1 / INTRODUCTION

1.1. ON THE SITUATION OF THE PROJECT OF THE ENLIGHTENMENT

The illustrious project of the Enlightenment, once the heart of European philosophy, has, in our day, a bad reputation. Those who commit themselves to the Enlightenment quickly meet with distanced reactions. For don’t we owe the destruction of nature to the Enlightenment? Isn’t the Enlightenment, with its demand to bring societal process under the control of rational planning, the project that is responsible for the societal systems of coercion? And isn’t the Enlightenment the project that set free processes of technical innovation, the consequences of which we cannot even begin to estimate? And isn’t a reason at work behind all of these processes that the Enlightenment thinkers promised would bring a release from our self-imposed immaturity?

Even if this diagnosis paints too bleak a picture, the discourse of modernity appears to be characterized by the development of aporias that overbid one another whereby each encompasses the other. These have condensed into an opaque, interwoven nexus, constituting a general crisis of the Enlightenment, which has recently increasingly been mistaken for the end of the Enlightenment. However, such a bleak view of the Enlightenment is hardly one that has only been offered by our contemporaries. Long before it had become the fashion to skirt problems by attempting to withdraw into a new era with the help of the prefix “post,” Adorno, for example, attempted to depict a violence of the Enlightenment impulse, which he had traced back to the very capillaries of conceptuality itself. This sort of attempt, for which Adorno is initially only to serve as an example, has always been connected with the problem that the diagnosis of the violent structure of the Enlightenment remained dependent on a medium that, for its part, belonged on the continuum of the Enlightenment, and thus could not be exempted from the scope of the critique. Thus, any self-critical process like this also entails the danger that the critique will rob itself of its own normative foundations. In this aporetic structure, however, only one motif reappears, albeit in a radicalized form, that was already connected to the earliest formulations of the concept of the modern, as found initially in Hegel. The process known as the Enlightenment has been subject to increasingly radical critique since its beginnings, and it has attempted, even in the development of aporetic structures, to strike a learning pose to critique. Precalculating the costs of the dominant concept of reason has always been an element of the Enlightenment.

In the wake of the end of the Enlightenment and the modern, the context needed for a self-critically developing Enlightenment threatens to unravel. Reaching back to the example of Nietzsche’s generalized—heuristically explosive, but methodologically implosive—suspicion¹ that reason is intrinsically empty and thus only an instrument of the will to power, demasking critique has been established as the form for the fundamental critique of the program and process of the Enlightenment. There are essentially four possible reactions characterizing how this type of critique has been taken up by theories of the Enlightenment.

If the demands of such a demasking critique are accepted, there remain possibilities: First, defensive immunization is possible; this can be done, for example, by ratcheting down the theoretical claims. Second, it is possible to integrate the perspective of demasking critique into the Enlightenment concept; here, however, the theoretical attempts become increasingly aporetic. If the demands of the critique are rejected, then there are further possibilities: either third, a basis must be developed that grounds the critique of the critique and that, for its part, cannot be (completely) encompassed by the critique that has been rejected, or fourth, the critique must be reconstructed within broadened parameters that are able to maintain the connection with the project of the Enlightenment. Irrespective of whether one is of the view that each of these four theoretically possible reactions leads to intelligible viewpoints, in any case critique would generate a development that, as a whole, can be understood precisely as Enlightenment. In such a perspective it appears to be irrelevant whether the critique develops within the parameters of Enlightenment theories or outside the parameters of such theories.

In the 1980s, however, it appears that this scenario fundamentally changes. It no longer presents the situation of the project of the Enlightenment, but only one perspective of the situation: for many, the case recently brought against the Enlightenment has, in its final appeal, now been closed. Others think that it is a process that is no longer relevant because foundations for such a process, in their immanent violence, have become a clear sign of the guilt of the accused.² Hereby, the perspectives centered around the Enlightenment appear to have become mere variations among other possible perspectives; and with this change in the point of view, images shift abruptly between perspectives of the Enlightenment as an incomplete project (Habermas), as a grand récit (Lyotard), as a catastrophic dynamic³ (Baudrillard), as a historical program with exaggerated foundationalist demands (Rorty), or as a bankrupt host of old European thought (Luhmann). Today it no longer appears possible to once again reorient the kaleidoscope of perspectives, redirecting them to the Enlightenment, simply by understanding the Enlightenment as a learning process and interpreting all critiques of this learning process, for their part, as a part of the learning process. For a more basic cause of the diagnosed decentering of the Enlightenment is the inability of the program to learn. Radical critics of the Enlightenment find this above all at two levels, which can be labeled “theory and violence” and “theory and self-reference.”

a. The motif showing the project of the Enlightenment to entail an implicit moment of violence is widespread, irrespective of the cleft between critics and defenders of the Enlightenment. The diagnoses motivating the basic skepticism toward or the turn away from the project of the Enlightenment range from an emphasis on the “will to power” to an emphasis on the preponderance of the general over the individual,⁴ the claim of ostensive metadiscourse to hegemony,⁵ and the structural power of the code.⁶ If, however, on the basis of such diagnoses, the medium of the Enlightenment is conceived as intrinsically connected to violence, then the (refined or sublimated) continuance of the project of the Enlightenment can only be feared; so theoretical work has to be joined to powers external to the Enlightenment. To this end every topos that is opposed to the alleged structures of the Enlightenment is offered: in opposition to the terror of the concept of unity, the other of reason stands as the “nonidentical,”⁷ the “incommensurable,” the “heterogeneous,” or the degraded;⁸ and art, ecstasy, deconstruction, and catastrophe are called to its aid.

b. The second, in a certain sense unavoidable, reproach to the tradition of the Enlightenment is that of self-referentiality, which in normative and legitimating contexts is sharpened to an allegation that it involves a circular self-presupposition. Theories of the Enlightenment appear to be dependent on presupposing something that they are precisely intending to portray. According to this diagnosis, the specific scientific knowledge of the Enlightenment “cannot know and make known that it is the true knowledge without resorting to the other, narrative, kind of knowledge, which from its point of view is no knowledge at all. Without such recourse it would be in the position of presupposing its own validity and would be stooping to what it condemns: begging the question, proceeding on prejudice. But does it not fall into the same trap by using narrative as its authority?”⁹

Aside from the implied crude identification of “scientific knowledge” and “true knowledge,” which is foreign to most (modern) philosophies of science, and aside from the fact that the diagnosis that the Enlightenment is involved in a self-presupposition is among the recurring forms of critique of an enlightened Enlightenment, the fact that something is a self-presupposition is a problem only in a certain sense, a sense that comes into purview if we turn our attention to the ability to subject theoretical constructions to criticism. If the self-presupposition is no longer able to be criticized in an explicative or normative context without at the same time accepting the theory in which this structure is generated, then a theory program exists to which there are apparently no alternatives as long as one adheres to the criticism. This situation can be described more precisely as follows: if the presuppositions of a theory are joined with the explanatory claim of this theory such that the relationship between the presupposition and the claim can no longer be criticized without presupposing that this relationship is valid, then the theory forms more or less its own environment and loses every external reference. Even criticism of the theory would be nothing more than an “externalized” self-reflection of this theory. Such a parthenogenic concept of reason, however, that indiscriminately includes the critique of itself, and that because of its totalizing self-referentiality must develop everything out of itself, belongs in essence among the fossils of the history of philosophy; only with great effort could it be restored as an adversary to be taken seriously today. In connection with problems of self-presupposition, what “remains is a formal peculiarity that one finds in the concept of rationality and perhaps nowhere else: The concept of rationality must be subsumed autologically, must be formed rationally.”¹⁰ This, however, by no means prevents concepts of rationality from being established in different ways, and thereby the avoidance of structures that solidify in self-reference.

The theoretical approaches that, in the marketplace of ideas, not long ago found (and still find) their customers under the marketing label “postmodern” and that in many cases, with their critique of the violence and self-reference of the Enlightenment, refer more to a caricature of the current conceptions of the Enlightenment¹¹ than to systematic problems, require this distorted background in order to plausibly show that they have somehow escaped the continuum of the Enlightenment learning process. But even the fact that theories that are included as part of the radical critique of the Enlightenment by no means stand in a consistent relationship to the Enlightenment, but oscillate between the “revision of the modern” and the discrediting of the medium of the modern—discourse¹²—does not restore a connection between making suggestions and critique, which is prescribed in the development of a robust concept of the Enlightenment. For precisely this concept—understood as a narrative of legitimation—has not only lost its credibility,¹³ but, by being “demasked” as a narrative, has forfeited the status that secured the conditions for its immanent critique. If one presupposes, like Lyotard, that scientific knowledge “requires that one language game, denotation, be retained and all others excluded”¹⁴ and that it legitimizes itself only by recourse to narrative forms of knowledge,¹⁵ which for their part are “never subject to argumentation or proof,”¹⁶ then the validity of the critique of science and thus of the Enlightenment would be disconnected from argumentation and would thereby leave the realm of reason-supported intersubjective argumentation. A critique conceived in this way has to face up to the charge that it itself is nothing more than a metanarrative. To fail to see a problem even in this is to begin the transformation from philosophy to entertainment.

The question that arises in the face of the above considerations is how, in this situation, an examination of forms of the radical critique of reason, the subject, and the Enlightenment can be established without degenerating into a witless exchange of opinions. It appears to me that what is necessary in order to do this is, first of all, a hard breach:

1. Some of the postmodern approaches repeat mistakes of the radical critique of the Enlightenment that have long been identified, and they presume rationality concepts and theoretical claims—undoubtedly also for reasons of rhetorical enhancement—that are seldom defended today in the ways that would suit the postmodern critique.¹⁷ Habermas, who is stylized as a hydra of rationality theory,¹⁸ years ago wrote:

Just as it always has, philosophy understands itself as the defender [of rationality] in the sense of the claim of reason endogenous to our form of life. In its work, however, it prefers a combination of strong propositions with weak status claims; so little is this totalitarian, that there is no call for a totalizing critique of reason against it.¹⁹

2. Such a radical critique of reason is only able to be brought to its recipients at the cost of performative contradictions, because, if it is to be an intelligible and correct analysis, for its part, it must raise claims of validity that it purports to demask as violent impertinencies:²⁰

Is legitimacy to be found in consensus obtained through discussion, as Jürgen Habermas thinks? Such consensus does violence to the heterogeneity of language games.²¹

In light of the grave problems associated with this type of critique, we can hardly expect the dispute with it to be the linchpin of a philosophical theory that is concerned with a broadened concept of rationality. Rather, the radical critique of reason constitutes only a part of the background against which this attempt will be made, whereas precisely those theories that are the object of postmodern critique constitute the real—and to put it clearly, the only—sensible links.

3. It is certainly also clear that, against the background of the evident weaknesses of the postmodern critique, it is not possible simply to speak of the affirmation of the Enlightenment. Rather, the attempt to break from the continuum of the learning process must be taken seriously as a symptom that even the most developed conceptions of the project of the Enlightenment may display grave weaknesses. What may be gained from the reception of the postmodern variations of the critique of reason and the Enlightenment is not a foundation for critically examining the theories of rationality or modernity. Rather, it is an increased sensitivity to the costs of those theories that today represent the tradition of the Enlightenment. If “postmodern knowledge” were limited to this sensibilizing function, then the following formulation from Lyotard could nearly be accepted without qualification: “Postmodern knowledge is not simply a tool of the authorities; it refines our sensitivity to differences and reinforces our ability to tolerate the incommensurable.”²² Luhmann, who here should be taken as a representative of the functionalist distance to the emphatic project of the Enlightenment, also accentuates this sensitivity to differences. Developing his thought against the background of the concept of difference, Luhmann shares Lyotard’s diagnosis that there is no métarécit, indeed “because there are no external observers.” As he continues, “Whenever we use communication [!]—and how could it be otherwise—we are already operating within society.”²³ But in Luhmann’s view as well, the benefits of the postmodern are limited to the diagnostic:

The proclamation of the “postmodern” has at least one virtue. It has clarified that contemporary society has lost faith in the correctness of its self-description[s]. . . . They, too, have become contingent. . . .²⁴ What is important here is not the emancipation of reason but emancipation from reason. This emancipation need not be anticipated; it has already happened.²⁵

However, Luhmann’s distance, especially from the utopian moments of the Enlightenment process, does not go so far as to abrogate cooperation with the project of the Enlightenment: For “even this ‘bifurcation’ [of the process of modernity in spontaneous decomposition and utopian renewal] can be understood as unity, namely as applying the process of learning to the little understood phenomenon of modern society.”²⁶

A similarly supported view has by now also become widespread among prominent theoreticians of the postmodern, namely that “postmodernity is not a new age, it is the re-writing of some features modernity had tried or pretended to gain, particularly in founding its legitimation upon the purpose of the general emancipation of mankind. But such a re-writing, as has already been said, was for a long time active in modernity itself.”²⁷

Regardless of what comes into purview from a postmodern perspective, the most developed theories of the Enlightenment and of rationality entail problems that provide occasion enough for dissatisfaction with mere justification. These theories deserve a fundamental and detailed critique precisely because they have taken it on themselves to develop robust and criticizable basic principles of critical social-theoretical reflection that go beyond the best-known, but most effective, forms of totalizing critique. Insofar as the bleak picture sketched out at the beginning presents only a murky variation of the view of a situation in which, if we remove the postmodern visors, we can find a link for continuing the project of the Enlightenment.

1.2. WHAT IS ENLIGHTENMENT?

The question “What is Enlightenment?” appears antiquated in a climate of declining theoretical weightiness and inflationary declarations of its obituary. Given the lack of prospects of theories that believe they are able to do without this concept, I believe there are nevertheless no alternatives to proposing a clear, contoured concept of the Enlightenment and to checking whether this is able to hold out against critical objections that are developed in the context of a critique of reason.

Agreement about the meaning of the concept of Enlightenment is limited essentially to the characterization of it as a societal process. However, far-reaching differences become apparent as soon as an assessment of this process is at issue. Besides the pessimistic assessment of postmodern concepts sketched out above, two different basic conceptions can be differentiated.

When using the word enlightenment in a rather descriptive way, it characterizes a process that has developed (contingently) within the parameters of a certain culture. In contrast with this distanced observer perspective, concepts that normatively charge the idea assume an internal perspective, which is a result of their participatory partiality. Odo Marquard, for example, whom I would like to mention as an exemplary representative of the distanced description of the process, characterizes the Enlightenment as the development of a de-emotionalized, epistemic stance: “The Enlightenment is the tradition of the routinized courage to an unagitated sobriety.”²⁸

“Sobriety,” as the stance that arises in the Enlightenment process, which is to secure both a distance from possible goals of action as well as a distance from emotional attachment to them, simultaneously has the function of easing the weight of the goals of the tradition, which has been passed on, and the objective of making these goals themselves the object of passionless reflection: “In short, Enlightenment sobriety, that is, the usance of modernity, is the exoneration of the absolute.”²⁹ This anti-utopian impulse is sharpened in Luhmann’s descriptive concept of the Enlightenment. Luhmann radicalizes the process of exonerating the absolute, and it becomes a process of generating contingency. With the internal differentiation of modern societies, the impossibility of privileged perspectives becomes increasingly prominent; for anything that can be seen can also be seen differently. The emancipatory impulse, which is intrinsic to the Enlightenment until then, loses the character of a goal that is to be sought and is transformed into a process that unavoidably arises as an epiphenomenon of technical modernization.

Technology, in its broader sense, is functional simplification, that is, a form of the reduction of complexity that can be constructed and realized even though the world and the society where this takes place is unknown. It is self-assessing. The emancipation of individuals, even irrational individuals, is an unavoidable side effect of this technologizing.³⁰

However, as already said, here this is no longer a matter of “the emancipation of reason but emancipation from reason. This emancipation need not be anticipated; it has already happened.”³¹ In this perspective, too, the Enlightenment remains a learning process. From the perspective of the observer, however, it is one in which, as binding descriptions of the world are unrelentingly deconstructed, any knowledge thought to be secure or any goal thought to be justified is again shown to be contingent. But within this learning process, strong intuitions and normative ideas are no longer stable foundations of the Enlightenment concept; they—like everything else—are subject to observations from standpoints that can be freely chosen, and they consequently lose their status as foundational points of orientation. Orientation on any of the kernel ideas on which the observations can be centered is only one of many possibilities. Privileged standpoints, which could function as starting points for theories of the Enlightenment, can only be taken up at the price of an antecedent narrowing of one’s perspective. When asked about the problem of a lack of a certain standpoint of his theory or an intuition that stood behind it, in a 1985 interview, Luhmann answered:

I would say that the structure of my theory is rather open for very diverse areas of society, rather open to both the positive and the negative sides of modern society. Habermas’ problem is that his thinking is organized in reference to a moral duty of social theory. That lends this theory indeed a special attraction, but at the same time also indicates it boundaries. From the point of view of such a position, for example, one cannot see the money economy, law, or political machinery critically or can only view these with negative, critical depreciation.³²

Luhmann further specifies the openness of his theoretical perspective while denigrating Habermas’s concept, which is centered on the intuition of undistorted intersubjectivity and is committed to a “utopian perspective of reconciliation and freedom.”³³ He does so by indicating the different starting point of his theory and its primarily descriptive tasks:

I, for example, think it is more fruitful, not to begin theories with unity, but with difference, and not to allow them to end in unity (in the sense of reconciliation), but, how should I say this, with a better difference. . . . As a stance for research, the idea behind this is that one could do it all much better than any of the interpretations so far. So, one must withhold moral judgment or critique until one sees how it is possible to think about modern society.³⁴

Understood against this background, the Enlightenment is not a committed and normatively grounded intervention. Rather, it is an observation that abstains from moral judgments, descriptively accompanying the societal processes, which generate ever-new configurations. It is an observation that is prepared to reflect the reduction of complexity that is intrinsic to any observation and, by introducing new differentiations, to construct it more complexly.

