or eternal rules— to maintain the independence of research programmes against any oppressive ideologies. After all, ideologies should submit themselves to criticism. Once their function as (positive or negative) rationalizations of practice is acknowledged, the question may be raised whether or not there are "preferential reasons" which make certain ideologies more appealing than others. Among the least appealing ideologies are those which are deceptively, or perhaps wickedly, claimed not to be ideologies at all. In ideological matters, however, the best policy is still honesty.

To be sure, anything that is said about practice, including the statement that practice has a primacy over theory, is part and parcel of some theory. A kind of circle, not dissimilar to the so-called "hermeneutic circle" seems here inevitable. Now, although I do not believe that this circle can be eliminated altogether, its disturbing character can be made somewhat less annoying by noticing that the theory in question is not, strictly speaking, a scientific theory, but an "ideological theory," and one that honestly acknowledges its ideological and, as it were, "preferential" character.

Philosophy has been on the verge of dying a number of times, but each time she has surprised everyone, including those philosophers who declared her defunct, by producing yet another "tentacle" in her octopus-like structure. One of the tentacles of philosophy stretches out towards science to the point that there is a certain rather welcome continuity between philosophical theory and scientific theory. Perhaps another tentacle stretches towards ideology. Whether ideology and science can get together, at least to discuss their respective aims, through philosophy, is an open question which is philosophically worthwhile to pursue further.


  1. "Extratheoretical contexts" may impinge on theories in many different ways. To begin with, a distinction can be made between formation and development, which are diachronical, and structure, which is synchronical. Furthermore, whereas formation and development can be accounted for genetically, such is not necessarily the case with structure. Secondly, X may impinge upon T in various ways: it can be the (direct, indirect, concomitant) cause of T or the frame within which we are supposed to understand T. In the latter case we may call X a "context", C. Still, C can be a context of T in so far as the conceptual framework of C includes the conceptual framework of T, or in so far as appeal to C, or to parts of C, may allow us to understand T, or parts of T, etc. Needless to say, notions such as "conceptual framework" and "understand", are in need of clarification. I hope that the "context" within which these pages are written is reasonably clear.
Ferrater Mora, José. “On Practice.” American Philosophical Quarterly 13 (January 1976): 49-55.
back to The Essayist