Foundational Development Without Foundationalism
Wouter van Haaften and Ger Snik
University of Nijmegen, The Netherlands
Liberal education aims at the development of autonomous, critical thinking in diverse domains of rationality. Autonomous critical thinking is defined as the ability and the inclination to take an inquisitive attitude towards prevailing opinions. Liberal education is not primarily concerned with the transmission of contents of thought (what to think, what we think) but with the development of certain forms of thinking (how to think). Education should not consist in the passing-on of tradition-bound beliefs, but rather, foster the acquisition of general and fundamental principles which guide the responsible person in his or her thinking, judging and acting, and which may enable him or her to break out of the web of convention.
This ingredient of general and impartial principles can be found in many formulations of the idea of liberal education. It involves a distinction between "the present and the particular" and "the fundamental and the general,"1 between contents of thought that are held to be specific to particular cultures or traditions, and tradition-transcendent forms of thinking which make it possible for persons to take a position towards, and not be confined to, their cultural backgrounds. Such transcendent forms of thought have been discovered and made explicit in what are called critical traditions. In these traditions it is recognized that the validity of a judgment is not dependent on its source. Judgments are seen to be fallible in principle. Impartial criticism makes it possible for dominant ideas to be corrected.
Roughly, this is the Enlightenment concept of education.2 It has come under attack from several corners. Many authors have pointed out that in this picture the values of tradition and of the virtues have been severely underestimated. This criticism is correct in our view. The ideas and ideals of liberal education need to be broadened and adjusted in this regard. Peters3 and Scheffler4 have forcefully argued that liberal (moral) education is not incompatible with, but in fact, presupposes character education and the transmission of tradition.
A more threatening critique is that which denies the possibility of tradition-transcending principles.5 This critique strikes at the roots of the classical concept of liberal education.6 The central point is the argument against foundationalism. Liberal education seems to presuppose the foundationalist idea that there must be some absolute ground, guaranteeing the independence and certainty of at least some of our basic convictions and standpoints. Otherwise there could be no objective and general principles at all. If there is no such solid ground we shall have to accept that each group can maintain its own criteria of right and wrong.
It remains to be seen whether this reasoning is correct. The arguments against foundationalism must be acknowledged, but it might still be possible to justify general and fundamental principles in a non-foundationalist way. Some authors (for example, Scheffler) suggest that objective and impartial principles can be the result of "evolving traditions." They consider it possible to evaluate and justify such principles. The pivotal question then is, How and by what criteria this can be done without relapsing into foundationalism? In this paper we will follow up Scheffler’s suggestion and discuss some intriguing features of what may then be called foundational development.
Although various forms of foundationalism may be distinguished,7 the standard version resembles the "principium rationis sufficientis,"8 claiming that each proposition requires sufficient reasons for its justification; that these reasons in turn should be sufficiently grounded in more basic propositions; and that the most basic propositions or "ultimate origins" must be absolutely certain and unshakable, they must constitute a "fundamentum inconcussum."
There are two serious arguments against foundationalism: first, it is logically impossible to dig up ultimate origins in such a deductive procedure. If a principle is justified, it cannot (yet) be ultimate; and if it is ultimate, it cannot be justified.9 A second argument is that we cannot get out of our conceptual frameworks in order to compare them with the true and unconceptualized reality. We can have no such Archimedian point of reference for comparing ways of thinking.10
Surely these criticisms affect what is presupposed in liberal education, namely, that we can discover fundamental and general principles which transcend tradition-bound conceptualizations. If that thesis turns out to be illusory, we can only abandon the classical ideals of liberal education and embrace framework relativism. We then have to accept that people construct reality in different ways, depending on contingent and socially determined principles, each group using its own criteria of truth and justice, without our being able to refer to impartial meta-criteria to distinguish right principles from wrong ones.
As said above, we think that foundationalism is an untenable position. The critique is sound. We cannot escape from conceptualizing reality, so that there will be no "ultimate origins" for us to be found outside of our conceptual frameworks. Neither shall we be able to hit absolute ground in a "Letztbegründung durch Deduktion."11 However, this is not enough to force us to give up the idea of liberal education as based on tradition-transcending principles. Authors such as Scheffler and Siegel have argued that although there are no invariant and a-historical "fixed foundations," it may nevertheless be possible to find some general and impartial vantage ground. They have tried to pave a third way between foundationalism and relativism as a basis for liberal education towards critical thinking. Scheffler’s attempt was to re-coin the notion of objectivity.12 Siegel pleads for a "non-foundationalist absolutism."13
The central idea in both authors is that "standards of rationality can change over time." General principles of rationality have been discovered in "evolving traditions," covering diverse fields of rational argument, and they may well be further developed.14 At this moment we use the best principles we have, but we should not pretend that these are final and definitive. Even the most basic principles are fallible. They may need correction or substitution.