Richard Rorty, to introduce a final distinguished advocate of a primarily descriptive view of the Enlightenment—starting from a pragmatic deconstruction of central values of Western rationalism (e.g., “objective truth”)—also maintains a distance from the normative claims of the Enlightenment and describes it as a historically contingent process. He is convinced that in the wake of a pragmatic deconstruction, philosophy ought to leave unproductive theoretical debates behind and accept the view that philosophical beliefs are not the cause of societal processes, and that, beyond their contextual usefulness, theories cannot be shown to be true. Against this background, Rorty pleads for a “light-minded aestheticism . . . toward traditional philosophical questions,” which should serve the same end “as does the encouragement of light-mindedness about traditional theological topics”:³⁵

Like the rise of large market economies, the increase in literacy, the proliferation of artistic genres, and the insouciant pluralism of contemporary culture, such philosophical superficiality and light-mindedness helps along the disenchantment of the world. It helps make the world’s inhabitants more pragmatic, more tolerant, more liberal, more receptive to the appeal of instrumental rationality.³⁶

Only if we put aside the pathos of philosophy and accede to a perspective of purposively rational reflection—in accord with which we evaluate the use value of theories, institutions, and social practices, oriented toward praxis—will we escape the danger of becoming entangled in irrelevant academic questions. What is relevant—leaving aside entertainment value—are in any case the practical consequences of theoretical proposals, not their abstract claims to validity.

For the pragmatist, by contrast, “knowledge” is, like “truth,” simply a compliment paid to the beliefs which we think so well justified that, for the moment, further justification is not needed.³⁷

As a result of the radical cut in the breadth of validity claims to concrete cases, both a concept of universal transcultural rationality³⁸ and a concept of an encompassing project of the Enlightenment have become superfluous. In the final analysis, only concrete individuals in their respective life situations can decide whether the societal condition in which they live offers suitable tools for satisfying their needs or not. Here it seems to me to be problematic that, after the abandonment of the philosophical critique of social practices, the theory maintains only an elevated stance with view to liberal society, which, for its part, represents the milieu of evolutionary change and, against the background of our needs, is either retained or not. Individuals can learn from failure, “but they will not learn a philosophical truth, any more than they will learn a religious one. They will simply get some hints about what to watch out for when setting up their next experiment.”³⁹

Beyond both the cool distance of Rorty and the sobriety of Marquard, as well as Luhmann’s theory, which is oriented on an exact description of societal processes, in the debate that centers around the concept of the Enlightenment there are normatively constructed theories, which, while not basing their concept of the Enlightenment on transhistorical or “objective” norms, do, however, argue with norms that assume a prominent position within the theoretical framework. The works of Mittelstraß and Habermas provide exemplary cases of such views.

Mittelstraß characterizes the Enlightenment as a process of confronting the existing conditions with normative criteria. In the examination of the culturalist concept of rationality, as it has been developed by Kambartel in connection with the concept of a form of life found in Wittgenstein,⁴⁰ he emphasizes the intervening character of critical reason that is engaged and that realizes itself in the process of the Enlightenment.

In contrast to the specific affirmative character of a culture of reason, in accordance with Kambartel, I maintain that reason can in principle only become real as a form of resistance, as a critique of existing circumstances and as a projection of different circumstances, as enlightenment. The differentiation between rational and irrational circumstances, which we cannot get around if our talk of reason is not to become empty and vacuous, is a distinction that we bring into the world.⁴¹

Mittelstraß worries that by redirecting the Enlightenment project toward descriptive tasks and by focusing on the implications of the rationality for societal reality, the critical distance of the Enlightenment to the existent reality and to what is described will be lost. For “if the culture of reason as a universal culture is itself normative, as it is now called, precisely then is the appeal to its reality affirmative.”⁴² Consequently, he insists “that reason cannot be extracted from the circumstances, but rather must always be asserted against them.”⁴³ As the form of the Enlightenment, critique remains dependent on the basic principles of a normative concept of reason, for “reason is an idea. And ideas are not defined. They are rather determined in reference to ideal demands, postulates.”⁴⁴ The “Enlightenment idea of reason” that Mittelstraß is concerned with counters the existing conditions in a culture with a judgment about its existing practices that can be reformulated into demands on the participating subjects. “Reason is not an attribute of bourgeois culture or some other culture, but a form of evaluation of these cultures.”⁴⁵

Against this background, the Enlightenment is a process that—based on the suggestions of individuals⁴⁶—aids in the development of rational conditions, and these conditions, which are to be created, must be able to achieve general acceptance. They must “prove themselves to be developed in comparison to the existing conditions and as suited to serve as the basis for future justified developments.”⁴⁷

Habermas surely is the most prominent advocate of a normatively constructed theory of the Enlightenment. His Enlightenment perspective is fully connected to the intersubjective sphere of linguistic communication, and in two senses: on the one hand, linguistic communication is the medium in which the process of the Enlightenment is carried out; on the other hand, the conditions that must be achieved in order for understanding to occur simultaneously serve as the basis for the normative criteria on which the Enlightenment, as a learning process, can orient itself. Socioculturally, the process of the Enlightenment is carried out through a process in which a medium of reflection is socially implemented that makes possible the intersubjective examination of validity claims that are raised in linguistic action. Here, at the level of collective imagination, the process sets out from the “linguistification of the sacred.”⁴⁸ In the course of this, the traditional and unproblematic presuppositions lose their self-evident status and become the objects of critical debate, the objects of argumentation. The reflexive problematizing of traditional models of integration and orientation at the collective level corresponds to a process of decentralizing one’s egocentric understanding of the world at the individual level.⁴⁹ In this process, discursive socialization yields to an ability to change perspectives and exchange roles.⁵⁰ A linguistic understanding or even more specifically the “communicative sociation”⁵¹ of the individual is essential for both forms of learning that are basic to the Enlightenment; and the differentiating of the further development of linguistic interaction into forms of communicative action that are solely committed to examining validity claims that are made, i.e., to discourses, is at the same time also the moving force of a dynamic of rationalization that is characteristic of the Enlightenment process.

In comparison to Mittelstraß’s view that the rational stands in a constitutive opposition to existing conditions, the normative aspects of the Habermasian Enlightenment concept are subtly interwoven into the existing conditions; for noncoercive discourse, as the medium for the-matizing and critical examination, is an idealized construction whose basis—action oriented toward understanding—underlies linguistic interaction even in its most contorted forms. Enlightenment thus does not appear at an abstract distance from what exists; rather, as an advocate of noncoercive understanding, it can assure itself of the normative implications even of everyday communication by pointing to the conditions that make understanding possible at all. A communicative rationality that unfolds in the process of the Enlightenment can, supported by the analysis of the conditions of communicative action, rely on the fact that, whatever the accidental limitations the process of understanding may be subject to, an irreducible telos of mutual understanding is inherent to linguistic communication.⁵²

Communicative reason, too, treats almost everything as contingent, even the conditions for the emergence of its own linguistic medium. But for everything that claims validity within linguistically structured forms of life, the structures of possible mutual understanding in language constitute something that cannot be gotten around.⁵³

The final authority in the process of the Enlightenment here remains solely argumentative discourse, to which those involved in the process of the Enlightenment are to unreservedly submit, allowing their hypotheses and theories to be examined by others.

The vindicating superiority of those who do the enlightening over those who are enlightened is theoretically unavoidable, but at the same time it is fictive and requires self-correction: in a process of enlightenment there can only be participants.⁵⁴

1.3. HOW TO ACHIEVE THE ENLIGHTENMENT

As a participant in the Enlightenment in this sense, in the following reflections I want to contribute to the understanding of the concept of the Enlightenment, which while not attributing too much weight to utopian elements, does not exhaust itself in the sobriety of an epistemic attitude. I attempt to develop an Enlightenment concept, which tries not to settle for a “levity in dealing with philosophical topics”; for doing so leads all too easily to the elimination of the presuppositions for that very levity. Rather, I attempt to hold fast to the perspective of intersubjective learning, responsive to suggestions and critique. To this end I assume the following preliminary provision:

(A1) The Enlightenment is the social process in which we develop and learn to understand our ability to understand.

At the center of this proposition lies the idea that the process of the Enlightenment should be characterized on the basis of competencies that are so fundamental that any possible self-understanding of social actors is connected to these competencies. It is clear that (A1) is initially nothing more than a thesis that stands in need of further justification; it appears to be especially questionable how (A1) stands in relation to those conceptions of the Enlightenment that place the process of the Enlightenment in an intimate relationship to the realization of rational structures, be they rationalizable beliefs, effective problem solving, just social structures, or what have you. For the justification of their proposals, normatively committed Enlightenment positions are dependent on theories of rationality, with the help of which it can first be shown what we can view as a rational belief, problem solving, or social structure. However, Enlightenment, understood in this way, is then a process that is dependent on an explicit theory of rationality. Viewed more closely, it becomes clear, however, that (A1) is not a noncommittal thesis, for irrespective of which conception underlies the characterization of the rational, thesis (A1) yields restrictions that are authoritative for all conceptions of rationality. These restrictions come to light if we attempt to characterize a minimal conception of the Enlightenment that is compatible with (A1).

1.3.1. A Minimal Conception of the Project of the Enlightenment

If the process of the Enlightenment is understood as the social implementation of competencies of understanding, then, by characterizing requirements for social processes of understanding that cannot be undercut, we can determine basic aspects of the process of the Enlightenment more precisely; on the basis of such a minimal conception it will not be necessary to center the project of the Enlightenment on wide-reaching and specific objectives. Such requirements are:

(A2) Any theory of the Enlightenment

1. must, as a normative theory, be normatively consistent;

2. cannot, as a theory with implied descriptions of human competencies, make decisions in advance about specific competencies that might limit self-understanding; in short, it must be anthropologically open;

3. must, on the one hand, be able to place the conception of its own theory of understanding in the context of historical conceptions of understanding (tradition/historical competence) and, on the other hand, must not consist of a mere reformulation of these (progressiveness);

4. must be universal insofar as it must include any possible form of understanding.

It may not be so obvious that precisely these—not exactly weak—demands be considered the minimal criteria of a concept of the Enlightenment. Thus in what follows I will comment on and attempt to justify the specific criteria. The general strategy here is to eliminate those theories of rationality that do not meet the restrictions following from (A1) from being considered a normative basis for the Enlightenment process.

1.3.1.1. On Normative Consistency

Under the keyword of normative consistency I want to thematize problems of normative logic that are connected to the constitutive normative character of all rationality theories. Insofar as all rationality theories make evaluative standards available, they are at the same time faced with the problem that the means that they use in their construction must not contradict the evaluative standards to which they owe their normative content. On closer examination, problems of normative consistency are posed at two levels.

If the Enlightenment is dependent on normative theories of rationality, then the critique based on rationality theory must be precluded from contradicting the very normative basis to which it owes its normative justification, i.e., Enlightenment critique must be protected from self-application that is autodestructive.⁵⁵ If, for example, the medium of the critique itself cannot be viewed as anything other than an instantiation of that which is to be criticized, then the critical status of an intervention employing this medium collapses. If a conception of the Enlightenment exhibits a normative inconsistency of this sort, this does not merely mean that the critical enterprise is up against particular difficulties, which are to be supported once it is finally specified what the Enlightenment is; rather, the result is that a conception like this is itself no longer rational. For good reasons cannot be provided for engaging in a practice that inevitably reproduces the very thing that the practice counters. In other words: it is incomprehensible what orienting oneself on a theory of rationality like this would mean.

However, it is just as clear that the problem of autodestructive self-reference cannot be solved simply by removing the normative basis from the scope of those things that can be criticized on its basis. For in this way the normative foundation would in principle be immunized and above all disconnected from further learning processes. However, if a concept of rationality requires that its normative basis be immunized, it limits the process of learning by bracketing certain objects from criticism. Then, however, it implies two incompatible postulates: the postulate of justifying normative foundations, which consists of nothing other than dealing with anticipated criticism, and the postulate of exempting the normative basis from justification. However, one cannot rationally (verstehend) adhere to postulates that are so incompatible. Between the autodestructive character of the critique and the immunization of its normative foundation, for a theory whose basis falls prey to its own critical potential, the possibility of self-criticism must be secured by developing a follow-up theory whose normative basis falls within the scope of the criticism, and yet whereby the criticism is held at bay. On the other hand—and that is the second aspect of normative consistency—on the basis of a desire to make the normative basis comprehensible, circular structures must not be built into the conception that end up being identified with the presuppositions and the performance.

1.3.1.2. On Anthropological Openness

The criterion of anthropological openness ought to raise awareness of the implications of theoretical concepts: beyond having a design value for the theory, basic concepts also have consequences for the openness of the anthropological design. For if the contribution of a theoretical construct comes at the cost of the possibilities available to people in developing images of themselves, then the possibilities for self-interpretation are sacrificed to the more primary need for stable theories. A theory of rationality that is to satisfy the criterion of anthropological openness thus must not secure its epistemological stability by constituting itself or its components in a way that predetermines the possible scope of self-conception.

The problem that I am attempting to address with the criterion of anthropological openness becomes clearer if we keep in mind that anthropological considerations and theoretical considerations about rationality are interwoven in a peculiar and often intimate way. For, on the one hand, the emphasis on specific abilities, with the help of which, from an anthropological perspective, we attempt to differentiate the human species from others, almost always also forms the basis for the exemplification and explication of rationality; on the other hand, theoretical considerations of rationality have anthropological implications. Reference is made to these in the attempt to provide evidence of the suitability of conceptions of reason in species-specific dispositions or practices. For example, the function that the ability for planned labor⁵⁶ or that the availability of language⁵⁷ take on in anthropological contexts when determining criteria for the species is complemented in each respective case by a corresponding concept of rationality (Vernünftigkeit); they become a prescriptive amalgam that runs the danger of becoming dead weight in the process of the Enlightenment. The criterion of anthropological openness thus has the function of testing whether rationality concepts are sufficiently general and do not owe their plausibility to the fixation of anthropologically prominent species competencies. For the type of beings we understand ourselves to be and the type of beings we consequently partially are should continually be able to be newly determined in the course of the Enlightenment process.⁵⁸

1.3.1.3. On Historical Competence and Progressiveness

The criterion of historical competence and progressiveness is meant to ensure that a concept of reason, like every other idea that has made a contribution in the continuum of Enlightenment ideas, is placed in relationship to the work that has already been performed and the breadth of ideas that have been developed; against this background, its specific contribution is to be elucidated. To put it technically, the criterion requires, on the one hand, that new concepts be compatible with the old ones. This functions to make access to proposed concepts more probable by conjoining them to known ones. On the other hand, it expresses the expectation that a contribution proceed beyond the stand of the discussion thus far reached as long as the issue at hand remains problematic. There is a simple reason that the criterion of historical competence is a necessary criterion for rationality theories. For the resources that are available to us for understanding any possible new rationality theory can by definition only originate from already understood, known theories. Insofar, the criterion of historical competence should only secure conditions for the possibility of understanding theories, which are fundamental for the process of the Enlightenment.⁵⁹

The criterion of progressiveness prohibits the mere citation of existing theories, thus reformulations that are not interpretations. For we demand of interpretations that are to document the understanding of a contribution to the process of the Enlightenment that they be informative by showing, in their own formulations, that it is plausible that this contribution can be made comprehensible with the help of reasons.

