Chapter 6
It is doubtful that any single interpretation can bring together the many themes running through the unusually rich German idealist debate. Different interpretations stress different aspects of the debate in this period. This book has argued that cognitive constructivism is central to Kant’s critical philosophy as well to post-Kantian German idealism, hence to German idealism in general.
German idealism has now receded into history. It did not invent, but only reformulates, develops, and illustrates cognitive constructivism, which began before and continues after it. The legacy of German idealism lives on through cognitive constructivism. Constructivist epistemology attracts little attention in the current debate. The Swiss psychologist Jean Piaget—perhaps the most distinguished twentieth-century cognitive constructivist—formulates a theory of genetic epistemology to explain the mental development of children.1 In place of scientific realism, Van Fraassen’s constructivist empiricism rejects scientific realism by suggesting that our scientific theories are at best only empirically adequate.2
Cognitive constructivism is more frequently criticized than studied, but also is frequently rejected, often on superficial grounds.3 To begin to assess this approach, we must shift the focus from German idealism to the underlying cognitive problem. The interest of this cognitive strategy lies in a possible solution to the Parmenidean form of the cognitive problem, which turns on the normative criterion of metaphysical realism. “Realism”—which refers to the ontological independence of the cognitive object with respect to beliefs, knowledge, conceptual frameworks, and so on—is the dominant element in the cognitive debate. The type of realism one accepts leads in turn to specific epistemic strategies.
The Western cognitive tradition is determined through an early commitment to what later became metaphysical realism. “Metaphysical realism” is any form of the cognitive claim that knowledge requires a cognitive grasp of reality, or the mind-independent external world as it is, not merely as it appears. Though not under that name, metaphysical realism runs throughout the entire Western tradition at least since Parmenides. His ontological claim that what is, is and cannot not be—hence cannot change—influentially points to the grasp of mind-independent reality as the standard of knowledge. This standard runs throughout the entire later tradition.
Parmenides influences Plato—who adopts the view that to know is to know reality—and through him the later debate. Parmenides’s claim for the identity of thought and being presupposes at least three claims: first, there is mind-independent reality; second, thought (or thinking) can grasp mind-independent reality as it is beyond appearance; and third, when we know, thought in fact grasps mind-independent reality as it is. The second claim can be restated as an identity of (cognitive) identity and (ontological) difference. Since the normative conception of knowledge determines the strategy, acceptance of the Parmenidan criterion of mind-independent reality naturally led to an effort over many centuries to formulate epistemic strategies appropriate for this goal.
There is more continuity than discontinuity between the ancient way of framing the cognitive problem through Parmenidean realism and modern efforts to resolve it. In the ancient tradition, Plato, who identifies two main cognitive strategies, rejects representationalism when featuring intellectual intuition.
The modern debate features numerous efforts to reverse the Platonic rejection of representationalism through a causal theory of perception by linking together an idea in the mind to the mind-independent world. In a causal theory of perception, the idea in the mind is the effect of which the mind-independent world is the cause. Rationalism and empiricism—the two main forms of the new way of ideas—both feature a cognitive relation between ideas and reality. In both cases, cognition depends on the relation of ideas in the mind to mind-independent reality—more precisely, on the backward (or anti-Platonic) inference from the effect to its cause.
Cognitive representation is problematic since there seems to be no way to know that representations match up with, correspond to, or otherwise correctly represent the cognitive object. The main criticism is simple but devastating. If access to the cognitive object is possible only through representation, then there is no way to know how a representation relates to what it represents, hence no way to argue for the success of a representational approach to knowledge. This criticism, which seems to vindicate Plato’s rejection of the backward inference from effect to cause, undermines any and all modern forms of representationalism.
Kant, who uses representational language, is apparently unable to decide between representationalism and constructivism. He rejects intellectual intuition (which Plato favors) while formulating a constructivist approach to cognition, which is incompatible with any form of cognitive representation. Yet he never directly considers epistemic foundationalism, which is the main modern cognitive innovation.