Now whether this is a viable route depends on one crucial question, How can changing standards of rationality themselves be justified as the right basis for judgment and evaluation? If such standards are not fixed, the question of why we should be entitled to apply the specific standards of this moment is even more pressing. If such standards are discovered in evolving traditions, how can it be argued that a later, or the most recently developed, set of principles is more appropriate than earlier ones? What are the "second order" criteria to be used here, and how can they be rationally assessed?
It is quite understandable that people going through a development in view are convinced that their later ways of thinking are better, particularly in case of fundamental changes in perspective. The same is certainly true for changing standards of rationality. The question is, however, whether and how such pretensions can be justified, avoiding the pitfalls of what has been called the "Münchhausen Trilemma" (circularity, endless regress, or an arbitrary stop in the argumentation)15 as well as the genetic fallacy. Scheffler and Siegel have suggested an interesting direction, but we feel that their solutions to this crucial question are not entirely satisfactory.16 Without claiming that we can solve the problems involved, we shall review what seem to be the most salient features of an approach along these lines.
If we accept the idea that fundamental principles of rationality are not fixed but developed in evolving traditions, and that, in principle, they can be further developed in the future, leading to new and qualitatively different principles, then we should first notice the implication that it will never be decidable in advance what are the best or the most appropriate principles in any field.
This means two things: In the first place, neither was it in the past, nor is it at present, nor will it be in the future, possible to foresee with certainty what principles may be accepted as valid later on in this development. We should therefore be modest about today’s attainments. We may feel sure about principles that we have for good reasons accepted, but at the same time we must recognize that we may -- again for good reasons -- have to give them up at a later time.
Second, the same is true at the meta-criterion level. We do have a global idea of what would make some standard or principle of rationality better than its forerunner, namely, that it somehow must lead to more appropriate forms of reasoning and judgment. But we shall have to recognize that the criterion to decide in which regard we may, at a later moment, be able to speak of "more appropriate," is not given beforehand. Thus, the idea of evolving fundamental principles of rationality implies that it is not possible at any moment to know in advance for certain what the final and definite correct principles are, or will be, nor is it possible to know what the decisive meta-criterion to determine this will be.
An admittedly extreme example can make this clear: the break-through of the insight that consistency is a basic principle of rationality that we should adopt in any form of reasoning, as compared to the situation in which this principle is/was not yet accepted. To us, this is an entirely evident principle, so much so that we may easily overlook that even this principle once had to be discovered and, in any event, that this principle also requires argumentation. We regard it as a basic element in what rational judgment can be, whatever the subject. We would find it absurd to abandon this principle. And what is more, whoever once recognizes the principle cannot even sensibly argue that it should not be held on to (once this stage of the development is reached, we also find that we cannot accept or reject principles without argument). Thus, it now must (logically!) be accepted. But whoever does not see this point can freely babble as he or she likes. Of course, we would think this improper, and to people at our stage of rational development, we can more or less explain why. We say: in the end we are right, not (s)he. But that is of no concern to him or her. (S)He is not going to be bothered by what, for us, has become an unshakable principle.
Now, this is, of course, an extreme case -- as if there were only two steps in the development of rationality, from irrational to rational. Certainly, all kinds of subtler reconstructions can be made, covering more aspects and more stages in rational development. But the example may help to make clear that the criterion for choosing the right "principles of rationality" is not fixed once and for all. In the same vein we shall have to accept that we do not have the last word in this matter. In principle, even the principle of consistency might once be superseded however counterintuitive this of course is to us.
Principles of rationality, the most prominent of which is consistency, can be seen as foundations in the full sense. They are constitutive for how we think and what we think. They create, and at the same time, delimit the field of our logical possibilities. Consistency is essential to what it means to be rational. To us it is a defining criterion of rationality. Once we have seen the relevance of this principle, we feel that it can only be abandoned on penalty of irrationality, that is, of leaving the domain of rationality as such. In that sense the principle is constitutive of that domain as conceptualized at a certain moment.