1.3.1.4. On Universality

The most suspect of the criteria of a theory of rationality is certainly the claim to universality, precisely in the context of a conception of Enlightenment processes. Against the background of the now widely propagated view that demands that are supposed to be connected with philosophical theories ought to be abnegated, the criterion of universality surely awakens the impression of a search for new masterly thinking. In the following I would thus like to illustrate more clearly why I think such a criterion belongs to the minimal criteria needed for the concept of rationality. Concerns about constructing a universalistic concept of rationality are primarily based on two caveats.

If “universalistic” is supposed to mean “culturally invariant,” then it is objected that, hidden behind the well-intended requirements of a universalistic concept are nothing more than ethnocentric or cultural-centric overgeneralizations of standards that are in fact relativistic, and that, in the wake of these, there is a leveling of the specifics of cultural developments. In this criticism—which is motivated, or at least supported, by the obvious economic dominance of the centers of Western capitalism and the destruction of traditional ways of life in other political economies, which are subdued and have compatibility forced on them—the Western and merely allegedly universal concept of rationality is reproached for having an element of violence immanent to it that extinguishes what is foreign, and the source of the evil is perceived precisely in the demand for universality. From this perspective, the alleged universality of the rationality concept becomes the theoretical background for a cultural imperialism that subjects what is foreign to its own models.

A further caveat against the universalistic conception of rationality views a rationality concept that is allegedly universal as imperialistically imposing itself on differing spheres of human activity. In accord with this caveat, it is questionable whether all areas, thus whether all types of action or cultural spheres also within a culture, may be viewed similarly from the standpoint of a unified concept of rationality. Put more precisely, in accord with this perspective, it is questionable, first, whether all of the object areas are accessible by reason, and second, should this be affirmed, it remains unclear whether we here are dealing with one reason or different types of rationality, each of which is to be assigned to a specific object area.

At the basis of both of these elements of suspicion is the common motive to contrast the diversity of types of action, social practices, institutions, traditions, and cultures with the alleged hybrid demand of universalistic rationality theories, whose grasp of the factual plurality of phenomena leads to a homogenization that levels the differences under the aegis of a monomaniacal concept of reason. However, in view of this critical perspective, it is questionable what it could mean to learn from it. For if we are supposed to be able to learn something from this perspective, then it is only under the presupposition that we can understand what the arguments of critical intervention mean and what changes in our practices they suggest. However, apart from the reconstruction of possible motives, there are serious obstacles precisely to understanding this.

The alleged homogenizing universalism is most radically opposed by a relativist interpretation. In the light of this often-criticized variation, we are faced with the alternative either of interpreting statements that claim the relativism of all propositions as statements with truth-value, which are self-refuting, or of describing such statements not primarily as expressing truth claims, but as having expressive goals, which, however, then certainly remain deficient with a view to theoretical demands.

Because of the methodological lethality of radical relativistic positions, to be able to ensure a continued hearing of the caveats about universalism that arise, a contextualist reformulation is needed, such as the one Rorty, borrowing from Wittgensteinian motifs, has in mind; it emphasizes the need to save the antifundamentalist impulse of relativism in the process of disarming it.⁶⁰ Rorty rebuffs two untenable interpretations of relativism as being self-contradictory or “overstretched”; namely, those that assert the indifference to all standards and those that assent to a plurality of standards that correspond to the diversity of procedures for justification. He then, however, characterizes an acceptable interpretation of the relativistic caveat as one that insists that, beyond describing the procedures for justification established in (our) society, there is nothing sensible about truth and rationality that can be said. This interpretation, which is contextualistic since it is ethnocentric, does not view itself as epistemologically secured, but expects evaluative expressions to be context dependent in a way that is supposed to be comparable to indexical expressions.

However, even if, against the background of this analysis, which has taken leave of inconsistent incommensurability theses, we grant that the formulation of context-transcending standards is connected with a God’s-eye perspective that we cannot assume, we cannot avoid the question regarding the perspective from which the differences can be diagnosed. For even a modest, context-sensitive theory of rationality must aim at differences in its diagnosis of cultural differentiations that are initially differences from the perspectives of those doing the diagnosis, and thus stated from their own cultural perspective; beyond this, however, the act of stating differences presupposes that they are related to an object that is in some way comparable. For otherwise it would not be at all clear whether the matter of concern was a (completely?) different kind of entity or process, or something that simply is not one of this specific type of entity or process. If, however, the diagnosis of allegedly essential differences is part of the kernel of contextualist relativism, this more moderate relativism implies the universalistic hypothesis that differences concern something comparable as well.⁶¹

The argumentation strategies sketched out at the beginning can only be sensibly understood as a critique of certain formulations of a universalistic concept of reason. For, on the one hand, a congruous methodic relativism is a methodologically self-contradictory position, from which nothing follows but the unacceptability of this position.⁶² On the other hand, positions that insist on a basic divergence of cultural developments or human spheres of activity, and thus reject universalistic concepts of rationality, themselves lay claim to a latent form of universalism insofar as the divergence that they advocate can only be considered valid against the background of a universalizing hypothesis that can secure the comparability of what is divergent. However, if this latent universalism is itself a necessary condition for the diagnosis of culturally and ethnically significant differences, then the real problem connected with universalistic conceptions of rationality is not to be found in their universalistic claims or implications, but in the specific characteristics of the concepts of rationality themselves, which only more or less meet the universalistic claims.

The task, then, that presents itself in connection with the universalistic implications of conceptions of rationality consists in designing the concept of rationality so that it can avoid being unmasked as, in fact, particular. This demand is compatible with a theory of rationality aiming for universality as a regulative idea and the failure of every concrete formulation of a rationality concept in reference to this demand. But that can only be shown from the perspective of a rationality concept that is more universal than what is criticized.

The universality criterion should thus here be understood as indicating that it allows the criticism of the particularity of a concept of rationality that is propagated as a real universality; this can be done without necessarily being oriented immanently on the normative structure of what is criticized. This criterion makes a distance from (one’s own) standards possible; it is analogous to ideology critique, but does not share its presuppositions, which require that the normative basis for the critique be adopted from its objects.⁶³ Recalling the fact that one’s own standards are a product of a process that is subject to contingent influences, a self-distancing is possible that is guided by the criterion of universality. For this, the concept of universality must be designed to be supported by necessary presuppositions of critique and the generation of arbitrary alternatives to what is being criticized.

Both of the preceding considerations can be connected to the following argument and radicalized. For it can be shown that universalism is in a certain sense necessary, and indeed necessary for dealing with what is other—and probably endangered by the universal—or what is allegedly “completely other.” If we want to understand activities or practices that we are not familiar with, we must initially impute the standards we use for orientation to those whose behavior we want to understand.⁶⁴ It is clearly inevitable that we start with the differences that we are familiar with and on which our familiar concepts are based; for initially we simply have nothing else to work with. If, on the basis of these standards, we fail to understand activities that we are not familiar with, we can modify our schemes for differentiating things (our differentiational sets) and base these on other hypothetical standards. But even these types of standards, which there have been reasons to modify, remain our standards, and they have to remain ours if they are to assume their function in our process of interpreting and understanding. For if the set of modified standards were (normatively) completely incompatible with the set of standards that we started out with, then it would not be possible to provide any reasons that would be able to explain the transition from set A to set B. For every hypothetically modified set of standards, the following criteria thus must apply:

a. In order to secure its status as a set of standards, we must be convinced that we are able to explain our own actions with its help.

b. We reservedly assume that the hypothetical standard is universal. Our reservation, however, is limited: we cannot allow the reservation about our standards to be all-encompassing, because doing so would deprive us of the possibility of understanding (ourselves and others or the other).

In the face of these conditions, it is clear that every concept of rationality must prove its worth for two perspectives at the same time. Confronted with the phenomena that we are attempting to understand, it must hypothetically allow modification, and it must retain the ability to be connected to the standards with the help of which we explain our own actions as rational.

This weak concept of universalism—weak because it is necessary—is not to be confused with cultural imperialism, because this universalism is reciprocal and because every formulation of a hypothetical-universalistic standard counts on its own provisional character—not in the form of a global suspicion, but with the consciousness of the partial fallibility of every single formulation of rational standards. What the theory of rationality can adopt from the relativistic or contextual critiques of reason thus does not take the form of an argument, but the form of a gesture, a gesture that says essentially: “Do not hold the formulations of the standards of rationality that you are assuming to be valid under all conditions and for all time.” And “learn to achieve a reflexive distance from your standards by viewing them from the perspective of alternative interpretations!” Yet the only form in which this gesture can be taken into account is as the insight that each and every assumption of universality is fallible. But an assumption of universality can only be falsified with the aid of a better position—that is to say with a constructive vote of no confidence.⁶⁵

If we presume with the normatively laden views of the Enlightenment that it is only possible to determine the process of the Enlightenment with the help of normative criteria, which, for their part, are capable of being justified, then it is clear that a theory of such criteria moves directly toward a theory of rationality. With provisions (A1) and (A2) I have specified a perspective that makes it possible to consult different and competing theories of rationality with a view to which conceptual resources they make available for an adequate, comprehensive concept of understanding.

Because the outline of my study is anything but obvious, in the following section, after offering a relatively schematic historical overview, I want to again take up those systematic demands that rationality theories can raise.

1.4. ORIENTING REFLECTIONS ON THE THEORY OF RATIONALITY

Beyond the path of a totalizing critique of reason, Habermas’s assessment “that philosophy in its postmetaphysical, post-Hegelian currents is converging toward the point of a theory of rationality”⁶⁶ seems to have found broad acceptance.⁶⁷ Even at first glance, however, it is clear that the rediscovery of the topic of rationality is not connected with a movement toward a common concept of rationality. So, as in connection with the concepts of the Enlightenment, rationality theories with normative claims can initially be differentiated from those that reject the normative implications of the rationality concept or that at least consider implicit (vernunftinterne) normative determinations of the rationality concept to be unsustainable.

In order to examine the group of views that lie between emphatic normative theories and normatively disinterested theories, in the following reflections I would first like to cover the systematic possibilities that exist regarding the basic conception of rationality theories. To simultaneously embed these aspects of theory construction in a historical context and in this way to depict strands of development that are taken up by current works in this area, I will—leaning on the work of Herbert Schnädelbach⁶⁸—draw out some lines along which the discourse about the definition of rationality has developed. In connection with this, I will situate some paradigmatic conceptions of our day against the background of the historical development and discuss systematic problems of rationality theory.

The history of the concept of rationality can—following Schnädelbach—be described as a history of the subjectivizing of reason, which converges with the history of the decentering of reason and rendering it less normatively potent.⁶⁹ Schnädelbach describes the steps involved in the history of the subjectivizing of reason as follows.

Following the objective logos of antique metaphysics comes the first step of the subjectivizing; this is objective subjectivizing, which is found in Christianity when the objective logos is implemented as a characteristic of a personal (personenhaften) God. This process passes into the phase of subjective subjectizing, which is raised by Descartes’s reconstitution of what is objective in his methodic, subjective doubt and that continues through Kant’s reconstruction of reason as a subjective capability. In this process, Hegel’s absolutizing of reason seems like a final rebellion of the objective view of reason, centered on a reformulated logos, undercut by the tradition of an “empirical interpretation of the aprioris of reason” that prepares the setting for the post-Hegelian decentering.

What comes after idealism is no longer a philosophy of reason, but a metaphysics of the irrational . . . , existential thinking and then naturalism and objectivism in all its forms. What they have in common: reason is not denied or rejected out of hand. It is pushed into secondary ranking; it is no longer the core or the essence, but . . . a function of something which in essence is not reason.⁷⁰

The results of this latter phase of the decentering process, which simultaneously serves as a preparation for the preceding⁷¹ described forms of postmodern demasking critique, can be illustrated in reference to positions that subordinate reason to specific authorities external to reason. Schopenhauer, for example, views reason as a subordinate tool of the will. Nietzsche and Freud follow him in this. Naturalism views reason as a natural fact, as a product of evolution. Sociologizing, in contrast, views it as a social fact. And, finally, existential philosophy (Kierkegaard, Heidegger) understands reason to be a phenomenon that is subordinate to the specifics of people, namely, existence and decisions.⁷² In all of these cases, reason is defined by its function, as an instrument of authorities external to it. Rationality is, in a manner of speaking, a competence for solving problems that reason-extrinsic authorities use in an attempt to solve problems according to their own goal-setting standards.

According to Schnädelbach, the rationality at the gate of this decentering process, which is still going on today, has remained a philosophical topic because of nineteenth-century neo-Kantianism, which insisted that reason is “not simply one object among the other objects of the sciences, because it has something to do with what makes scientific objectivity possible in the first place.”⁷³ With the concept of reflection, which has been connected with the resistance to a voluntarist and emotivist neutralization strategy since Kant, procedures become the focal point of considerations about rationality theory.⁷⁴ If, namely, judgments about the status or the function of reason itself always rely on reason, then the explication of what reason is can orient itself on the procedures of justification. Procedural concepts of reason thus do not rely on authorities external to rationality, whose importance in contexts of justification itself has to be justified; rather, the procedures of justification themselves become a normative foundation. The kernel of such approaches is, as Schnädelbach plastically notes, the idea “to use reason, which is the medium of critique, at the same time as its standard.”⁷⁵ This is the enormous advantage of procedural conceptions. Reason need not be brought into play as a static capability, but, as the very process of critique, it is open, and it is in this way to be connected to learning processes and Enlightenment. Committed to this insight, today above all Habermas has once again taken up this central element of Kant’s critical philosophy and reconstructed it in his theory of communicative reason in the concept of discourse as a procedure for the intersubjective examination of truth claims that are raised.⁷⁶

Today if we once again want to place rationality at the center of philosophical reflection, on the one hand—and here I decisively agree with Schnädelbach—this must occur with a consciousness of the results of the “demythologizing history of reason.”⁷⁷ As a result of this, we view reason as a concrete human capability: “Without the realization of the human disposition, ‘rationality,’ there is nothing rational in the world; everything else is bad metaphysics or misleading metaphorics.”⁷⁸

On the other hand, however, simultaneous with the subjectivizing—which the approaches that subordinate reason to authorities external to it make clear—the process in which it loses normative potency occurs:

Insofar as we can survey the history of reason, it can be described as a history of the normative neutralizing of reason, which, in the end, offers pure functionalist rationality conceptions; the transition from a substantive conception of reason to a functional conception, which has only really been carried out today, has a long prehistory.⁷⁹

Procedural concepts, which share with the functional concept a temporal aspect and a tendency to dynamization, but not normative vacuousness, thus initially appear to be the only means by which to develop a rationality concept with normative ambitions today. Indeed, in connection with the normative neutralization of rationality, Schnädelbach has drawn attention to a striking convergence between functional and procedural conceptions of reason:

The borders between a purely procedural and a functionalist rationality concept can no longer be drawn if the procedural concept no longer includes an element of substantialist rationality, i.e., if the rules that establish the procedures of norm justification themselves are normatively neutral.⁸⁰

If, in the development of my project, I adhere to a procedural rationality concept, then a goal of this project must also consist in identifying rules that, in a manner to be more precisely determined, are not arbitrary.

The preliminary results of the development of rationality concepts are summarized in figure 1.1.⁸¹ However, here it is necessary to point to one aspect of the project of providing a foundation to a rationality concept that relies on elements external to reason that Schnädelbach does not illuminate in his presentation of the process by which the normative element of reason loses potency. For there is a methodological problem connected with the strategy of generating a form of reason that is internally normatively laden and is simultaneously able to be reconstructed as a motif of functionalistic concepts. That is to say, if substantialist concepts of reason—be they objective, subjective, or procedural in nature—are not able to fall back on authorities external to reason for a normative definition of the concept, then the only remaining possibility is to conceive of rationality such that the concept functions simultaneously as the basis of norms as well as the concept for their justification. In opposition to this inevitable circularity of substantialist concepts of reason, it is possible to call on the intuition that only authorities external to reason are able to break through the circle of self-sufficient self-justification. Reason must, as it were, be brought into connection with something other than itself. For “the process of reason, if left to its own devices, resembles a self-regulated machine whose functioning is no longer controlled by a particular end.”⁸²

The versatile forms of the problem with such “sui-sufficient” conceptions of rationality can be detected in the decentering history of rationality. Here, two forms of self-sufficiency are to be distinguished. The problem of connecting reason to purposes or goals must be differentiated from the question of which requirements external to reason rationality has. It is possible, for example, to point out that, for its part, rationality has historical requirements that are not themselves rational.

What we consider to be our reason did not always exist. Our individual and collective capacity to think and to act reasonably obviously was generated, and it has changed—whether in the wake of natural or cultural evolution of the development of language and communication, interaction and production. As something generated and changing, it has a history.⁸³

Even this admittedly rough sketch sensitizes us to an alternative that each theory of rationality that is to be developed must address: on the one hand, if we decide on the substantialist-procedural interpretation, then an authority must be found that makes it possible to break through the circular nexus of relations of justification internal to reason and to provide reason with a task that points beyond the justification of justifications. On the other hand, if we take a functionalist path, then it is necessary to avert the danger of reducing reason to a loose-leaf collection of hypothetical imperatives.

