Hegel opens the first paragraph of his Introduction to the Phenomenology of Spirit by introducing a model of cognitive faculties that he supposes will be most familiar to his readers in its Kantian form:

Knowledge … tends to be regarded as the instrument with which one takes hold of the Absolute or as the medium through which one discovers it. [PG 73]

He thinks no account that has this general shape can meet basic epistemological criteria of adequacy. By showing that, he hopes to make his readers appreciate the need for an alternative model, which he will then supply.

The general character of his complaint against construing cognitive faculties on the instrument-or-medium model seems clear enough. He offers a twofold summary. That model leads to “the conviction that there is”

“[A]n absurdity in the Concept of even beginning a process of knowledge designed to gain for consciousness that which is in-itself.”
“[A] strict line of demarcation separating knowledge and the Absolute.” [PG 73]

The first objection alleges that theories of the sort he is addressing must lead to a kind of skepticism: a failure to make intelligible the idea of knowing how things are in themselves. The second complaint points to a diagnosis of the reason for this failure: the model excavates a gulf separating consciousness from what it is consciousness of.

He expands on both these points. He fills in the charge that instrument-or-medium theories lead to skepticism by saying:

[I]f knowledge is the instrument to take hold of the absolute essence, one is immediately reminded that the application of an instrument to a thing does not leave the thing as it is, but brings about a shaping and alteration of it. Or, if knowledge is not an instrument for our activity, but a more or less passive medium through which the light of truth reaches us, then again we do not receive this truth as it is in itself, but as it is in and through this medium. In both cases we employ a means which immediately brings about the opposite of its own end; or, rather, the absurdity lies in our making use of any means at all. [PG 73]

In either case, there is going to be a distinction between what things are for consciousness (the product of the exercise of cognitive faculties) and what they are in themselves (the raw materials on which the cognitive faculties are exercised). Something about the character of this distinction, Hegel seems to be arguing, is incompatible with what things are for consciousness according to such a picture counting as genuine knowledge of how things really are (“in themselves”).

He elaborates the problem diagnosed in the second conviction, given in the preceding. It is that the instrument-or-medium picture

presupposes notions about knowledge as an instrument and a medium, and also the notion that there is a difference between ourselves and this knowledge; but above all, it presupposes that the Absolute stands on one side and that knowledge, though it is on the other side, for itself and separated from the Absolute, is nevertheless something real. Hence it assumes that knowledge may be true despite its presupposition that knowledge is outside the Absolute and therewith outside the truth as well. By taking this position, what calls itself the fear of error reveals itself as a fear of the truth. [PG 74]

It is apparently of the essence of the instrument-or-medium model to see such a “difference,” “separation,” two “sides” of one divide, and to understand the job of cognitive faculties to consist in bridging that divide.

This, he thinks, is precisely the predicament that calls forth an inquiry into the nature of the transformation effected by the exercise of cognitive faculties. But he claims that it is a mistake to think such an investigation can remove the difficulty.

To be sure, it does seem that an acquaintance with the way the instrument functions might help overcome this difficulty. For then it would seem possible to get the truth in its purity simply by subtracting from the result the instrument’s part in that representation of the Absolute which we have gained through it. In fact, however, this correction would only lead us back to our point of departure. For [i], if we remove from a thing which has been shaped by an instrument the contribution of that instrument to it, then the thing (in this case the Absolute) is for us exactly as it was before this now obviously superfluous effort. Or [ii], were the Absolute only to be brought a bit closer to us by an instrument, perhaps as a bird is trapped by a lime-twig, without being changed at all, it would surely laugh at this ruse if it were not, in and for itself, already close to us of its own accord. For in this case knowledge itself would be a ruse, pretending through its multifarious effort to do something other than merely bring forth a relation which is immediate and thus effortless. Or [iii], if the examination of knowledge, which we now represent as a medium, makes us acquainted with the law of light-refraction in the medium, it is likewise useless to subtract this factor from the result; for knowledge, through which the truth touches us, is the ray of light itself rather than its refraction; and if this be subtracted, we would be left with no more than an indication of pure direction or empty place. [PG 73]

The argument here seems to be that if there is a gulf separating how things are in themselves from how they are for consciousness that requires the operation of cognitive faculties to bridge it or reunite the two sides, then all that investigation of those faculties can do is reinstitute the gulf, or separation.