The principle of consistency is very general in that it pervades all areas of reasoning. But, as Scheffler indicates, there are different evolving traditions in which principles of rationality are discovered. General forms of thinking can also be developed in more specific fields, like morality, art, science, philosophy. In each of these domains (and in others) we may in their evolving traditions point to changing standards of rationality. In each of these areas these principles are constitutive of the relevant domain -- determining what is moral versus what is a-moral, scientific versus non-scientific, and so on.
We may differentiate here between foundational and non-foundational principles, in the following way. Foundational principles reflect what is taken as constitutive of the domain in question. They distinguish, for instance, the moral from the amoral (from what is not morally relevant at all). Within the domain of morality (as so delineated by the foundational principles) we use other, non-foundational principles (namely, moral principles) to distinguish the moral from the immoral. In the same way we may differentiate between the "principles of rationality" that are constitutive of what is rational, as opposed to what is irrational, and other, non-foundational principles (which may be seen as "rational principles" according to the foundational principles) to distinguish good arguments from bad ones. To clarify this point often a comparison is made with the game of chess where we can differentiate between rules that are constitutive of the game as such, distinguishing legitimate moves from moves that would not be part of the game and chess rules (within the so defined game) telling good moves from bad ones. But the main point here is that the foundational principles of rationality and of morality may have gone through fundamental changes in the past and can go through fundamental changes in the future.
Now, the idea (in what admittedly is our interpretation of Scheffler’s and Siegel’s writing) is that the developing principles of rationality in the various evolving traditions have led to, and may still lead to, more appropriate forms of reasoning. In our example, the development of rationality has led to the foundational principle of consistency, which forms a general and impartial basis in the evaluation of traditions. Similarly, the development of foundational principles of moral thinking has led to certain fundamental insights about the character of responsible moral judgment in terms of which such diverse issues as political systems and styles of education can be criticized. We need not hypothesize that children in their development rehearse the history of mankind to acknowledge that they will have to learn some of these constitutive principles of sound reasoning in such domains. But maybe they sometimes have to pass through comparable stages in their individual development in order to attain the intended level.17
In whatever way the evolving traditions or foundational developments in the different domains are reconstructed, the pivotal issue remains how foundational principles can be justified, or more precisely, how new foundational principles can be justified as more appropriate than what have come to be considered superseded ones in view of the deep qualitative differences between such (sets of) principles. Although we cannot into this question extensively,18 we would like to make some observations about what is involved. We concentrate on the justification of the principle of consistency.
Apel’s transcendental strategy19 is particularly apt here, because this principle of rationality is demonstrably presupposed in any serious form of reasoning about whatever subject, and in favor of whatever position. The strategy consists in pointing out that any potential opponent must already have accepted this principle as soon as he or she might start arguing against it. In opposition to Kant, Apel calls this line of argument transcendental-pragmatic, because it consists in clarifying what implicitly has been accepted in the praxis of any argumentation. Apel’s point is, that we have a foundation here, but it is not a "Letztbegründung durch Deduktion" but a "Letztbegründung durch Reflektion" -- which means that the conclusion can be reached by reflection on what is presupposed in the argumentative praxis without being caught in a deduction, and without having to step outside of the conceptual frameworks involved. In other words, we have a foundation while avoiding the problems of foundationalism. Moreover, as Apel rightly stresses, there is no circularity involved here because the strategy does not hinge upon a deductive argument.
Coming back to our earlier example, this type of reasoning can also be used as a transcendental-genetic argument in defense of the later stage in the development or genesis of rationality in which the principle of consistency has come to be acknowledged. Once this stage of an evolving tradition has been reached, it can be justified as more appropriate than the stage in which it was not yet accepted. This development has led to an adequate form of rationality and to a better recognition of what it is to be rational. A similar line of argumentation is possible in defense of certain fundamental principles in the more specific domains.20 Notice that the reflective (non-deductive) character of this argumentative strategy also enables us to avoid the naturalistic and the genetic fallacy. The evaluative development claim that the more recent stage is the most appropriate is not deduced from its factual character, nor is it based on the fact that it is the later stage.
On the other hand, the possibility of this form of justification is bound to the later stage in this particular development. It is understandable and acceptable for the person who has reached the new stage. Of course, the argument is valid in principle for the person in the foregoing stage as well, but he would not be able to recognize it (otherwise he would have moved to the later stage). And, as remarked before, in principle we may develop towards a later stage in which we are for good reasons going to prefer still other principles that might replace or subsume the now accepted standard of rationality. But what this theoretical possibility amounts to we cannot foresee (otherwise we would have moved to that later stage).