FIGURE 1.1 Concepts of rationality according to Schnädelbach (with additions)

1.4.1. Positions in the Current Debate on Rationality

Calling to mind the positions in the current debate on rationality, one discovers a complex field with fault lines along which it is possible to make out relativistic, culturalistic, universalistic, normatively ambitious, normatively mitigated, more reconstructive, and more constructivist positions.84 In what follows, in proceeding through some paradigmatic positions, I attempt to sort through the field of argumentation to gain an idea about which form a theory of rationality ought to take, which normative claims it ought to make, and what its explanative basis ought to consist in.

Initially we can presume that we ourselves—in contrast to what philosophical reflection about rationality suggests—actually are quite able to decide what is rational and what is not, and that we can do this without having well-worked-out criteria of rationality. We have obviously not acquired this familiarity with reason by memorizing systems of rules for rational action (otherwise the philosophical question of the kernel of rationality would hardly be so pressing), but because of the fact that we are practically incorporated into a more or less rational social practice. In his concept of reason as an aspect of our orientation practice, Friedrich Kambartel proceeds on this intuition.

1.4.1.1. Rationality as an Aspect of Our Orientation Practice

On the basis of the analysis of two different principles of rationality (rational self-interest and rational discourse), Kambartel reconstructs a cultural practice of judgment in order to show that reason cannot be understood as a criterion, a general principle, or a set of rules. If, in the face of concrete problems, we “are in a position as learned rational beings to deliberatively judge a series of different criteria, albeit connected to one another, in diverse ways,” that is, we are in a position to weigh the suitability of the criteria for the problematic situation, in doing so we are not able to refer to higher criteria in compliance with which we might be able to secure the adequacy of our judgment regarding the suitability of the criteria of judgment.⁸⁵ In short: we have to get by without universal standards of adequacy. Thus “reason is dependent upon our power of judgment and upon the life experience that in general underlies it.”⁸⁶

As a proponent of the concept of rationality that attempts to accommodate the view that, with the lack of general criteria of adequacy, life experience is at the basis of reason, Kambartel reconstructs culture as the framework within which we—as participants in a common, not individual, practice—educate ourselves to become rational beings. As we begin to participate in cultural practices, we become acquainted with the “various definitions of the rational” that are at work within a culture of reason and that, under the presupposition of “‘local’ suitability,” belong to the “grammatical network of this culture.”⁸⁷ In this, the various definitions of rationality illustrate the diversity of the “grammatical forms for overcoming a particular (individual, subjective) orientating practice.”⁸⁸

The obvious advantage of such argumentation consists in the fact that rationality, with the help of a culturalistic reconstruction, need not be presented to the societal practice from a problematic external perspective, but can be viewed as already existing in the social world. However, this advantage—if one agrees with Mittelstraß—comes at a high price: “If the culture of reason as a universal culture is itself normative, as it is now called, precisely then the appeal to its reality is affirmative.”⁸⁹ The rejection of a normative interpretation of the rationality concept, which could be explicated criteriologically, “makes reason into an attribute of relationships (Verhältnissen) rather than understanding it as a medium of critique.”⁹⁰

Beyond these costs, which were calculated by Mittelstraß, there are further ones that are caused by the soft contextualism of Kambartel’s concept. For what precisely does it mean to speak of different definitions of the rational (in different contexts)? In any case, the step beyond a particular orientating practice alone cannot serve as a sufficient criterion for categorizing an orientation as rational, for transindividual orientation practices, like orienting important decisions on the basis of horoscopes, can eventually be irrational. Criteria are thus needed with the help of which it is possible to characterize a general orientation practice as rational.

With a view to the argument that insofar as it is not possible to find general criteria of the rational, because “definitions of the rational” are really only meaningful in their respective contextual relationships, it is necessary to ask how it is possible to speak of a reason that transcends specific contexts if the contextual interpretation is not compatible with more basic criteria of the rational. In other words: if we admit that acting beings act rationally in specific contexts, then we impute elementary rationality principles to the actors, principles that can never be contradicted in their contextual specification. Otherwise talk of reason simply would be disparate. Any specification of criteria of rationality that aims to be adequate must be understood as a process that, technically speaking, maintains downward compatibility to fundamental criteria of rationality. The contextualization of reason that is required is thus not a diversification but a specification. Let us then once again take up the idea that Mittelstraß places at the center of his critique of Kambartel.

1.4.1.2. Reason as an Ideal

Mittelstraß understands criteria of rationality as criteria that are developed in opposition to existing affairs and thus cannot be reconstructed from them: “Reason is an idea. And ideas are not defined. They are rather determined in reference to ideal demands, postulates.”⁹¹ Mittelstraß does not refer back to sociocultural practices as the source of reason, but confronts these with outlines of rational affairs. His concept is not determined in a reconstruction of what exists but in contrast to it.

Reason is thus not something that is everywhere to be found, but—according to the Enlightenment thinkers—something that can always be awakened, in every head and in every life. In any case, reason is always an achievement.⁹²

This procedure is connected to an experimental character of normative rational postulates that by no means require that reason be exhaustively defined or that a principle of rationality be formulated. However, in Mittelstraß’s view, beyond the particularity of the postulate of reason, there is an “idea of the unity of reason,” which can be articulated in “developments that (increasingly) bring about a rational life, that is, in rational developments.”⁹³ As criteria for affairs that are to be directed to rational developments, Mittelstraß mentions general reasonableness and the possibility for consent to the affairs, their level of development in contrast to naturally emerging affairs, and their suitability to serve as a “basis for future justified developments.”⁹⁴

However, it remains problematic that, for their part, the criteria for rational development cannot be criteriologically justified. Like the particular rationality postulates, the criteria of general reasonableness and of the possibility for consent retain an obvious contingent character. Conditions for justified consent—for example, as in Habermas’s attempts to depict them on the basis of indissoluble connections between meaning and validity, which are consequently to be differentiated from contingent suggestions—are not, however, joined to procedures with the help of which suggestions can be reliably verified.

A more interesting impulse is contained in these reflections, namely, the impulse toward a criterion of development; this can be read as a demand to assess suggestions in reference to their consequences for the possibility of shaping future affairs. So, rationality can be understood as a criterion that prescribes the possibility of future action under conditions where there are an abundant number of alternatives. Here too, however, a strong justification of the criterion is lacking.

The reflections so far contrast two different ways of viewing reason, each of which is poignantly debunked in the respective work of Kambartel and Mittelstraß. While Kambartel, in a rather reconstructive perspective, insists that reason is not something that we have to be lectured about from a position independent of our practices, Mittelstraß emphasizes the idea that, for the sake of its critical function, reason must be something that, of necessity, retains a certain distance from our de facto practices. From his rather constructivist perspective, he emphasizes that if we view reason as something that is, in a manner of speaking, inherent to our practices, then the normative character of reason gives way. Both positions appear to capture something correct, but given their incompatibility, they cannot both be simultaneously correct. However, we must thus be able to expect from a more developed theory of rationality that, in the process of a more or less complex mediation, it does justice to both the normative character of reason and also to the fact that we—as potential addressees of the norms of rationality—view ourselves as beings who are in fact at least partially already reasonable. Habermas’s theory of communicative action can be viewed as an attempt to do justice equally to both of these elements.

1.4.1.3. Rationality and Reaching Understanding

Habermas’s attempt “to secure a concept of reason by means of formal pragmatics, that is, by means of an analysis of the general characteristics of a communication-oriented action”⁹⁵ is certainly one of the most ambitious conceptions of a reconstructive rationality concept. Here, too, in his universal pragmatic construction, Habermas refers to a social practice, but this case is more complex.

First, Habermas constructs his rationality concept cognitively: “When we use the expression ‘rational’ we suppose that there is a close relation between rationality and knowledge.”⁹⁶ Here rationality describes a disposition of individuals, who, on the basis of their knowledge, attempt to successfully achieve their objectives. A decisive differentiation of the mode for applying the propositional knowledge can be added to this first characterization, which defines rationality in reference to its application in descriptive knowledge.⁹⁷ According to whether this knowledge finds application in goal-oriented action or in communicative action, which is oriented on processes of uncoerced agreement in the medium of argumentative speech, we can distinguish a form of cognitive instrumental reason from a form of communicative reason.⁹⁸ As a consequence of the reference to knowledge—underlying both types of action is “fallible knowledge”⁹⁹—rationality is simultaneously placed in a context of justification and critique. This is why Habermas’s reflections lead “in the direction of basing the rationality of an expression on its being susceptible of criticism and grounding.”¹⁰⁰ For we judge the rationality of actions with regard to the possibilities for justifying the claims that are raised by them. The ability to connect such claims with actions is doubtless a competence of subjects who have communicative and action competencies. If we understand the concept of rationality as a competence of subjects, that is, we use the term dispositionally, then individuals who act with instrumental rationality must be imputed with other competencies than individuals who act oriented toward understanding.

Before these competencies are more precisely investigated, Habermas first broadens the parameters of actions oriented toward understanding; against the background of our way of applying the predicate “rational,” the earlier introduced example of truth—as a claim that is raised in communicative action—proves insufficient. For “normatively regulated actions and expressive self-presentations have, like assertions or constative speech acts, the character of meaningful expressions, understandable in their context, which are connected with criticizable validity claims. Their reference is to norms and subjective experiences rather than to facts.”¹⁰¹

It is now possible to connect the view of the expanded cognitive concept of rationality to a reconstructive procedure that investigates the competencies that must be attributed to actors oriented toward understanding. Habermas calls the research program that takes up this task universal pragmatics. In analyzing the necessary conditions for communicative competence, universal pragmatics concerns itself with the reconstruction of a “pre-theoretical knowledge,”¹⁰² which underlies the ability to linguistically communicate. What is fundamental here is that communicative competence is differentiated into an ability that underlies the generation of grammatically correct propositions and the competence that, based on this communicative competence to follow rules (or communicative rule competence), puts the speaker in a position to use propositions “as elements of speech, that is, for representational, expressive, and interpersonal functions.”¹⁰³ Only on the basis of this communicative rule competence are speakers in the position “to embed” their speech in “a well-formed sentence in relation to reality.”¹⁰⁴ To do this, a speaker must refer to an object; that is, she must determine some propositional content that reproduces a fact or experience; she must express herself by expressing her intentions; and she must express herself in such a way that an interaction arises between those involved in the communication, which has a form acceptable to the participants and thus conforms to recognized norms and self-understandings.¹⁰⁵

This means, however, that beyond fulfilling the demand for comprehensibility, “a successful utterance must satisfy three additional validity claims: it must count as true for the participants insofar as it represents something in the world; it must count as truthful insofar as it expresses something intended by the speaker; and it must count as right insofar as it conforms to socially recognized expectations.”¹⁰⁶ A possible agreement among those involved in the communication “is measured against exactly three criticizable validity claims; in coming to an understanding about something with one another and thus making themselves understandable, actors cannot avoid embedding their speech acts in precisely three world-relations and claiming validity for them under these aspects.”¹⁰⁷ The validity claim of truth that is raised in constative speech acts, with which the cognitive concept of rationality began, is thus only one special kind of validity claim that can be raised in speech acts.¹⁰⁸

As a result of the universal pragmatic reconstruction the concepts of understanding and meaning, on the one hand, and the concept of validity, on the other, complement one another. The core idea of truth semantics—that “to understand a sentence means to know the conditions under which it is true,” as this was developed by Davidson and Dummett in connection with the early Wittgenstein¹⁰⁹—is transformed. This is done as the dimensions of validity are now broadened to include the “variety of illocutionary powers” and as a result of the speech act theoretical transition from the analytical unity of a “sentence” to the unity of a “speech act.” The transformation is to the following formulation:¹¹⁰ “We understand a speech act when we know what makes it acceptable.”¹¹¹ And “a speech act may be called ‘acceptable’ if it satisfies the conditions that are necessary in order that the hearer be allowed to take a ‘yes’ position on the claim raised by the speaker.”¹¹²

In this way Habermas gains a pragmatically reconstructed concept of communication that includes not only the exchange of information, but also aspects of action coordination; this concept is at the same time connected to a normatively laden theory of meaning, which connects meaning to justifiable (einlösbar) validity claims and thus establishes “an inseverable composite of the meaning and validity of speech acts.”¹¹³ Herewith, the possibility of understanding speech acts is connected to the fact that criticizable validity claims are raised in speech acts; in contrast, speech acts to which this does not apply cannot be understood.

Not all illocutionary acts are constitutive for communicative action, but only those with which speakers connect criticizable validity claims.¹¹⁴

As a result of the outlined transformational steps, the explication of what constitutes rationality can draw on the competence of communicatively active subjects to use language with communicative intent, a competence that speakers realize by making sense on the basis of criticizable validity claims. If the connection between meaning and validity that communicative actors draw on in interaction oriented toward understanding is to be made the basis of a concept of rationality, it must, however, be shown that communicative action is the basic form of linguistic action. So, it must be shown in the next step “that the use of language with an orientation to reaching understanding is the original mode of language use, upon which indirect understanding, giving something to understand or letting something be understood, and the instrumental use of language in general, are parasitic.”¹¹⁵

In order to show this, Habermas ties into Austin’s speech act theoretical differentiation between locutionary, illocutionary, and perlocutionary acts; here, for Habermas, besides the locutionary act, with which speakers express matters of fact, the difference between illocutionary and perlocutionary acts is of special relevance. While the speaker expresses her intention with the help of illocutionary acts by determining the specific use she makes of a proposition, the perlocutionary acts serve to generate effects from the hearer, which the speaker attempts to elicit through a goal-oriented attitude. Figure 1.2 presents an overview of the internal differentiation of the speech act.¹¹⁶

By means of speech act theory, in a further step, the illocutionary element of the performative speech act can now be identified as a component of a speech act that is constitutive for action oriented toward understanding insofar as it is used to help characterize the modi of the speech act—i.e., to characterize it as maintaining, promising, suggesting, etc.—that relate the communicative meaning of a proposition to the intentions of the speaker.

While the “illocutionary aim a speaker pursues with an utterance follows from the very meaning of what is said,” the perlocutionary aim of the speaker “does not follow from the manifest content of the speech act; this aim can be identified only through the agent’s intention.”¹¹⁷ As in the case of teleological action in general, to reconstruct the perlocutionary intention of a speaker, we must “refer to a context of teleological action that goes beyond the speech act.”¹¹⁸ For perlocutionary effects can only be made out in relation to descriptions of states of affairs in the world,¹¹⁹ that is, in relation to descriptions that transcend the parameters of interpersonal relationships. However, a condition for the success of perlocutionary action is that at the same time the speaker hides her strategic intention behind an illocution.

FIGURE 1.2 Aspects of speech acts

Social action that is aimed at perlocutionary effects thus already presupposes the parameters of communicative action whereby it is shown that the modus of speech action that is oriented toward understanding is “the original modus” of language use. Hereby the intuition “that a telos of mutual understanding is built into linguistic communication”¹²⁰ is given a firm foundation with the help of universal pragmatics. Habermas only uses the term communicative action for the type of interaction that remains free of such strategic components and “in which all participants harmonize their individual plans of action with one another and thus pursue their illocutionary aims without reservation.”¹²¹ Figure 1.3¹²² situates communicative action in relation to other types of action and the intentions or validity claims that correspond to them.

In connection with the rational theoretical considerations, the following aspects are of decisive importance:

1. The view of a rationality concept can no longer be explicated by drawing on the instrumentality of problem-solving action, but is rather to be developed against the richer background of communicative action, which spans the dimensions of comprehensibility, truth, rightness, and sincerity.¹²³

2. If meaning and validity are brought into such a close relationship to one another that the constitutive meaning of a speech act is connected to an inherent claim to validity, then the judgment of the rationality of a person who formulates an expression can be joined to a monitoring procedure that tests the justification of the validity claim raised. Habermas views discourse as one such procedure. What is decisive here is that discourse is constitutively connected to the sphere of intersubjective understanding; for only in the framework in which action is oriented toward understanding, thus in which meaning can emerge, can validity claims constitutive for meaning be verified. Speech acts can be considered justified if their validity claims achieve intersubjective, shared recognition in a process of undistorted argumentative debate.