I think we can see in these passages the general shape of an argument. But it is hazy, and it is hard to discern both the exact outlines of the class of views it targets and just how the criticism of them is supposed to work. (The haziness of the argument is due partly to the compression of its exposition, and partly to the metaphorical terms in which it is conducted.) To fill in the details, one would have to specify what criteria of adequacy for epistemological theories Hegel is insisting on, what class of theories he claims cannot satisfy those criteria, what features of those theories are responsible for that failure, and how, exactly, the argument for that conclusion works. In the rest of this chapter, I offer one way of sharpening along these four dimensions the argument Hegel is putting on the table here, and an initial characterization of the shape of the alternative model that Hegel proposes to replace the instrument-or-medium model.

To get a better specification of the range of epistemological theories that fall within the target area of Hegel’s argument (metaphorically labeled as the “instrument-or-medium” model), it will help to begin further back. (This is the method we will come to see as at the heart of both Hegel’s practice and his theory: recollection, whose slogan is Reculer pour mieux sauter.) The theories he is addressing are representational theories of the relations between appearance and reality. Representation is a distinctively modern concept. Premodern (originally Greek) theories understood the relations between appearance and reality in terms of resemblance. Resemblance, paradigmatically one of the relations between a picture and what it pictures, is a matter of sharing properties. A portrait resembles the one portrayed insofar as it shares with its object properties of color and shape—for instance, of nose, ear, and chin (perhaps as seen from some perspective). The thought behind the resemblance model is that appearance is veridical insofar as it resembles the reality it is an appearance of. Insofar as it does not resemble that reality, it is a false appearance, an error.

The rise of modern science made this picture unsustainable. Copernicus discovered that the reality behind the appearance of a stationary Earth and a revolving Sun was a stationary Sun and a rotating Earth. No resemblance, no shared properties there. The relationship between reality and its appearance here has to be understood in a much more complicated way. Galileo produces a massively effective way of conceiving physical reality in which periods of time appear as the lengths of lines and accelerations as the areas of triangles. The model of resemblance is of no help in understanding this crucial form of appearance. The notion of shared property that would apply would have to be understood in terms of the relations between this sort of mathematized (geometrized) theoretical appearance and the reality it is an appearance of. There is no antecedently available concept of property in terms of which that relationship could be understood.¹

Descartes came up with the more abstract semantic metaconcept of representation required to make sense of these scientific achievements—and of his own. The particular case he generalized from to get a new model of the relations between appearance and reality (mind and world) is the relationship he discovered between algebra and geometry. For he discovered how to deploy algebra as a massively productive and effective appearance of what (following Galileo) he still took to be an essentially geometric reality. Treating something in linear, discursive form, such as “ax + by = c” as an appearance of a Euclidean line, and “x² + y² = d” as an appearance of a circle, allows one to calculate how many points of intersection they can have and what points of intersection they do have, and lots more besides. These sequences of symbols do not at all resemble lines and circles. Yet his mathematical results (including solving a substantial number of geometric problems that had gone unsolved since antiquity, by translating them into algebraic questions) showed that algebraic symbols present geometric facts in a form that is not only (potentially and reliably) veridical, but conceptually tractable.

In order to understand how strings of algebraic symbols (as well as the Copernican and Galilean antecedents of his discoveries) could be useful, veridical, tractable appearances of geometric realities, Descartes needed a new way of conceiving the relations between appearance and reality. His philosophical response to the scientific and mathematical advances of this intellectually turbulent and exciting time was the development of a concept of representation that was much more abstract, powerful, and flexible than the resemblance model it supplanted. He saw that what made algebraic understanding of geometric figures possible was a global isomorphism between the whole system of algebraic symbols and the whole system of geometric figures. That isomorphism defined a notion of form shared by the licit manipulations of strings of algebraic symbols and the constructions possible with geometric figures. In the context of such an isomorphism, the particular material properties of what now become intelligible as representings and representeds become irrelevant to the semantic relation between them. All that matters is the correlation between the rules governing the manipulation of the representings and the actual possibilities that characterize the representeds. Inspired by the newly emerging forms of modern scientific understanding, Descartes concluded that this representational relation (of which resemblance then appears merely as a primitive species) is the key to understanding the relations between mind and world, appearance and reality, quite generally.

This was a fabulous, tradition-transforming idea, and everything Western philosophers have thought since (no less on the practical than on the theoretical side) is downstream from it, conceptually, and not just temporally—whether we or they realize it or not. But Descartes combined this idea with another, more problematic one. This is the idea that if any things are to be known or understood representationally (whether correctly or not), by being represented, then there must be some things that are known or understood nonrepresentationally, immediately, not by means of the mediation of representings. If representings could be known only representationally, by being themselves in turn represented, then a vicious infinite regress would result. For we would be able to know about a represented thing only by knowing about a representing of it, and could count as knowing about it only if we already knew about a representing of it, and so on. In a formulation that was not extracted explicitly until centuries later by Josiah Royce, if even error (misrepresentation), never mind knowledge, is to be possible, then there must be something about which error is not possible—something we know about not by representing it, so that error in the sense of misrepresentation is not possible. If we can know (or be wrong about) anything representationally, by means of the mediation of representings of it, there must be some representings that we grasp, understand, or know about immediately, simply by having them.