In the modern debate, skepticism, foundationalism, and metaphysical realism are closely linked. Foundationalism—which exploits the analogy between a building, which is constructed on supposedly unshakeable foundations, and the correct epistemic strategy—is a peculiarly rigorous form of the causal theory of perception invoked to avoid skepticism in arguing for indefeasible cognitive claims. “Epistemic foundationalism” refers to any strategy for knowledge based on an initial principle or set of principles known to be absolutely certain. Modern foundationalism aims to overcome skepticism through metaphysical realism in meeting the Parmenidean standard.
When foundationalism comes into the tradition depends on what one understands it to be. Plato’s suggestion that mathematics and natural science depend on first principles, whose truth can be grasped through dialectic, is perhaps an early form of foundationalism. According to Aristotle, to avoid either an infinite regress or circular reasoning, the premises on which demonstration is based must either be demonstrable or not require demonstration, since in their role as first principles they cannot be demonstrated but are self-evidently true.
Aristotle argues that there must be such principles in order to have knowledge, which Descartes undertakes to demonstrate through his conception of the cogito. In comparison, Descartes can be said to improve on the Aristotelian approach in identifying a cognitive first principle, or foundation for cognition.
Geometry depends on axioms or postulates whose truth is assumed for purposes of demonstration in order to deduce theorems. Descartes takes the geometrical model further in making apodictic claims for knowledge. According to Descartes, there is a single foundationalist principle, or unshakeable Archimedean point, known to be true, and from which the remainder of the theory can be rigorously deduced. If a theory is rigorously deduced from a principle known to be true, then the theory is also true.
The deceptively simple Cartesian foundationalist cognitive approach is theoretically interesting but difficult to defend. This model depends on a causal theory of perception to argue from an idea in the mind to the mind-independent external world. The key argument suggests that if the cogito exists, then at least some of its ideas are also true. In this respect, Descartes makes a famous two-step argument. First, the cogito necessarily exists, since its existence cannot be denied. Second, clear and distinct ideas are true since God would not deceive me, and we can further reliably determine which ideas in the mind are true.
This argument has generated an enormous debate. Beginning with Arnauld, commentators often suggest that the Cartesian argument is undermined by the so-called Cartesian circle. This criticism consists in pointing out that Descartes appears to rely on the existence of God, which he has not yet demonstrated, to infer clear and distinct ideas are true. Descartes’s most important response is that clear and distinct ideas do not depend upon God to validate them. Suffice it to say that the argument about the Cartesian circle has never been decided. Other criticisms include the revival of circular demonstration and attention to the distinction between certainty and truth.
In the twentieth century, a new form of foundationalism emerged in the Vienna Circle. According to Carnap’s protocol theory, sentences about physical objects are not translated into sense data but into protocol sentences in order to weave a seamless web between immediate experience and natural science. Carnap’s initiative quickly led to a complex debate with Neurath, Quine, and later others, including Rorty. Neurath’s objection that there were in fact no protocols impelled Carnap to reformulate his position in ideal language. This was opposed by Quine’s denial of the basic Kantian distinction between analytic and synthetic propositions in shifting toward holism. Following Quine, Davidson and Sellars each separately criticize foundationalism. Davidson refutes empiricism on the grounds that a belief can only be grounded through another belief; Sellars influentially attacks the so-called myth of the given. Rorty, who seems to take analytic foundationalism as the standard of knowledge, argues against the pervasive idea of knowledge understood from a Baconian perspective as a so-called mirror of nature.
Scientific realism functioned for a time as a successor to the Parmenidean approach. It was featured under the influence of the Vienna Circle positivists in the first half of the twentieth century. Scientific realism, which presupposes a basic distinction between the so-called folk view and the view of modern natural science, suggests that only natural science describes the world as it is. This approach quickly assumed a dominant role after Carnap, Reichenbach, Hempel, and others came to the US. It just as quickly lost interest through the increasingly widespread disaffection with such related positivist doctrines as reductionism, physicalism, verificationism, and so on. Few if any thinkers currently accept Vienna Circle positivism in its original form. Yet the strong realism on which it insisted is still widely accepted as the standard of empirical knowledge, despite the evident inability to formulate a convincing argument for knowledge of mind-independent reality.