Every consensus rests on an intersubjective recognition of criticizable validity claims; it is thereby presupposed that those acting communicatively are capable of mutual criticism.¹²⁴

FIGURE 1.3 Types of action in Habermas

Rationality, we can infer, is the ability of competent speakers to generate speech acts that are criticizable. For achieving the conditions of criticizability secures the meaning of the speech act and the possibility of learning. The intrinsic normativity of the concept of meaning thus refers to discourse as a procedure for verifying validity claims; and it can be connected to processes of societal learning. The concept of rationality is thus, on the one hand, related to a universal pragmatic core of necessary conditions, which avoids historicizing reason; on the other hand, it is coupled to a procedure of critical testing that can both avoid the reproach of being affirmative and be open to processes of historical evolution.125

3. At the same time, the concept of rationality is not abstractly contrasted to the societal conditions, but can draw on the analysis of processes of understanding in which, even under contingent conditions, traces of reason can be made out that, simply as a result of the meaningful content of speech acts, can be related to an intersubjective procedure for critically testing validity claims. At the same time, here the necessary implications of such a procedure can, as normative instances, be mobilized against the existing forms of understanding (Verständigungsverhältnisse).¹²⁶ Philosophy, which is essentially a theory of rationality, can thus appear “as the defender of rationality in the sense of the claim of reason endogenous to our form of life.”¹²⁷

The advantage of this conception is that a critique based on the normative concept of rationality need not accept the accusation that it is affirmative, nor that it merely differs abstractly from what exists; rather it can draw on potentials that dwell within the existing practice. In contrast to the theoreticians of the old Frankfurt School, Habermas need not view the societal institutions as emptied of all traces of reason,¹²⁸ such that points of connection for societal learning processes are fruitless from the outset.

The obvious advantages of this conception, however, also entail problems. Here I would just like to single out four central difficulties. I would like to go into (1) the problem of the unity of the concept of reason, (2) the problem of universality, (3) the problem related to normativity, and (4) the problem concerning the foundations for theoretical considerations of meaning.

1. Among the central impulses of the Habermasian rationality theory is a differentiation between the dimensions of rationality that arises as the concept of action is differentiated. While concepts that are oriented on an (allegedly) homogeneous type of action can be reproached for giving rise to a concept of reason that is one-sided and overly simplified because of this orientation, it must be possible to question an internally differentiated concept of rationality with respect to its internal homogeneity. Habermas established a formal connection between the specific elements of reason that correspond to validity claims: “Between the differentiated moments of reason there is now only a formal connection, namely the procedural unity of argumentative grounding.”¹²⁹ Because reason, which in “metaphysical worldviews . . . was projected as a substantial unity,” could not be divested from the process of differentiating spheres of validity, “the concept of an objective reason itself fell prey in the end to the rationalization of worldviews.”¹³⁰ Habermas contrasts this historically disavowed substantial unity with a procedural one: “In principle, when substantive reason comes apart into its different moments, reason can retain its unity in the form of procedural rationality.”¹³¹

The kernel of this procedural unity is the process of argumentative justification. If we follow this view, however, we meet with difficulties; Habermas himself has called attention to the fact that validity claims of sincerity that are raised in expressive action are not accessible to argumentative justification:

The sincerity of expressions cannot be grounded but only shown; insincerity can be revealed by the lack of consistency between an utterance and the past or future actions internally connected with it.¹³²

And despite the prospect of the procedural unity of reason, Habermas does not hesitate to admit that solely “the truth of propositions and the rightness of moral norms and the comprehensibility or well-formedness of symbolic expressions are . . . universal validity claims that can be tested in discourse.”¹³³ Even if this difference can be softened terminologically, given that such expressions are viewed as open to criticism, but not as open to discourse,¹³⁴ it is still questionable what the repercussions of these qualifications are for the procedural unity of reason. For this unity would lose sight of specifically modern elements of aesthetic expressivity that Habermas had announced are to be appropriately considered for a nondistorted view of reason.¹³⁵ The kernel of the Habermasian concept of reason, it must be assumed, is so tailored to processes of intersubjective linguistic understanding that, in those places where the subjective sphere is not linguistically related to the process of intersubjective processes of understanding, it becomes nontransparent for a theory of rationality.

2. In my view it thus also remains questionable whether the rationality concept, despite its internal differentiation, is not too narrowly applied; for its basis—the connection between meaning and validity—is developed solely in the context of linguistic understanding. The problem connected with the methodological decision to analyze rationality in the context of linguistic action is that forms of action that cannot easily be reduced to linguistic action are relegated outside the scope of the rational.

3. Rationality, as a disposition of subjects capable of action, can be explained as the ability to raise criticizable validity claims. Here rationality is not initially connected to successful justification, but to speech acts being understandable, that is, that they fulfill the condition of being criticizable. Beyond this basic ability, which Habermas understands itself to be a validity claim,¹³⁶ rationality also consists in the ability to de facto justify validity claims that are raised. Rationality, in this narrower sense, is expressed in types of behavior (Verhaltensweisen) for which there are good reasons. At the same time, however, the existence of good reasons is dependent on a procedure for verifying them. We cannot know whether reasons are good reasons before we have verified the reasons on the basis of justified procedures. In addition, however, such justification procedures must be amenable to verification processes.

Habermas has indeed referred to ideal conditions of intersubjective verification procedures, but he rejects, for good reasons, an ultimate justification of verification procedures such as Karl-Otto Apel has attempted to carry out.¹³⁷ The cost of this decision is that the procedure he uses must itself be viewed as fallible, and it is only with difficulty that such a view, to which alternatives might be conceived at any time, can, for its part, be plausibly made an unshakable normative foundation.¹³⁸ In addition, Schnädelbach has rightly drawn attention to the fact that a procedural concept that disallows any form of fundamentalism and thus waives substantial characterizations of reason can no longer be differentiated from functionalist procedures.

4. In my view, among the most problematic decisions that Habermas makes about theory construction is the decision to make theoretical considerations of competence the groundwork for the theory of meaning. Initially the specific achievement of universal pragmatics consists in the reconstruction of those competencies of competent speakers that enable them to use linguistic entities in order to say something. And Habermas was faced with two sorts of competencies in his reconstruction, namely, linguistic and communicative competencies.¹³⁹ Here, the first competence consists in being able to generate well-formed linguistic entities; the second consists in embedding grammatically well-formed entities in relationships with reality and in using them in this way for purposes of representation. For my reservations about relying on a theory of competencies to conceptualize the theory of meaning, what is decisive is that an ability to follow rules underlies both aforementioned competencies: namely, those rules that must be followed in order to generate well-formed linguistic entities and those that must be followed in order to successfully engage in communicative action within an intersubjectively shared context of norms of action, values, and conventions. However, the concept of rule competence that is taken up here raises the question of whether a strategy based in theoretical issues of competence, which makes explanatory use of rules, does not presuppose far too much in order to avoid an infinite regress in the theory of meaning. For after all, those who not only have a causal disposition to behave in a certain way in certain situations, but also can follow rules, must already have means to understand the rules, must already master an (interpretation) language in order to be able to follow rules. Should the mastery of the (interpretation) language, for its part, be based on a rule competence, then the regress would be inevitable; consequently, the function of the theory of meaning as a normative basis for the theory of rationality would be jeopardized.

Regarding this problem of regress, however, Habermas seems to be hopeful that a developmental pattern—analogous to the moral development—in which the rule competence evolves as a product of an “evolutionary succession of stages of communication” can end the regress.¹⁴⁰ But to me it seems questionable how such an evolutionary succession can be shown to be plausible if a step has to be constructed within the succession to indicate the transition from a speech practice that is stochastic to one that is constituted by rule competence. For this step would presuppose that, at the preceding stage, a language had to be developed with the help of which the rules of the developed stage could be formulated and interpreted. In short: at the evolutionarily lower stage, the language would already be presupposed; without its help, the conventions of meaning and the rules of the next, higher stage could not be agreed on. If, for the sake of testing this, one here assumes the perspective of an evolutionary paradigm, then, in reference to Chomsky’s reflections, it will be possible to clarify what the problem of an evolutionary transition consists in. As is well known, Chomsky supposes that behind the diversity of natural languages there is a species-specific generative capacity, and the grammars of natural languages ought to be understood as individuating manifestations of this. He does not hesitate to characterize the structures that provide this generative capacity as a “mental organ.”¹⁴¹ In a perspective like this it is no surprise if the biological structure of organisms and—by virtue of their reproductive history—their organs are viewed within the framework of evolutionary theoretic explanations as a substrate of evolutionary processes. It is only questionable how we are able to move from a structural-functional description of these organs to a description of a rule-integrated practice. No clarity is gained regarding the problem of regress by introducing an oxymoron like a “mental organ.” The elegant connection of the mental and the functional or of the physicalistic vocabulary ignores the explanatory deficiency; it does not overcome it. For if we assume that actors follow rules when acting, we implicitly always presuppose that, in principle, actors always have the possibility not only of explicating rules that guide action, but also of breaking them. The language in which actors formulate the rules that they follow provides for the possibility of changing, suspending, or breaking rules. However, the research specifically in the area of generative grammar points out that this is not the case for the assumed species competencies. The “mental structure” innate to humans, which Chomsky irritatingly characterizes as rule systems,¹⁴² only makes the development of individual grammars possible within certain boundaries. But these fundamental programs need not be understood in order to be effective; nor is it possible to elude them when developing a language.¹⁴³

If my argumentation is solid and it is not unproblematic to use the rule concept for a theory of meaning, yet at the same time it is also impossible to defuse it with evolutionary theory, then it remains to be considered whether it is possible within the framework of Habermasian theory to substitute another, less problematic concept for it. However, this strategy is not simply there for the taking; for the close connection between the Habermasian theory of meaning and the concept of a rule is not due to a theoretical confusion. It is rather the result of a decision to join understanding, and consequently the theory of meaning, to conditions for acceptability, which for their part are connected to intersubjective normative contexts integrated by rules.¹⁴⁴ If strategies from evolutionary theory do not provide a plausible way to end the regress, and substituting the rule concept is not an option because of the intimate connection between meaning and validity, then the task arises of developing a “theory of meaning without a (presupposed) meaning,” and consequently a theory of meaning that is able to explain meaning without taking up a vocabulary that implicitly presupposes a concept of meaning, be that in the concept of meaning something in particular, of intending, or of rule following. Donald Davidson posed this task with an impressive radicality. Before I attempt to extract from his texts a concept of rationality that is based on a thinner theory of meaning than exists in Habermas’s view, I want to give voice to another objection to Habermas’s ambitious theory of communicative action, namely, to one that is based on Stefan Gosepath’s theory. Gosepath thinks especially that Habermas’s attempt to wrench moral norms from the conditions for understanding is exorbitant; he contrasts it with a view of instrumental reason reminiscent of Hume’s.

1.4.1.4. Rationality as Reflected Smartness (Klugheit)

In his detailed and extremely precise work, Aufgeklärtes Eigeninteresse—which I would like to introduce here as an example of a theory of rational interest—Gosepath¹⁴⁵ combines a reconstructive procedure that secures intersubjectively shared ideas by analyzing our use of the concept of what is “rational” with a constructivist element regarding the characterization of rules, which if followed, ought to guarantee rational action. The delineated rules take on the character of a normative construction insofar as the definition of rational action does not emerge solely from the analysis of our use of the concept, but relies on systematizations and reductions that prepare the ground for a criteriological interpretation of rational action. Gosepath emphatically argues for a concept that explains rationality in reference to justification. In doing so he advocates the internalist thesis that we can only judge the rationality of an activity in reference to the process of generating the activity and in reference to a person’s motives, goals, and state of knowledge.

Gosepath pursues his project of developing a robust concept of rationality in two steps: first, he is led by the presupposition that “the question of the meaning of the term ‘rationality’ . . . can only be answered by analyzing the way the word ‘rational’ is used.”¹⁴⁶ In connection with this analysis, he attempts to show “that there is a uniform meaning of the word rational or reasonable” that can be pronounced “well-grounded” for all situations in which it is used.¹⁴⁷ In a second step he then evaluates whether the concept of being “well-grounded” allows for the formulation of a robust rationality concept.¹⁴⁸ Gosepath hopes to find the answer to this question in a procedure that is sketched out as follows. As candidates for x in a predication of the type “x is rational,” he first tests beliefs, actions, desires, goals, norms, and expressions, whereby beliefs and actions are the main candidates; at the same time a differentiation between theoretical and practical rationality is achieved.¹⁴⁹ If x represents persons, institutions, and so on, then one is speaking of rationality in a derivative sense, which can be referred back to the rationality of actions. For both theoretical and practical rationality, rationality is understood as “the following of internal production rules (for one, for beliefs; for another, for actions), with results that are internally justified.”¹⁵⁰ The result of the research aimed specifically at beliefs and actions is, on the one hand, a rule principle of theoretical rationality, which operationally defines rational beliefs on the basis of requirements for belief formation:

It is rational for S to hold the belief that p, at time t, if and only if this belief emerges from or is maintained by particular rules for the formation of beliefs.¹⁵¹

On the other hand, it is a rule principle of practical rationality that formulates requirements for a rational choice among alternative actions:

An action is rational for S at time t if and only if the action is carried out on the basis of a practical consideration of S regarding t concerning what the best thing to do is in the situation in which the person involved finds herself, and this consideration is made according to the rules of practical rationality.¹⁵²

The content of these models of justification can now be specified using rules of rational belief formation. For theoretical rationality, Gosepath provides rules of theoretical rationality that I would like to characterize as smartness rules (Klugheitsregeln). Besides demanding the maximization of true beliefs and the minimization of false and irrelevant beliefs, these rules require the consideration of all sensible hypotheses, the cautious weighting of the evidence, and the choice of the best hypothesis with reference both to the intentions of the person reflecting on the matters and to the expected epistemic uses.¹⁵³ Against the background of these smartness rules, the judgment about the rationality of beliefs can nonetheless come to differing results. All rules remain relative, for example, regarding background presuppositions. This limitation initially corresponds to the conditions to which we connect the evaluation of the rationality of beliefs; for we would not allow beliefs based on views that the believers are not epistemically accountable for to be drawn on in judging their rationality. Because of this, the question is whether, beyond the relative sense of rationality, a stronger “absolute concept of rationality” can be developed. To clarify this question, Gosepath first differentiates the relativity of the rationality of processes of belief formation: he posits a relativity regarding the cognitive starting point of a person and a relativity regarding the rules of rationality. However, on the basis of the investigation of the relativity of the existing beliefs, it is necessary to analyze the relativity of rules; for “the cognitive starting point that, with regard to beliefs, is considered rational, i.e., justified, can . . . only be legitimated relative to the validity of certain rules, methods, laws, and goals.”¹⁵⁴ But regarding the relativity of rules, three versions are possible. Under the presupposition of a starting point, we can judge the rationality of belief formation with reference to

a. the rules accepted by a person (the radically subjective variation);

b. general rules (the radically objective variation); or

c. rules that the person has accepted after careful consideration (the reflective subjective variation).

If we apply one of these variations to our judgment, then at the next step the problem of substantiating the rules poses itself: if it is assumed—as in the radically subjective variation—that the rules are in principle dependent on the subject or the context, then a relativistic stance is taken that, however, itself would have to be relativized; consequently, it is hardly possible to explicate precisely what the position says. If, in the face of the performative contradiction that radical relativism becomes entangled with, it is pointed out that relativism is also subject to the laws of logic, then the objective variation comes into view. However, with recourse to the rules of logic (the principle of contradiction), only a subset of rules—namely, pure formal rules—are determined. Gosepath thus suggests taking the reflective subjective rules as the basis.