The latter, however, should not be regarded as a shortcoming; rather, we suggest, that this, itself, is one of the intriguing features of foundational development. It is in line with what was observed before, namely, that we can only justify principles in the best way possible for us now under simultaneous recognition that this may turn out not to be the best way possible in the long run.
In connection with the foregoing we would like to call attention to one further curious feature that seems to us typical of (at least some forms of) foundational development. It concerns the status of the meta-criterion appealed to in the evaluation of foundational development stages when we retrospectively reconstruct this type of development towards what we now regard as crucial for the notion of rationality as such. We can now see that the meta-criterion we use in judging the later stage as more appropriate than the prior one(s) coincides with the principle of rationality that is characteristic of that same later stage. The newly accepted principle appears to also function as our meta-criterion in the comparison. What we in our stage of rational development have come to see as the best possible principle of rationality is now regarded as the best available meta-criterion as well. Or the other way around: only the new stage yields the meta-criterion in terms of which the rival stages are evaluated.
If this is correct, the status of the meta-criterion is rather peculiar in that it is not something common to both stages compared (as Scheffler21 demands for the evaluation of different paradigms). Nor is it totally independent in the sense of a point of reference outside of any conceptual framework (as, according to its critics, foundationalism would require), or independent in the sense of an impartial third party (for instance, some higher-level coordinating framework).
Without being able to argue this here, we surmise that this is also true in case more than the two stages of our example can be reconstructively distinguished.22 The general, first person structure of this type of justification then would be the following. When at any moment S(n) represents our highest reached stage, then, in the light of the foregoing and depending on how many qualitatively different stages are retrospectively discerned, all stages S(n-m) [with m > 0] may be judged less adequate; while no stages S(n+m) can be judged. Thus we may conclude that S(n) is the most appropriate and the most adequately justifiable stage at this moment while realizing that we are participating in an evolving tradition.
Does this peculiar status of the meta-criterion make the evaluation circular? The answer seems to be, Yes and no. Let us very briefly mention three points. First, the argument should not be reversed. The thesis is not that any foundational development will automatically lead to a better stage. What we wanted to show are some of the peculiarities that seem to be involved if a justification of an evaluative foundational development claim is attempted. Second, in that case, the later stage need not necessarily be evaluated as better than its forerunner. It may be so evaluated. Perhaps there are also foundational changes that cannot be argued to be an improvement. However, this does not seem to be the case if a transcendental argument can be appealed to as in the example given in the section "Undecidability in Advance." Here the counter-position would, indeed, undermine itself. Third, it must be admitted that there is a real kind of circularity involved in the argumentation towards the person in the earlier stage. For the justification of S(n) does presuppose the specific meta-criterion that is only in S(n) accepted. We may have to swallow this as an awkward, but at the same time, inevitable and (perhaps also) clarifying characteristic of foundational development. It is, after all, the development of really new and unknown perspectives.
Foundationalism is an untenable position. However, we do not need to reject the idea of liberal education with its typical presupposition that there can be tradition-transcending principles. Although, as Scheffler has pointed out, there are no fixed foundations because even the most basic principles must be regarded as fallible in principle, it may be acknowledged that in evolving traditions ever more appropriate forms of critical thinking and principles of rationality have been, and still may be, discovered.
In this paper, we have elaborated some aspects of what this may involve. The pivotal question is, how we can rationally evaluate principles of rationality that are developed in such evolving traditions? This problem is particularly intriguing for foundational principles, as distinguished from other principles by their being constitutive of the domain of conceptualization in question. We have concentrated on the argumentation in favor of the principle of consistency, which may be seen as the basic principle constitutive of rationality as such, and we have observed that a transcendental-pragmatic or transcendental-genetic strategy is suitable here. Consistency was our example, but we suggest that a transcendental strategy is applicable to some of the most fundamental principles in more content-specific domains, such as morality as well. The prominent advantage of this type of justificatory strategy is that the genetic fallacy and some other problems like endless regress and, in certain respects, circularity can be avoided.
Circularity remains, however, in the justification of the new foundational principles towards persons who have not yet reached the new developmental stage and its characteristic insights. This has to do with the peculiar situation we find ourselves in when justifying evaluative claims to foundational development. The stages we went through were not predictable in advance. They can only be reconstructed in retrospect. Neither was the meta-criterion for the comparative evaluation of the stages fixed beforehand. It was the outcome of our most recent stage. We think, however, that this may be inherent to foundational development.