The shift away from foundationalism in the mid-twentieth century did not diminish the interest in metaphysical realism. Kant, who relies on the results of experience, justifies the Copernican turn through the fact that no progress has ever been made toward metaphysical realism. Yet this fact seems not to discourage those interested in metaphysical realism, who strive toward knowledge of the mind-independent real as it is.
The Parmenidean view that to know means to grasp the mind-independent real remains popular in recent debate. Thus Frege thinks that absolute ontological permanence is a necessary condition of cognition.4 Dummett holds that anti-realists, who cannot guarantee the truth-value of every sentence, must deny bivalence.5 According to Boghossian, there is in fact a way the world is and we can know it.6 None of these claims surpasses Platonic realism about universals. Now as in Kant’s time, the difficulty remains the same. If we admit we cannot intuit, represent, or found claims to know reality as it is, we must either make cognitive claims while abandoning any pretense of knowing reality or accept skepticism. This problem is not alleviated if we suppose there is a way that things are in independence of us and that knowledge requires us to know them as they are. The traditional approach, which lies in adopting metaphysical realism as our standard, fails in practice. For no one in the tradition of Western philosophy has ever formulated a convincing argument to show that we can grasp the mind-independent real. Now as in Kant’s time, the most promising alternative is a constructivist approach that, in turning away from metaphysical realism, takes empirical realism as its cognitive standard.
Kant rejects intellectual intuition and representation, hence metaphysical realism, before embarking on the Copernican turn. We might now—some two centuries after Kant—desire to appeal to more recent cognitive approaches. There is no way to anticipate philosophical ingenuity. Yet we can dismiss any version of the claim to cognize reality. Plato, who seems to equate cognition with knowledge of reality, apparently suggests that philosophers can see, hence cognize, reality, or that this must the case if there is to be knowledge. Yet there does not seem to be any way to cognize reality. Related claims—such as the regularity of experience, or the ability to predict experience—do not show that we in fact grasp reality.
Cognitive constructivism, which follows the Kantian suggestion that objects must conform to the mind, avoids skepticism deriving from the failure to cognize reality. This suggestion offers an approach to knowledge without imposing the impossible Parmenidean standard. Though constructivism is the most promising among current approaches to cognition, it appears unacceptable for two reasons: it directly contradicts the venerable effort stretching back to the origins of the Western tradition to cognize the world as it is, and it is often described in imprecise ways—or at least in ways that do not directly relate to the problem of cognition. For instance, Rawls, who appeals to “constructivism” in his moral theory, never defines the term, which in turn undermines his own position as well as this general approach.7
Kant identifies an important argument for constructivism by noting that the effort to know mind-independent reality has not progressed. If this is correct, then the current interest of cognitive constructivism seems to be threefold: first, since we cannot show we know reality as it is, this goal remains regulative but cannot be constitutive; second, constructivism is useful in enabling us to avoid epistemic skepticism; and, third, constructivism correctly describes cognitive praxis.
The first argument has already been made. Since there is more than one possible cognitive standard, the second argument obviously depends on how one characterizes it—theories of cognition rely in all cases on an account of what counts as knowledge. This argument depends on accepting a revision of the widespread normative philosophical conception of knowledge. The Parmenidean view that cognition requires the identity of thought and being has often been interpreted as mandating indefeasible claims to know. We can distinguish between the many forms this argument has already taken and the underlying normative commitment to an ahistorical conception of knowledge.
The Western philosophical debate centers on an ongoing search, spanning the entire tradition, for cognitive rigor. In ancient philosophy, this commitment motivates the Socratic search for ethical universals, through the distinction between knowledge and opinion, the Platonic view of philosophy as the science of sciences, capable of justifying itself through dialectic, anhypothetical, Aristotelian essentialism, and so on. The modern debate features such cognitive approaches as representationalism as well as cognitive foundationalism. In our day, the commitment to rigorous cognition, illustrated by the Husserlian conception of philosophy as rigorous science, takes many forms. They include: scientific empiricism; logicism, or the reduction of mathematics to logic; semantics, or the problem of reference further linked to the analysis of language; scientism, or the view that natural science provides the most authoritative—or in a more extreme version, perhaps the only authoritative worldview; and so on.