On the basis of an anthropological constancy—that all people have goals¹⁵⁵—theories, and thus systems of beliefs, can be judged as contributions to the solution of problems. The criterion that Gosepath thus proposes is once again a comparative one. It states “that those standards are rational that . . . more successfully solve problems than competing standards.”¹⁵⁶ With this proposal, three things are able to be achieved: First, it can be maintained that the critique of a concrete problem-solving attitude takes up a motive of the actor—namely, her interest in solving the problem; and this interest must, second, at the same time also include an interest in improving the standard. This also ensures that the standards are open for learning and critique so that the historical development of standards of rationality is compatible with the criterion. With problem solving, as a tertium comparationis, it is possible in a cautious manner, third, to provide for cultural invariances in the standards of rationality, and to avert the danger of radical relativism. Nonetheless, “rational reasons . . . thus remain subjective reasons, which, however, must be criticizable, but internally, that is from the standpoint of the person that is being criticized.”¹⁵⁷ In summary, the following can be maintained:

The rationally formed belief is, after careful consideration by the subject, the most effective means by which to reach the chosen goal.¹⁵⁸

Insofar as Gosepath’s proposal in the area of theoretical rationality connects the rationality of beliefs to goals that are chosen, the problem of justifying standards of rationality is now transferred to the possibility of justifying goals, desires, and interests. Gosepath devotes himself to the investigation of goals, desires, and interests in his considerations of practical rationality; it, analogously to theoretical rationality, is differentiated into a relative practical rationality, which presumes the given goals, and an absolute rationality, which could serve to justify goals, desires, and interests. For Gosepath’s internalist concept of rational action, two aspects connected to Davidson’s work are central:

There are two ideas that are implicit in the concept of acting on reasons: the idea of a cause and the idea of a justification. A reason for action is a justified cause.¹⁵⁹

This conception of rational action presupposed, a person (P) acts rationally, according to Gosepath:

1. If, on the basis of the belief that her goal Z can be reached by means M, P chooses that action that employs M to achieve Z.¹⁶⁰

2. If, in a situation in which the means to achieving a goal are otherwise of equal value, P chooses the action that achieves the goal to the greatest degree by the least difficult means.¹⁶¹

3. If, in the case where the unintended disadvantages of an action surpass its expected uses, P refrains from the action.¹⁶²

4. If, from among all the means to reaching a goal that come under consideration, P chooses the one “that infringes on other goals (in quality and quantity) least—above all if those are higher goals—and that fosters other goals most or fulfills them at the same time.”¹⁶³

5. If, when there are incompatible goals, P chooses the means that make it possible to reach the greatest number of P’s goals “in line with how desirable they are.”¹⁶⁴

6. If P orders the goals hierarchically in such a way that they are “complete and consistent.”¹⁶⁵

7. If, in her attempt to organize the goals in a certain time period, P chooses the plan that achieves her highest goals or the same ones to the greatest degree by the least difficult means, and where if P has another plan that can achieve all the goals of the earlier plan and at least one further goal, P prefers that plan.¹⁶⁶

8. If, when goals are roughly of equal value, P carries out those actions or pursues those plans that maximize the probability of success, and when there is the same probability of success, the result will be more highly valued.¹⁶⁷

9. If, when there is lack of surety about her goals and the way they are to be achieved, P initially makes plans such that all the possibilities remain open, but in doing so only reflects on the matter as long as the use of the consideration is worth the trouble.¹⁶⁸

Relative practical rationality can thus be understood as compliance with smartness rules, which, for given goals, guarantee the most effective, most useful, least disadvantageous, and most conflict-free realization of interests, and, where there is lack of surety about the goals, attempt to prevent the chances for achieving interests from being limited.

Subsequent to the investigation of instrumental (or goal-relative) rationality, the question is whether, beyond justifying actions relative to the desires and objectives of the actors, it is also possible to justify desires and objectives themselves. Reflection of this kind is in a certain sense already connected with every practical decision insofar as the actors have to choose between competing goals.¹⁶⁹ However, because the criteria that can be used for such reflections must in the final analysis always be found by the individual actors, and in this respect they remain goal relative, for Gosepath norms have only hypothetical validity. For: “The final authority for justification consists in the freedom and the interests of the individual.”¹⁷⁰

According to the internalist interpretation, it is only rational for an actor to follow a norm if she has a reason to follow the norm, namely, the desire to follow the norm. Because Gosepath interprets norm-compatible action as a special case of teleological action, the relationship between the individual and the norm is in the final analysis subject to the rules of relative practical rationality; thus, rational action conforming to rules cannot be traced back to structures of duty but solely to the interests of the individual that affect motivation. Rationality is thus the instrumental ability to orient decisions on the effective realization of goals.

Every type of rationality consists in the decision about the most effective means to reaching certain goals. That also applies to the goals themselves, which can only be justified in reference to higher ones.¹⁷¹

Indeed, Gosepath’s argumentation entails a weighty presupposition, namely, individual freedom; one can ask whether this freedom has presuppositions of its own that might serve as the basis for justifying categorical norms. For if we understand freedom as the possibility of individuals to choose between alternatives,¹⁷² then the existence of such alternatives could be viewed as a value that takes a particular position in reference to the freedom of the individual, a position, namely, that is characterized both by the fact that the existence of choice is a general constitutive condition of rational action and that it is also a result of realizing rational competencies. Because these action alternatives are of necessity alternatives for the acting persons, these alternatives do not exist independently of the competencies to individually specify the possibilities; consequently, we can postulate that, independently of the interests of individuals, it is necessary to maximize the possibilities for action and description. One can object that in the last analysis the number of these possibilities only measures the factual interests of individuals. However, insofar as the interests only come into purview from the perspective of the competencies of persons to individually individuate these in a process of self-interpretation, it is possible, relying on rational conditions of autonomy, to incorporate processes for generating the means for self-interpretation into the characterization of rationality and to understand the development of these as a norm that is independent of interests. We could thus investigate the conditions for the possibility of choice and connect a substantive concept of rationality, rich in content, to the ability to maintain or extend the conditions for choice. Gosepath, however, does not pursue this course. He ends his reflections with a rather sobering assessment:

I . . . do not see how the concept of rationality can be coupled with a conception of the good life by recognizing certain substantial claims about the choice of human goals and the right form of cooperative life together. Because we are, in my view, forced to reject such more substantial conceptions, there remains only a rationality concept that resembles the Humean one and that is so formal that it does not allow the rational or irrational ordering of norms and values, societies, and worldviews.¹⁷³

Even if one shares this assessment in reference to the “true goal of humankind,” Gosepath may draw consequences that are too broad insofar as he decouples certain forms of social interaction and consequently, in a broad sense, forms of cooperative life that are necessary for the development of the competence of self-understanding, from the concept of rationality. However, it would at the very least be strange if the rationality concept was neutral about the conditions for the possibility of self-understanding, even if we must assume that these conditions must be assumed to be at least partially fulfilled in order to speak of rationality.

If, as Gosepath himself maintains, rationality is connected with autonomy insofar as it is the final authority for justification, then the question emerges regarding how this connection can be made plausible with the help of a rationality concept empty of substance. For if the procedural norms that explain and ground rationality have the character of hypothetical imperatives, there would only be a connection if the individuals have an interest in autonomy. At the same time it appears that we can only speak of actions if we presume that the actors are at least partially autonomous; with this, however, a conceptual and not merely contingent connection between rationality and autonomy seems to emerge. Rationality is not simply an instrument of autonomous persons, but a condition for their autonomy because self-determination requires a relationship to the self that can only be fulfilled by actualizing rational competencies. In short: rationality is not only an instrument relative to the interests and the autonomy of individuals; rather, it is a set of competencies that allows this autonomy and the interests to be individually specified in the first place. However, on the basis of Gosepath’s reflections, it is not possible to theoretically secure these views. Because Gosepath views the predicate “rational” as coextensive with the predicate “well-grounded,” he cannot interpret competencies that are initially competencies of (self-) understanding, and not competencies for justification, as an aspect of rationality.

After these, in part, rather tentative remarks, in closing I would like to point to a difficulty related to the explanative structure of Gosepath’s theory. Because Gosepath explains rationality in all nonderivative contexts as being well-grounded, it is not clear offhand how it is possible to avoid a circular explanation of what it is to be well-grounded by simply reverting to rationality. Amazingly, Gosepath treats this objection, which Schnädelbach had formulated in the context of the discussion of justification-centered rationality concepts, in a footnote in which he gives one pause to think that being “well-grounded” can be explained with recourse to procedural rules; moreover, explanations that draw on the rationality that they explain do not make the content of the explanation circular.¹⁷⁴ His reply thus looks more like a defense against an argument about self-application than a reaction to the diagnosis of circularity. Because the suspicion of circularity does not only apply to Gosepath’s theory and is of systemic interest, in the following remarks I will deal with it in more detail.

1.4.1.5. Intermediate Remarks: Rationality and Justification

The intimate relationship between rationality and justification that is drawn on by nearly all of the rationality theories outlined here—especially, however, by the theories of Habermas and Gosepath—was subject to critical examination by Schnädelbach in his article “Rationalität und Begründung.” Schnädelbach first shows that “fixating on the justification model of rationality leads [to] holding everything to be irrational as long as it is not fully argumentatively or discursively justified.”¹⁷⁵ With this, however, the “scope of the irrational would be broadened to gigantic proportions,”¹⁷⁶ and, in addition, such an explication of rationality promotes a scientistic view of reason. To counter this truncation of the concept of rationality, which identifies rationality with the ability to know or act for reasons, Schnädelbach brings three arguments to bear.

First, it can be objected that reason is not exhausted by the ability to answer why-questions. Besides the ability to justify knowledge or action, there are also rational abilities that are not mere justifications. Schnädelbach mentions the ability to engage in reality testing, as discussed in Freud; the ability to learn from mistakes and errors, as discussed in Popper; the ability to solve problems in nexuses of action that contain feedback processes, as discussed in Gehlen; and the goal-oriented choice of means, as discussed in Weber.¹⁷⁷ Second, rationality concepts oriented on justification must proceed from a pre-understanding that, for its part, “is much too complex to be fully exhausted by a justification concept.”¹⁷⁸ In relation to this pre-understanding, the explication of rationality with the help of a justification concept is just one among various alternatives, and it must show itself to be the most preferable alternative.

Whether these objections are strong objections can be rightfully doubted, for the competencies that Freud, Popper, Gehlen, and Weber bring into play may not completely be worked into justification competencies, but it does not seem very plausible that competencies for reality testing, learning, and (instrumental) problem solving could be characterized as these competencies without relating these to justifications; a pre-understanding is not sacrosanct.

More serious than these rather bland objections is, however, the third objection, namely, that rationality and justification stand in a circular relationship to one another in a rather strict sense insofar as justifications “themselves [must] be qualified as rational in order to be able to exemplify rationality.”¹⁷⁹ If we are confronted with the fact that someone has reasons for a rationality concept that is based on justification, this by no means implies that those reasons are good ones. But what good reasons are depends on the criteria used to characterize good reasons, typically on criteria as they are developed in a theory of rationality. If one now investigates possible reasons for a concept, it is possible to differentiate between material and formal reasons. Here it immediately becomes clear that material reasons that may be true, suitable, cogent, plausible, etc., are not sufficient, suitable candidates for justifying a concept of rationality or even for comprehensively characterizing rationality; for irrationality does not result from material falsity.

What is false is not irrational, because only something rational can be false; otherwise it is simply senseless. Those who are mistaken are thus not irrational; but only rational beings can err.¹⁸⁰

So after the recourse to material “qualities of reasons” has failed, because what is to be determined with their help “itself already must be rational in a broader sense of the word,”¹⁸¹ there remains only the examination of the formal characteristics of reasons. The formal qualities of reasons, such as correctness and stringency, that pertain to the relationship between reasons and justifications, remain—excluding material reasons—related to the framework of hypothetical validity. Formally good justifications are then those that satisfy the particular procedures of formal justification, as they are developed in formal logic or by discursively satisfying the validity claims raised. With regard to discourse theory, however, it can be shown that the rules and procedures according to which validity claims can be examined themselves could become the object of a rational discussion according to rules and procedures. We must thus, in connection with the attempt to determine rationality in reference to formal qualities of justification, “be satisfied with a procedural explication of reason, which, in addition, is itself in principle fallible.”¹⁸² According to Schnädelbach, there is no way to bypass the circular consequence that the explication of what is rational on the basis of the formal qualities of justifications presupposes rules and procedures that are fit for the task precisely because they are rational. In short:

Explications of “rationality” with “justification” are . . . thus necessarily circular because justifications themselves must always already raise a rationality claim in order to be identifiable as justifications.¹⁸³

In light of the fact that the discussion here is not about some tolerable form of circularity,¹⁸⁴ but it is claimed that this is a basic objection to an explanative strategy, it is irritating that Gosepath, for example, who must be directly vulnerable to the argumentation, does not appear to be particularly impressed by these considerations. And perhaps it is in fact not really completely clear how Schnädelbach’s text should be understood. Two interpretations are conceivable. One interpretation could be called a “self-imputing argument”; another could be characterized as a “strong circularity argument.” The self-imputing argument essentially means that if P wants to characterize rationality in reference to justification procedures J, then P must impute to himself that this explication itself is sufficient for J. Gosepath appears to understand the circularity argument in this or some similar sense, and in this form it in fact is more of a diagnosis of a peculiarity of any explanation of rationality than a fundamental problem. Schnädelbach could, however, also have the following much stronger argument in mind:

(P1) Any nonderived case of the predicate “is rational” can be replaced by the predicate “is well-grounded.”¹⁸⁵

(P2) Sentences of the type “X is well-grounded” are true if and only if there is a good justification for X.

(P3) Good justifications are based either on

a. material reasons for X or

b. formal reasons for X,

c. and there are no other kinds of reasons.

(P4) If reasons that are drawn on to justify X are materially false, it follows that X is not well-grounded, but it does not follow—in contrast to (P1)—that X is irrational.¹⁸⁶

(K1) Therefore: (in order to be able to accept [P1]) it is necessary to determine good justifications with the aid of formal reasons.

(P5) Formally good reasons are reasons that satisfy certain procedures of formal justification.

(P6) Procedures of formal justification are fallible.

(K2) Therefore: to characterize good procedures of formal justification, rationality criteria are needed—which amounts to saying that (P1) is circular.

This argument is not based on self-application, but it shows that the explanative resources that are drawn on to define rationality are precisely the resources that have to be characterized with the aid of the rationality concept; it is not a diagnosis of normative, but of explanative circularity. Something is explained with the aid of something else that can only be identified by presupposing the explanation. On the other hand, because it can hardly be disputed that there is a relevant connection between rationality and justification, the question arises as to how one should react to the argument. It is clear that one cannot avoid this circularity by exempting the particular instance of explication from all assumptions of rationality; for if we want to explain rationality, we must impute rationality to ourselves. Here, however, there is a significant limitation: we must not identify the particular instance of explication with rationality, because if we do there is no explicans independent of the explicandum.¹⁸⁷ The elucidation of what rationality is must not presume a comprehensive definition of rationality, but must limit itself to presuming rationality in some contingent pre-understanding:

It is not rationality in general, but rationality in a finally contingent interpretation that is the a priori that we encounter in the first person perspective as the highest or deepest point of our philosophical self-assurance.¹⁸⁸

Schnädelbach attempts to avoid the narrowness of a concept of rationality based solely on justification and its entanglement in circularity by interpreting rationality as an “open concept” that “cannot be completely explained for all contexts,”¹⁸⁹ because each thematization of rationality, for its part, can also be thematized without thereby constituting a threshold for self-reference. This, in my view, defensive approach appears strategically to be motivated, among other things, by the attempt to deal with the radical critique of reason, which aspires to unmask a concept of rationality that is oriented on justification as a scientistic truncation of reason.¹⁹⁰ A danger of this procedure consists in the fact that this “openness” passes over into indetermination and then implies that we accept something as rational that is incompatible with our respective standards of rationality; for nothing can appear to us to be rationally demanded that is simply incompatible with our de facto rationality standards. If, in addition, the openness is characterized with reference to the view that reason cannot be explained completely “for all contexts,” that is, by characterizing it in a manner somehow related to Kambartel’s skepticism about general criteria of reason, then one has to allow the question how it is possible to identify reason in the various contexts.

Schnädelbach makes it clear that “rational” must not be identified with “well-grounded” and that, instead of that, we should be on the lookout for an explanation of rationality that does not lead to explanative circularity. Because we are nonetheless to maintain the close connection between justification and rationality, rationality must be interpreted as something that is more comprehensive than justification competencies. Along with Schnädelbach, I thus conclude from the previous argument:

(K3) Therefore: there are rational competencies that are not justification competencies.

Schnädelbach, however, only implies what these comprehensive rational competencies might consist in, and, in particular, the problem remains whether we could accept the idea that these competencies, which, taken together, are supposed to explain the concept of rationality, constitute a collection of the most varied abilities; for if we accepted this idea, we would not be able to say why precisely these competencies, taken together, ought to characterize reason. I would now like to suggest that we count among the postulated competencies of (K3) all of those that make understanding possible and, among these, especially competencies for understanding that are not characterized by reasons. In my view these are forms of understanding that occur at a nonpropositional level; understanding of this sort need not be accounted for by recourse to reasons. Schnädelbach has also affirmed that if we explain rational competencies as the individuation of mental states with content, then competencies of nonlinguistic articulation are among the things that a sufficiently refined theory of rationality has to explain.¹⁹¹ In chapter 3 I will make an emphatic attempt to show the validity of this view on the basis of a theory of media and nonlinguistic understanding. First, however, on the basis of Davidson’s reflections, I will attempt to clearly portray what it means to explain rationality against the background of processes of understanding.