The result was a two-stage representational story that sharply distinguished between two kinds of things, based on their intrinsic intelligibility. Some things, paradigmatically physical, material, extended things, can by their nature be known only by being represented. Other things, the contents of our own minds, are by nature representings, and are known in another way entirely. They are known immediately, not by being represented, but just by being had. They are intrinsically intelligible, in that their mere matter-of-factual occurrence counts as knowing or understanding something. Things that are by nature knowable only as represented are not in this sense intrinsically intelligible. Their occurrence does not entail that anyone knows or understands anything.

As I have indicated, I think that Descartes was driven to this picture by two demands. On the one hand, making sense of the new theoretical mathematized scientific forms in which reality could appear—the best and most efficacious forms of understanding of his time—required a new, more abstract notion of representation and the idea that it is by an appropriate way of representing things that we know and understand them best. So we must distinguish between representings and representeds, and worry about the relations between them in virtue of which manipulating the one sort of thing counts as knowing or understanding the other. On the other hand, such a two-stage model is threatened with unintelligibility in the form of a looming infinite regress of explanation if we don’t distinguish between how we know representeds (by means of our relations to representings of them) and how we know at least some representings (immediately, at least, not by being related to representings of them). The result was a two-stage model in which we are immediately related to representings, and in virtue of their relation to representeds stand in a mediated cognitive relation to those represented things. The representings must be understood as intrinsically and immediately intelligible, and the representeds as intelligible only in a derivative, compositional sense: as the result of the product of our immediate relations to representings and their relations to representeds.

I want to say that it is this epistemological model that Hegel takes as his target in his opening remarks in the Introduction of the Phenomenology. What he is objecting to is two-stage representational theories that are committed to a fundamental difference in intelligibility between appearances (representings, how things are for consciousness) and reality (representeds, how things are in themselves), according to which the former are immediately and intrinsically intelligible, and the latter are not. The gulf, the “difference,” “separation,” the two “sides” of one divide separating appearance and reality, knowing and the known, that he complains about is this gulf of intelligibility. Hegel’s critical claim is that any theory of this form is doomed to yield skeptical results.

Of course, Descartes’s view is not the only one Hegel means to be criticizing. By using the phrase “a process of knowledge designed to gain for consciousness that which is in-itself,” Hegel seems to indicate that Kant, too, falls within his target range. He is thinking of Kant as having a two-stage representational theory. Cognitive activity needs to be understood as the product of both the mind’s activities of manipulating representations (in the sense of representings) and the relations those representings stand in to what they represent. Both what the mind does with its representations and how they are related to what they represent must be considered in apportioning responsibility for features of those representings to the things represented, as specified in a vocabulary that does not invoke either the mind’s manipulation of representations, or the relations between representings and representeds (that is, things as they are “in themselves” [an sich]), or to the representational relations and what the cognitive faculties do with and to representings. The latter for Kant yields what the represented things are, in Hegel’s terminology, “for consciousness”: contentful representings.

Kant’s theory is not the same as Descartes’s. But it shares the two-stage representational structure that distinguishes the mind’s relation to its representings and its relation to representeds that is mediated by those representings. Although Kant does sometimes seem to think that we have a special kind of access to the products of our own cognitive activity, he does not think of our awareness of our representings as immediate in any recognizably Cartesian sense. Awareness is apperception. The minimal unit of apperception is judgment. To judge is to integrate a conceptually articulated content into a constellation of commitments exhibiting the distinctive synthetic unity of apperception. Doing that is extruding from the constellation commitments incompatible with the judgment being made and extracting from it inferential consequences that are then added to that constellation of commitments. This is a process that is mediated by the relations of material incompatibility and consequence that relate the concepts being applied in the judgment to the concepts applied in other possible judgments.