Efforts to formulate a rigorous epistemic theory often approach cognition as ahistorical, beyond time and place, unrelated to opinion, unrelated to the historical moment, unrelated to subjectivity of any kind, and so on. The identification of cognitive rigor with a supposed ahistorical status is obviously intended to drive a wedge between cognition and history. Thus Kant notoriously claims the critical philosophy—which none of his contemporaries accepted without modification, and which Kant hastened to revise—would never require revision.
The widespread conviction that cognitive rigor is possible if and only if the cognitive object is unchanging leads to ahistorical cognitive claims. Thus Husserl likens historicism of any kind to cognitive relativism, hence to skepticism. Yet mind-independent reality—which ahistorical thinkers among us often presuppose—is a historical construct embedded in a theory of knowledge. The Higgs boson, for instance, is meaningful only within the conceptual framework of the standard view of matter. There is clearly nothing ahistorical about this subatomic particle, which literally depends on the theory for its existence.
The normative conception of ahistorical knowledge, which comes to us from ancient Greece, is one of the most hallowed philosophical traditions. The ancient dream of a final philosophical account of cognition partly depends on philosophical vocabulary. The familiar dualism between appearance and reality falsely suggests there is a way that reality is, and even that it can be known as it is. Yet “reality” is not independent of, but rather depends on the theory about it. There is simply no reason—other than the long-standing conviction that we know or ought to strive to know reality beyond change—to support the commitment to a normative view of cognition as ahistorical, hence as “unrevisable.” Though there is clearly a longing for a halcyon ultimate resting place beyond the tradition where things will never change (in short, a final scientific theory), there is also no reason—none at all, since we cannot compare our theories of reality to reality—to think that we are almost there, or are even getting closer to grasping it. We do not now know, and there seems to be no reason to think we will ever know, that we are, or perhaps later might be able to, carve nature at the joints. If history is our guide, there is no reason to think we will ever realize the dream of a final theory, which, hence, is regulative but cannot be constitutive.
The third point is the claim that constructivism correctly describes the cognitive process. Constructivism applies a version of the familiar scientific hypothetico-deductive model to cognition in general. The central assumption is that cognition depends on the experimental construction of a theory about the contents of consciousness. We examine our theory through further experience, and we modify it as needed. Scientific inquiry proceeds by formulating a hypothesis, which is then tested against experience; if it is falsified, is regarded as false. Constructivist cognitive claims, like other scientific claims, are not subject to confirmation but are fallible, hence subject to refutation.
A constructivist approach avoids a number of common but indemonstrable cognitive claims. Thus it makes no assumption that through the cognitive process we are getting at the real—or indeed, anything beyond the contents of consciousness. It further admits revisability in turning away from the familiar claim for apodicticity. And it neither claims to cognize nor seeks to cognize what does not change, hence lies beyond time and history.
Philosophy is, like other forms of cognition, a historical enterprise. It builds on prior efforts to resolve, solve, or at least come to grips with ongoing concerns. The implicit suggestion that philosophy only finally begins and in fact ends in the critical philosophy is clearly mistaken. Kant’s considerable achievement does not lie in bringing the debate to a close. It rather lies in reformulating a powerful constructivist approach to cognition. Constructivism is central to Kant, central to post-Kantian German idealism—hence to all forms of German idealism—and central to the cognitive problem after German idealism. Now as in Kant’s day, constructivism answers Plato’s reaction to Parmenides. Metaphysical realism—or in some views, realism tout court—reiterates the ancient but impossible Platonic conviction that cognition is cognition of reality. This theoretical claim in practice gives way to the view, central to German idealism and to selected later thinkers, that we know only what we in some sense construct.