1.4.1.6. Rationality and Understanding

Donald Davidson’s thoughts about the theory of rationality start, like Habermas’s investigations, from problems of the theory of meaning, but the basic concept that Davidson’s thoughts center on is not the concept of communication (in German, Verständigung, which is variously translated as communication, comprehension, or reaching a mutual understanding)¹⁹² but that of understanding (in German, Verstehen). Davidson, unlike Habermas, does not presume antecedent instances of a symbolically mediated lifeworld, but sets out from the “original setting (Ursituation) of the more recent analytic philosophy,”¹⁹³ which became well known through Quine’s Word and Object as the situation of radical translation.

To put oneself in this position means first of all to base the problems of understanding on considerably slimmer premises, for Davidson is not concerned, like Wittgenstein or Habermas,¹⁹⁴ with explaining “understanding” within the framework of an existing rule-bound communicative practice, but with making it plausible that understanding is possible under more adverse conditions in which neither language nor a form of life in the stricter sense need to be shared by speakers and interpreters. However, in contrast to Quine’s empirical student of language, who is entrusted with compiling a handbook of translation for the “translation of the language of a hitherto untouched people,” in which “all help of interpreters is excluded,”¹⁹⁵ Davidson starts from a situation in which the interpreter is no longer able to trust the “ordering power of language,” which is supposed to enable him to access what is meant.¹⁹⁶ In this situation of radical interpretation, the concern is not with producing a network of lexical relations between expressions of different languages that are already understood as ordered systems, but with simultaneously unlocking what the speaker believes and the meaning of his expressions.

From the analysis of this situation, Davidson hopes to be able to track all the presuppositions that must be fulfilled so that interpretation, i.e., understanding, is possible,¹⁹⁷ and one of his central insights is that a universal assumption of rationality is among these inevitable presuppositions; this has achieved renown as the principle of charity.

Before I more precisely present the situation of radical interpretation, as well as the consequences for interpretation and rationality theory that Davidson draws from his analysis, I would like to make it clear that Davidson’s strategy is not simply a result of the preference of analytic philosophy to work off of artificial situations that are tailored to one’s purposes—precise but irrelevant. It does not result from a philosophical style, but from his wide-reaching diagnosis that most relevant interpretive strategies are bound to fail. According to Davidson’s analysis, the basic problem, which cannot be solved within the framework of established theories of meaning, consists in the fact that, in order to formulate the theory, all of these theories have to make use of a concept of meaning that is only available as a result of these theories. According to Davidson, this includes any theory that bases its explication of the concept of meaning on specific intentions of the speaker, be these in the form of direct reference to communicative intentions (Grice) or in the form of an analysis of the research on social action contexts (Wittgenstein, Austin); for the reference to realistically interpreted specific intentions of speakers assumes that the theory of meaning is not needed in order to introduce the semantic vocabulary theoretically. However, if one presumes that actors or speakers have intentions independent of their self-interpretations and independent of the interpretations of others, then one makes use of semantic relations such as believing and interpreting as basic concepts of the theory of meaning, which are supposed to explain these semantic concepts in the first place. Davidson finds three basic maneuvers to be promising in this situation.

1. He advances an antirealistic theory of meaning, which avoids bringing intentional vocabulary into play when identifying what is understandable. The starting point of the theory of meaning, which in Davidson’s view is a subdiscipline of action theory, is expressive events, consequently sounds caused by (human) organisms that are physically describable, which are then described in the theory of meaning as a specific type of action. Davidson thus does not expect expressive actions to disclose themselves in the horizon of a lifeworld, but views expressive actions as products of a descriptive and interpretative perspective, which makes expressive events part of an intentional description. In the framework of this antirealist perspective, meanings have the same ontological status as interpretations; they are nothing other than products of the (self-)interpretation of expressions that can be reconstructed within the framework of a nominalist ontology.

2. He gives the theory of meaning the character of an empirical theory, which, on the basis of the elementary characterization of the concept of meaning, as a product of interpretation, applies the concept of meaning as a theoretical concept and does not attempt further elucidation from more conceptual analysis. As a result of the empirical status of the theory of meaning, the following questions play a fundamental role in this:

a. Which knowledge would enable the interpreter to understand expressions of the speaker of a language completely foreign to the interpreter?

b. How can an interpreter acquire the knowledge necessary for the interpretation?

A Davidsonian theory of meaning thus has the task of explaining how finite beings are in the position to understand a potentially infinite number of expressions; it has the task of identifying finite knowledge that places speakers and interpreters in the position to understand an infinite number of expressions. Davidson entrusts formal semantics with the task of identifying this knowledge.

3. According to Davidson, formal semantics offers a theoretical framework, with the help of which the knowledge needed by speakers and interpreters is able to be represented as a form of finite knowledge; for in connection with the work of Frege and Tarski, by means of formal semantics, it can, for one, be shown that the meaning of all possible expressions is dependent on a finite number of attributes of the constituents that are used in expressions, and, for another, it can be shown how the meaning of expressions can be determined with a view to the contribution of the expressions to the meaning of the propositions that they form. In Davidson’s view, to understand expressions as sentences that are formed by units and correspondingly to determine the meaning of components of sentences with recourse to their function for the meaning of a sentence both hinge on the assumption of truth-conditional semantics, that sentences can be described as truth functions.

In order to more precisely determine the place of the Davidsonian rationality theory, it is first of all necessary to bring to light the central implications of the above orientation and to reconstruct their systematic links with one another.

From what has been said, it follows that Davidson understands a theory of meaning of a language L to be an empirical theory that provides the truth conditions for sentences and expressions in L. This theory is formulated in a metalanguage whose objects are the sentences and expressions of the object language that is to be explained. With a view to the issue of whether something is able to be learned, the theory of meaning must present the truth conditions of expressions in L in a finite form. Davidson meets this demand by structuring the theory of meaning axiomatically. A theory of meaning for a language L first of all consists in

(TM)

a. a finite number of axioms that provide the semantic function of the expressions of the object language in which they, for example, determine their extensions;

b. a finite number of rules for combining expressions; and

c. a finite number of rules, with the help of which theorems can be deduced that provide the truth conditions of object-language sentences.

This is made more precise if, against the background of Davidson’s nominalism and his diagnosis of the implicit circularity of established theories of meaning, one asks which form the theorems may assume. It is first of all clear that the typical form for the specification of meaning

(Th₁) s means m

cannot be the form of the theorems for the theory of meaning because it is incompatible with Davidson’s nominalism, the view that truth has theoretical primacy over meaning, as well as the diagnosis of circularity. For (Th₁) allows, for one, that we treat meanings like entities, for example, if we replace m with singular terms, which are supposed to refer to meanings; however, for another, it does not make it possible to relate meaning to truth conditions, nor can it be formulated without the two-place predicate “meaning,” which is what one is supposed to be explaining with a theory of meaning in the first place. It is clear that the criticism applies both to theories of communicative intentions (Grice) and to pragmatic theories of meaning (of Wittgensteinian or Habermasian provenance); for specifications of meaning within the framework of these theories take on the form of (Th₁). Therefore, as a form for the specification of meaning in an intentionalist theory of meaning,

(Th_1i) s means m because A means m if A expresses s

must also be precluded from being a candidate for the theorem of a Davidsonian theory of meaning, and

(Th_1p) s means m because there is a rule in the speech community to which A belongs that says that s, in the situation in which A expresses s, means m

must be precluded as a form for specifying the meaning in a pragmatic theory of meaning. As a further candidate, Davidson thus also tests the theorem schema

(Th₂) s means that p,

which at least has the advantage of containing something like an implicit relation to truth conditions. The problem with (Th₂), however, consists in the fact that implementing it would involve us in “intensional springs”;¹⁹⁸ for the sentence-forming sentential operator “means that” spans an intensional context in which coextensional expressions cannot be exchanged salva veritate. If, however, as a result, the truth-value of the theorem of the theory of meaning can be false just because an intensional predicate is used, then (Th₂) cannot be the suitable form for the theorem of the theory of meaning without having consequences for the empirical status of the theory of meaning. Davidson’s own suggestion thus must do without intensional predicates, and it must simultaneously bring into play a systematic connection to truth conditions. His suggestion is:

(T) s is true if and only if p.

This results, first of all, in the following advantage: a T-theorem is formed with the aid of the pure extensional sentential operator “if and only if,” and a T-theorem produces a relation between two sentences that are capable of truth-value, namely, the sentence that is formed from the metalinguistic name “s” for a sentence in the object language and the predicate “is true,” as well as the sentence p, and it is thus in a form that satisfies the core requirement of truth conditional semantics. However, if a theory of meaning consisted solely in T-theorems, this would immediately result in a difficulty, namely, that the schema (T) produces an infinite number of theorems, which would conflict with the demand that the theory of meaning is to specify finite knowledge that must be available to the speaker in order for him or her to be able to speak and understand a language L. In view of the necessity that interpreters must be in a position to deal with a potentially infinite number of expressions, this characteristic of the schema (T) cannot be waived. The demand for finitude must be reconciled with other components of the theory of meaning. Davidson complies with this demand by designing the theory of meaning in a form corresponding to (TM), which shows for a language L “‘how the meanings of sentences depend upon the meanings of words’ if it contains a recursive definition of truth-in-L.”¹⁹⁹ In Davidson’s view, the model of such a theory can be found in Tarski’s theory of an extensional truth definition for formal languages. Tarski solved the problem for finite axiomatization by applying the method of recursive definition, i.e., by specifying all operations with the help of which complex sentences are formed from simple ones and indicating the way that the truth-values of the complex sentences are dependent on those of the simple ones. However, because halfway interesting languages contain not only sentences, but also open sentences (thus a bit of quantificational logic), a procedure is also needed that brings the truth-value of quantified sentences into connection with open sentences, which by virtue of the quantification become closed sentences. Tarski provides for this procedure by a recursive definition of “true-in-L,” which, for quantified sentences, makes use of a recursive definition of the relation of satisfaction (which can exist between open sentences and objects or a series of objects).²⁰⁰ However, while Tarski’s efforts aimed to provide an extensional definition of truth, Davidson assumes the concept of truth as an undefined basic concept in order to initiate an extensional theory of meaning.

The next step must now consist in clarifying whether the form of Tarski’s theory, which was developed for formal languages, can serve as a suitable model for the development of a theory of meaning for natural languages. In that case it would eventually be necessary to come to terms with the following difficulties: a procedure has to be found that copes with the particularity of natural languages, namely, that natural languages contain indicators or deictic expressions. Here, however, the truth-values of expressions containing indicators vary with the temporal and spatial coordinates of the expression. However, it is possible to accommodate this peculiarity of natural languages by reformulating the schema for T-sentences (with the aid of a three-place truth predicate) in the following way:

(T′) s is true in L for the speaker of s under the conditions of its utterance (i.e., at time t, at place o) if and only if p.

Besides these difficulties, which are easy to defuse, natural languages exhibit a series of further peculiarities such as unreal conditionals, adverbs, mass terms, belief sentences, imperatives, optatives, questions, and normative and modal sentences; for each of these, as Davidson himself sees,²⁰¹ it must be shown in detail how T-propositions can contribute to their comprehension. With a view to some of these particularities, Davidson’s theory of meaning must thus be seen as an open research project.²⁰²

An important characteristic of a theory of meaning, as an empirical theory, is that the T-theorems are empirically testable. For it is clear that a theory of meaning can only be a theory for the interpretation of the linguistic behavior of real speakers if there are criteria for testing the T-theorems. Davidson thus requires that the correctness of T-theorems is able to be confirmed by evidence that is accessible to interpreters with no previous knowledge of the language that is to be interpreted and that the way the T-theorems are interpretive, thus the way the expressions of real speakers can be interpreted with their aid, can be shown. For with a view to the problem of the interpretability of T-theorems, it must first of all be recognized that individual T-theorems are obviously too weak a basis for suitable interpretations insofar as a mere truth-value agreement between the object-language expression and the metalinguistic sentence provides too few restrictions for appropriate interpretations.²⁰³ With respect to the problem of evidence, Davidson admits that the truth of T-theorems cannot be solved independently of the problem of interpretation; for the evidence for T-theorems is evidence that the metalinguistic translation of an object-language sentence is correct. However, it is first of all clear that situations in which speakers express their sentences must serve as evidence for a T-theorem.

For both problems, however, there are solutions that have become known under the catch phrases of radical interpretation and holism. This is first defused by the fact that Davidson views T-theorems as lawlike sentences, thus as sentences that formulate causal law hypotheses. Here it is presumed that speakers think occasion sentences are true because there are causes in the world that affect (bewirken) whether speakers accept these sentences as true. However, insofar as the concern is with law-like connections, it can be precluded that expressions are connected with expressive situations in a contingent unique way, and interpretations are not subject to chance evidence.

The theory of radical interpretation, which starts out from a situation in which we cannot resort to knowledge of the language of the speaker or the help of a dictionary or translator, confronts the fact that for T-theorems of type (T′) we do not have independent access to the beliefs of the speaker and to the meanings of the expressions that she uses. In contrast, in order to determine those beliefs, we need assumptions about the meaning of the expressions, and, in order to determine the meaning of the expressions, we need assumptions about the beliefs. In order to cut through the knots of this reciprocal dependency, Davidson proposes that we set out from the content-unspecified assumption that the speakers think their sentences are true.²⁰⁴

With a view to mental states, holism asserts that these can only be individuated in a network of inferential relationships; and with a view to language it asserts that T-theorems are only interpretive relative to a net of linguistic expressions, i.e., in the context of a language and in the context of expressive situations. A theory of meaning must, however, also prove itself in that T-theorems must be deducible from it that are also true for sentences that contain indexical expressions.²⁰⁵

However, holism and radical interpretation also entail new problems; for against the background of the interdependence of meanings and beliefs, we now need criteria that allow us to make reasonable choices between competing interpretations. The principle of charity (PC), which Davidson views as a necessary presupposition for interpretation, is supposed to throw light on this problem. The (PC) requires that interpreters not ascribe absurd beliefs to the speakers, but rather ascribe beliefs that maximize the rationality of the speakers.

Why should we do that? In contrast, aren’t people often irrational? Perhaps, however, if we assumed that, we would not be able to understand them; for understanding presumes that we assume the conditions for understandability are fulfilled by what is to be understood. However, this amounts to saying that we assume that what is to be understood is that which we ourselves think is rational. For, otherwise, we could not integrate an interpretation into our system of beliefs. In short: what we ascribe to those we are trying to interpret must be a possible thought for us—something rational. So, what does the (PC) demand of us? In Davidson’s view, the following:

(PC)

(COH) Principle of Coherence: If you are interpreting a speaker, assume that he has the least possible number of contradictory beliefs (i.e., beliefs that you think are contradictory).

(COR) Principle of Correspondence: If you are interpreting a speaker, assume that the majority of his beliefs (according to your own standards) are true.

(COR) must be explained against the background of holism in the following way: Although holism asserts that the meaning of sentences depends on the entire system of beliefs, it is still necessary to preclude the possibility that such a system, while being coherent, is entirely empirically false. Occasion sentences thus are a suitable starting point for explaining the (COR). For they offer the only possibility that interpreters in the situation of radical interpretation can ascribe the speaker beliefs in a way that is intersubjectively accessible, namely, by referring to a causal relation to the world.

With the reference to the world Davidson brings the causal theory of reference of meaning into play in a way that he connected with an elementary learning situation in the later triangulation model. The triangulation model has, on the one hand, an antiskeptical function, which makes it possible for two beings in an intersubjectively accessible world to observe the correlation between expressions and the conditions for making expressions, and, on the other hand, the function of joining the theory of meaning to a process of language acquisition that is to be achieved intersubjectively. The language acquisition of a student is explained by the triangulation such that the student can create a relation between his reaction to an observable event in the world and the (linguistic) reaction of the teacher to this event.²⁰⁶ It is “a matter of two private perspectives converging to mark a position in intersubjective space”²⁰⁷ such that that in which the perspectives converge can be understood as a cause of a linguistic reaction. In the triangulation model Davidson attempts in a fruitful way to expand interpretationism, which in a situation of radical interpretation requires that the interpreter already have a metalanguage available with the help of which he can formulate T-theorems and evidence so as to suitably reconstruct processes of language acquisition. Here the social conditions clearly emerge through which beings develop mind.

Hence I believe there could not be thoughts in one mind if there were no other thoughtful creatures with which the first mind shared a natural world.²⁰⁸

With the triangulation model Davidson appears to have managed in a radical sense to connect the ability to think to social conditions; strikingly, he manages to do this without presuming that the ability to speak, which is the basis on which thinking is possible, can only be explained with recourse to a set of shared rules. Even though I do not think that the triangulation model accomplishes what it sets out to accomplish, namely, because its presuppositions are also far too excessive,²⁰⁹ I share with Davidson the belief that thinking is an ability that is acquired by internalizing social processes of interpretation.²¹⁰ In the place of rules there is, in the interpretationist perspective, a sufficiently stable linguistic practice of a speaker who is to be understood, on the basis of which hypotheses can be formed and tested. Thus for Davidson—in contrast to Habermas—knowledge of rules is not ascribed an elemental function for the possibility of understanding, but rather at best an economic one; for in general our willingness to continually produce complex analyses of linguistic behavior from scratch is quickly exhausted if it does not concern interpretations of things that we have a great interest in understanding.