For a response
to this essay, see Airaksinen.
1. See Charles Bailey, Beyond the Present and the Particular: A Theory of Liberal Education (London: Routledge & Kegan Paul, 1984).
2. See Immanuel Kant, Beantwortung der Frage: Was ist Aufkl rung? (1783, A481) in Werke in zehn Bänden Hrsg. W. Weischedel (Darmstadt: Wissenshaftliche Buchgesellschaft, 1983), Bnd. g, 53.
3. Richard S. Peters, Moral Development and Moral Education (London: Unwin, 1981).
4. Israel Scheffler, In Praise of the Cognitive Emotions and other Essays in the Philosophy of Education (New York/London: Routledge, 1991), 97-100.
5. For example, Richard Rorty, Philosophy and the Mirror of Nature (Oxford: Basil Blackwell, 1980).
6. Liberal education has also been defended from a more contextualistic perspective, but this position seems confused, cf. Wouter van Haaften and Ger Snik, "Allgemeinbildung als Entprovinzialisierung des Denkens," in Pedagogik und Pluralismus, ed. F. Heyting and H.-E. Tenorth (Weinheim: Deutscher Studien Verlag, 1994), 65-82.
7. See William Alston, Epistemic Justification (Ithaca/London: Cornell University Press, 1989), 19-56.
8. See Gottfried Wilhelm Leibniz, Monadologie (1714), Hrsg. H. Glockner (Stuttgart: Reclam, 1963), sect. 32.
9. Hans Albert, Traktat Über kritische Vernunft (Tübingen: Mohr, 1980).
10. Rorty, Philosophy and the Mirror of Nature.
11. Karl-Otto Apel, "The Problem of Philosophical Foundations in the Light of a Transcendental Pragmatics of Language," in After Philosophy: End or Transformation? ed. K. Baynes, J. Bohman and Th. McCarthy (Cambridge, Mass./London: MIT Press, 1987), 250-90.
12. Israel Scheffler, Science and Subjectivity (Indianapolis/New York: Bobbs-Merrill, 1967); cf. Alven Neiman and Harvey Siegel, "Objectivity and Rationality in Epistemology and Education: Scheffler's Middle Road," Synthese 94 (1993): 55-83.
13. Harvey Siegel, Relativism Refuted: A Critique of Contemporary Epistemological Relativism (Dordrecht: Reidel, 1987), 165.
14. Israel Scheffler, Reason and Teaching (London: Routledge & Kegan Paul, 1973), 78f.; Harvey Siegel, Educating Reason: Rationality, Critical Thinking, and Education (New York/London: Routledge, 1988), 59, 134f.
15. See Albert, Traktat Über kritische Vernunft.
16. Wouter van Haaften and Ger Snik, "Critical Thinking and Foundational Development," Studies in Philosophy and Education (1995, in press).
17. For example, Lawrence Kohlberg, Essays of Moral Development, vol. I: The Philosophy of Moral Development (San Francisco: Harper & Row, 1981) for the domain of morality.
18. See Wouter van Haaften, "The Justification of Conceptual Development Claims," Journal of Philosophy of Education 24 (1990): 51-69; Wouter van Haaften, "Conceptual Development and Relativism: Reply to Siegel," Journal of Philosophy of Education 27 (1993): 87-100; van Haaften and Snik, "Critical Thinking and Foundational Development."
19. Apel, "The Problem of Philosophical Foundations in the Light of a Transcendental Pragmatics of Language"; Karl-Otto Apel, "Fallibilismus, Konsenstheorie der Wahrheit und Letztbegründung," in Philosophie und Begründung, ed. Forum für Philosophie Bad Homburg (Frankfurt am Main: Suhrkamp, 1987), 116-211.
20. For instance in morality, see K. O. Apel, Diskurs und Verantwortung. Das Problem des Übergangs zur Postkonventionellen Moral (Frankfurt am Main: Suhrkamp, 1988); Ger Snik and Wouter van Haaften, "Moral Autonomy as an Aim of Education," in Morality, Worldview, and Law, ed. A.W. Musschenga, B. Voorzanger and A. Soeteman (Assen/Maastricht: Van Gorcum, 1992).
21. Scheffler, Science and Subjectivity, 82-83.
22. van Haaften and Snik, "Critical Thinking and Foundational Development."