The foregoing explanations should, in any case, have made it clear that an understanding of linguistic action is, in Davidson’s view, a process in which we ascribe intentional states to the beings that we interpret, which we understand as causes of linguistic action. Here, we presume that these states are propositional states and as such can only be individuated in a network of further propositional states. Because this network of propositional states is constituted by inferential relationships, the attempt to ascribe to a being a propositional attitude, which can function as a reason, implies that we ascribe rational organization to the network.

Thoughts, like propositions, have logical relations. Since the identity of a thought cannot be divorced from its place in the logical network of other thoughts, it cannot be relocated in the network without becoming a different thought. Radical incoherence in belief is therefore impossible. To have a single propositional attitude is to have a largely correct logic, in the sense of having a pattern of beliefs that logically cohere. This is one reason why to have propositional attitudes is to be a rational creature.²¹¹

In a manner of speaking, rationality is constitutive for having a mind. Herewith Davidson explicates a dimension of rationality that, for example, is misapplied in Gosepath’s instrumental conception, although it is among the necessary presuppositions of a mind that is supposed to be capable of instrumental considerations in the first place. If it is additionally supposed to be correct that having an individual propositional attitude implies the existence of a network of propositional attitudes and networks of propositional attitudes imply the implementation of rational relations, then it is possible to maintain that rationality in this sense is not something that can be demanded of a being. Rationality is rather a condition for the possibility of having propositional attitudes and—insofar as an antirealistic interpretation of propositional attitudes is our basis—for interpreting whatsoever.

In what follows I would like to more precisely investigate what it means for the concept of rationality to follow these thoughts. Here I would like to orient myself on the following questions: (1) If rationality is a presupposition for having propositional attitudes, to what degree and in what sense is rationality then still a normative concept, and what presuppositions must a being really fulfill so that it can really still be understood? (2) How is irrationality possible?

1. Davidson’s texts leave no doubt that successfully imputing rationality is a condition sine qua non for understanding. So, for example, he notes in “Radical Interpretation”:

If we cannot find a way to interpret the utterances and other behavior of a creature as revealing a set of beliefs largely consistent and true by our own standards, we have no reason to count that creature as rational, as having beliefs, or as saying anything.²¹²

However, how does one examine whether it is right to impute rationality? What are the detailed determinations of content or the criteria for rationality? By mentioning freedom from contradiction and truth, the quote indeed provides two criteria, but it is still not clear precisely which criteria a being must fulfill in order to be intelligible as a rational being. One may be suspicious of the circularity, above all, of the reference to the fact that an intelligible being must be largely free of contradiction, for if “largely” means more or less “sufficiently,” then it appears that it is not possible at all to explain freedom from contradiction independently of interpretability:

(RD) A being, B, acts rationally if and only if B’s action A is understandable, and A is understandable if an interpreter of the activity that constitutes A can describe it as an activity that B carries out for (largely good) reasons (intentions and beliefs).

This circularity, however, is not by chance, for a being acts rationally and is only understandable insofar as she acts in accord with our standards of rationality. However, are these the standards that Davidson notes in his articles? That is, standards like the following:

(RD′)

a. The principle of the requirement of total evidence: “give your credence to the hypothesis supported by all available relevant evidence.”²¹³

b. The principle of self-control or the principle of continence: “Perform the action judged best on the basis of all available relevant reasons.”²¹⁴

c. The principle of conservation: change as few expectations as possible if, under the condition of the constancy of some things, you begin to integrate recalcitrant things.²¹⁵

d. The elementary principles of decision theory (e.g., the transitivity of preferences).²¹⁶

e. The logic of sentence structure.

Hardly, for then could we not understand beings who, confused by the fact that some things in their surroundings are not as they expect, act pretty clumsy and draw on explanatory hypotheses that do not exactly correspond to the principle of conservation? After all, it is Davidson who tells us the (fictive) story of how he, after walking into his neighbor’s house, lost in his thoughts, attempted, with the help of numerous improbable assumptions, to explain why the furniture was so mad and why the neighbor was mixing a drink in his living room.²¹⁷ Through this or merely thereby is Davidson a being that we can no longer understand and thus who is also no longer rational? In the case of the continence principle, things may be different, for in a certain sense, part of the concept of best action is that one carries it out if one is ready to act. However, it is questionable even in such cases whether a being that does not follow, or only seldom follows, the continence principle is correspondingly incomprehensible.

The suspicion arises that, for Davidson, at least two levels of rationality are in play. On the fundamental level, rationality in a strong sense in fact characterizes the conditions for understanding; this appears to demand weaker criteria than those that are expressed in Davidson’s principles. Thus Alexander Becker has shown that in understanding language on the basis of a truth-conditional semantics, one can only assume the ability to follow modus ponens as a universal principle of fundamental rationality.²¹⁸ Rationalizing interpretations thus meet up against a limit of understanding if an interpreter is not able to suggest an interpretation of a being’s behavior that describes it as having intentional states that are largely organized in conformity to modus ponens. Only if this point is reached is it no longer necessary that we attribute the interpretation that a being is not fundamentally rational and that its action is consequently incomprehensible to the inability to rationalize the action.

2. Even if it is questionable whether Davidson in fact articulates fundamental conditions of understanding with the suggested principles (RD′), it is still clear that his reconstructive efforts are aimed at a constitutive dimension of rationality. However, against the background of these mentally constitutive conditions, the question of how irrationality is possible becomes all the more virulent.²¹⁹ Consequently, for Davidson the problem arises regarding how it is possible to ascribe something like irrationality at all in dealing with actors who are ascribed with propositional attitudes.

To explain irrationality we must find a way to keep what is essential to the character of the mental—which requires preserving a background of rationality—while allowing forms of causality that depart from the norms of rationality. What is needed to explain irrationality is a mental cause of an attitude, but where the cause is not a reason for the attitude it explains.²²⁰

Among the costs of Davidson’s philosophy of mind is thus that in cases in which we are confronted with genuine forms of irrationality—for example, when individuals carry on and knowingly act against their own preferences—we must presume the division of mind into two subnets that are further rationally structured internally. This consequence hardly appears plausible, for after all, the mental states of the subnets—according to interpretationist premises—can only be given content by the rational self-interpretation of the person who harbors these subnets. How then can the mind of a person be divided if the existence of the subnets, for their part, presumes a rational interpreter?

Under the premises of a thin theory of meaning, Davidson makes available a theory that promotes, with impressive radicality, the attempt to spell out rationality using the concepts of a theory of understanding. Even if his theory poses questions that, for example, in the case of genuine irrationality, suggest that, besides propositional attitudes, other mental states with content (other than perceptions, etc.) are also brought into play with content that is not dependent on interpretations, in the following reflections I would like to pursue the interpretationist perspective that Davidson was important in inaugurating. Because I do not see why such a perspective should remain limited to the analysis of linguistic behavior and the ascription of propositional attitudes, I will attempt to open it up to nonlinguistic communication processes in order to include dimensions for a theory of rationality that usually fall victim to the linguistic paradigm.

1.4.2. Summary

The passage through some paradigmatic contemporary theories of rationality has shown that rationality is defined as

1. the culturally conveyed competence of actors to orient themselves with the help of transindividual but context-varying standards in situations of action (Kambartel);

2. a set of normative demands, which, if followed, should promote the production of forms of life that find general agreement, are generally perceived to be acceptable and are thought to present a suitable basis for justified developments (Mittelstraß);

3. the competence of communicative actors to subject their communicative actions to universal validity claims and thus to fulfill the conditions required to criticize and, hence, understand them (Habermas);

4. the competence of actors to optimize their actions and beliefs relative to their well-understood individual interests (Gosepath);

5. a set of normative demands that are directed to interpreters who want to describe and understand the behavior of beings as actions and, to that end, ascribe a mental organization to those beings that substantially fulfills the individuating conditions for propositional states (Davidson, Dennett);

6. a competence of systems to reduce the complexity of their environment under conditions of self-maintenance and to assume a perspective that relates the selectivity of the system to the improbability of its existence—whatever that means (Luhmann).

Habermas’s concept allows us to solve the problem that we first of all have to view rationality as something that we ascribe to ourselves and that, to this degree, as Kambartel claims, is inherent in our practices; at the same time, as Mittelstraß emphasizes, it must be interpreted as something with a normative character and that can counter our practices. For Habermas’s anchoring of rationality theory in a pragmatic theory of meaning allows rationality to be described as a capability to generate and understand criticizable expressions. Inherent to our practices oriented toward understanding is thus simultaneously a capacity to fulfill rational competencies; so, too, there is an inherent possibility that each contribution to these practices will be an object of critique that measures the validity claim raised by the contribution in reference to its factual fulfillment. Habermas’s theory is problematic in the first instance because his theory of meaning is too ambitious insofar as its concept of rule following brings competencies into play that in my view can only be shown to be plausible if one presumes that the beings whose communicative competencies we want to elucidate with the concepts of rule following already possess a language that makes it possible for them to formulate, understand, and break rules.

Davidson’s considerations of rationality theory, like those of Habermas, are rooted in the theory of meaning, but Davidson’s theory of meaning can be understood as a theory that attempts to get by with minimal assumptions. Here, however, a strange shift in the normative perspective occurs. For in Davidson’s analytical perspective, rationality is first of all that which an interpreter must impute to a being so that it is understandable. However, what the interpreter here imputes to this being cannot establish a demand on the being that is to be understood, for it is only a possible addressee of normative criteria if it already fulfills these criteria in accord with the interpreter’s interpretation. In this sense, rationality, as a necessary assumption that makes understanding possible, has no normative function in the stricter sense, but a constitutive one. However, at the same time, Davidson also presents criteria for rationality that indeed have a normative character, but without it being clear how these criteria are related to the criteria that are constitutive for understanding and are directed at the interpreter. Although this ambivalence can be understood to be a result of the interpretationist composition of Davidson’s theory, in which self-interpretation is the anticipated interpretation of the other, in the face of this problem, one should manage to differentiate various levels of rationality that throw light on the question of the relationship between self-assumption and the normative demand, which Schnädelbach discussed under the catchword of necessary pre-understanding.

I thus would like to suggest differentiating the following levels of explication in the theory of rationality:

1. Basal rationality is that rationality that must be presupposed in order to be able to describe a being’s activity or behavior as an action. Basal rationality is thus introduced via an explicitly circular procedure: a being is minimally rational if it behaves in such a way that its behavior can be understood to fulfill the norms of minimal rationality—its behavior thus can be described as action. We decide whether minimal rationality exists with the aid of an all-or-nothing criterion; minimal rationality thus does not come in degrees. Four short remarks about that:

1. If the criterion for minimal rationality is not fulfilled, the behavior that is to be explained loses the status of action.

2. It is not easy to precisely determine the content of the criterion of basal rationality insofar as basal rationality is in a fundamental sense relative to the interpreter. However, perhaps it is possible to say that the attempt to understand an activity as an action has limits that can be specified by the fact that actions are activities that are based on reasons and that the existence of reasons has transindividual preconditions that come into play via the holism of the mental and of language. It is possible to specify the content of the requirements for basal rationality if one asks which principles must be presumed to be fulfilled so that it is possible to ascribe a being with intentions and beliefs that can take on the role of reasons for the actor. The holism of the mental now claims that it is not possible to ascribe to a being solely one intentional state insofar as the relational identity of the ascribed state is dependent on its inferential position in a network of further intentional states. Admittedly the restrictions for the inferential relations that are supposed to integrate such additional, necessarily ascribed intentional states into the system do not appear to be very comprehensive; rather, they appear essentially to be exhausted in that a being acts so that the network of intentional states that is ascribed to it is reconcilable with modus ponens.

3. The criterion for basal rationality is normative only in the special sense that fulfilling it legitimizes a way of speaking about a being. It cannot be required that a being fulfill the criterion. In short: the criterion has a different direction of fit than normal normative criteria.

4. It is possible to keep the ascription of basal rationality local in the sense that, for the purposes of explaining action, no assumptions about the entire network of the intentional states of a being must be made, and partial coherence may be sufficient for explanations of action.

2. Habitual rationality, in contrast to basal rationality, can be ascribed to those beings that participate in a social practice of interpretation and thus necessarily fulfill the criterion of basal rationality. If we, as outsiders to this practice of interpretation, gain the impression that this practice of interpretation is successful with a view to the ascribed problems of action coordination, then we are justified in the assumption that the participants in this contribute something: such beings follow factual norms that see to it that they are more often interpreted correctly (in relation to their interests). Norms that are generally followed by habitually rational beings—habitual rationality comes in degrees—cannot be specified in a canonical list with a claim to completeness. However, as a rule, habitually rational beings follow the principles named under (RD′).²²¹

3. Optional rationality is a form of second-order rationality. It assumes that the criterion of basal rationality and most principles of habitual rationality are usually or always fulfilled. Optional rationality makes the order of the network of intentional states itself the content of intentional states. It contrasts with habitual rationality, which indeed has consequences for the ordering of the network of intentional states insofar as it, in the interest of facilitating interpretation by others, adjusts the order of the network to mechanisms for successful understanding. The object of optional rationality, in contrast, is the optimization of the network with a view to coherence and comprehensiveness. The maximization of optional rationality is the basis for learning processes in the stricter sense and the development of a scientific attitude.

The criteria for optional rationality are normative insofar as achieving optional rationality requires that strategies are followed and beings are able to make the fulfillment of such criteria the object of their own intentional states, or not. Such norms require that the addressees

1. minimize episodic irrationality;

2. actively increase the overall coherence of the system of their intentional states;

3. increase the number of coherently correlated intentional states; and

4. expand their competencies for understanding.

The last-mentioned criteria for optional rationality of course articulate precisely those demands on learning processes that are Enlightenment processes in the sense that I have characterized them.²²² The only questionable matter is why these competencies for understanding should remain limited to competencies of linguistic (self-)interpretation.

1.5. WHY A THEORY OF RATIONALITY AS A THEORY OF MEDIA?

The theory of rationality converges to a theory of understanding. This is especially clear for thinkers like Habermas and Davidson, who operate within the framework of the linguistic paradigm and place the analysis of linguistic processes of understanding at the center of their work. However, in light of the fact that we possess highly developed forms of nonlinguistic communication, with art being in my view the exemplary case, this limitation is a restriction of the analysis of processes of understanding, which requires its own justification. At the same time, it is the reason that it is still necessary to achieve Habermas’s central goal of establishing a comprehensive concept of rationality that is not one-sided.

A first step in this direction could consist in establishing a concept of rationality that goes deeper than the level of linguistically communicative action. Habermas, who is skeptical of such “sub-surface migrations” of communication theory, has armed his theory against such attempts, which he views as threatening to relapse into mentalism. I attempt to contrast this with a theory of nonlinguistic understanding that nonetheless does not abandon the framework of the interactionist paradigm and to that degree does not think it is necessary to fear mentalism.

Habermas orients his work explicitly on the “conviction that language forms the medium for the historical and cultural embodiments of the human mind, and that a methodologically reliable analysis of mental activity must therefore begin with the linguistic expressions of intentional phenomena, instead of immediately with the latter.”²²³ The project I am following here, however, only shares one impulse of this belief: namely, that the study of the activity of mind is not directly accessible to a methodologically robust analysis. In contrast, restricting the analysis solely to the linguistic medium of the embodiment of mind seems questionable to me. I thus think it is sensible to search for an analogue to language that allows, first, the analysis of nonlinguistic forms of understanding within the framework of the interactionist paradigm of philosophy with which, second, it is possible to explain to what degree nonlinguistic forms of communication are processes of understanding and with which, third, it is possible to check and see whether these processes are based in something that is to be penetrated by a broadened concept of rationality. In agreement with some influential theories, I view media as such an analogue. In other words: without falling back into the immediacy of the philosophy of consciousness, with a study of media theory, I attempt to gain a view of the entire range of media through which the activity of mind can be expressed.

Given the lack of a comprehensive body of research such as that which has been developed since the beginning of the linguistic turn, a project of this type confronts disproportionately greater methodological difficulties than linguistically oriented analysis. An analysis aimed at all media must first secure its methodological basis, but in doing this it is not able to draw on presuppositions comparable to those that have been developed in over one hundred years of the philosophy of language. As I will show in the next chapter, it must indeed first develop a concept of media that can withstand the pressure of the demands of a theory with normative ambitions.