So Kant shares with Descartes the two-stage representational structure, but does not take over the idea that our relation to our own representations is one of immediate awareness.² His view would still fall within the range of Hegel’s criticisms, however, insofar as he maintains the differential intelligibility of representings and represented: that representings are as such intelligible, and what is represented is, as such, not.³ I call this commitment to a “strong differential intelligibility” of appearance and reality: the claim that the one is the right sort of thing to be intelligible, and the other is not. Kant has a new model of intelligibility: to be intelligible is to have a content articulated by concepts. It is the concepts applied in an act of awareness (apperception) that determine what would count as successfully integrating (“synthesizing”) that judgment into a whole exhibiting the sort of unity distinctive of apperception. But the conceptual articulation of judgments is a form contributed by the cognitive faculty of the understanding. It is not something we can know or assume to characterize what is represented by those conceptual representings, when the representeds are considered apart from their relation to such representings—that is, as they are in themselves. Insofar as Hegel takes Kant to fall within the scope of his criticisms, he is treating Kant as committed to a gulf of intelligibility separating our representings from what they are representings of, in the form of the view that the representings are in conceptual shape, and what is represented (how things are in themselves) is not.

Just to remind ourselves how much is at stake in Hegel’s criticism of two-stage representational theories of the relations between appearance and reality that are committed to the differential intelligibility of the relata, it is worth thinking in this connection also about Frege. For Frege, discursive symbols express a sense [Sinn] and thereby designate a referent [Bedeutung]. Senses are what is grasped when one understands the expression, and referents are what is thereby represented: what expressing that sense is talking or thinking about. A sense is a representing in that it is a “mode of presentation” [Art des Gegebenseins] of a referent. No more than Kant does Frege construe grasp of a sense as immediate in a Cartesian sense—according to which the mere occurrence of something with that sense counts as the mind’s knowing or understanding something. Grasping a judgeable content requires mastering the inferential and substitutional relations it stands in to other such contents. But like Descartes and Kant, Frege thinks that grasping senses, understanding representations as representations, does not require representing them in turn, and that representings (senses) are graspable in a sense in which what they represent is not (apart from the special case of indirect discourse, where what is represented is itself a sense). So if, as I have claimed, Hegel’s argument is intended to be directed at two-stage representational models committed to treating representings as intelligible in a sense in which representeds are in general not, then it seems Fregean sense-reference theories, as well as the Kantian and Cartesian versions, will be among the targets.

II. Genuine Knowledge and Rational Constraint

In order to see whether there is an argument of the sort Hegel is after that tells against theories of this kind—two-stage representational theories committed to the strong differential intelligibility of representings and representeds—we must next think about what criteria of adequacy for such theories Hegel is appealing to. In general, we know that what Hegel thinks is wrong with them is that they lead to skepticism. Further, he tells us that what he means by this is that such theories preclude knowing things as they are “in themselves.” I think what is going on here is that Hegel learned from Kant that the soft underbelly of epistemological theories is the semantics they implicitly incorporate and depend upon. And he thinks that two-stage representational theories committed to the strong differential intelligibility of representings and what they represent semantically preclude genuine knowledge of those representeds.

I call the criterion of adequacy on epistemological theories that Hegel is invoking here the “Genuine Knowledge Condition” (GKC). Obviously, a lot turns on what counts as genuine knowledge. But it is clear in any case that this requirement demands that an epistemological theory not be committed to a semantics—in particular, a theory of representation—that rules out as unintelligible the very possibility of knowing how things really are (“genuine” knowledge). This is what I take Hegel to mean when he says that epistemological theories of this kind show themselves as surreptitiously expressing a “fear of the truth.” I do not take it that the very existence of a contrast between how we know what is represented and how we know representings by itself demonstrates such a failure. His specific claim is that when that difference is construed as one of intelligibility in the strong sense—representings are intrinsically intelligible and representeds are not—then skepticism about genuine knowledge is a consequence. And he takes from Kant the idea that intelligibility is a matter of conceptual articulation: to be intelligible is to be in specifically conceptual shape. If this reading is correct, then Hegel’s argument must show that to satisfy the Genuine Knowledge Condition, an epistemological theory must treat not only appearance (how things subjectively are, for consciousness), but also reality (how things objectively are, in themselves) as conceptually articulated. Again, what could count as a good argument for this claim obviously turns on what is required to satisfy that requirement.

Both resemblance and representation models of the relations between appearance and reality have a story about what error consists in. That is what happens when antecedently intelligible properties are not shared, so that resemblance breaks down, or when there are local breakdowns in the globally defined isomorphism between the systems of representings and representeds. In the middle paragraphs of the Introduction, in which Hegel begins to present his alternative to two-stage representational epistemological theories committed to strong differential intelligibility of representings and representeds, the treatment of error looms large. (This is the topic of Chapter 2.) I think we can take it as an implicit criterion of adequacy Hegel is imposing on epistemological theories that they make intelligible the phenomenon, not only of genuine knowledge, but also of error. I call this the “Intelligibility of Error Condition” (IEC).

The Genuine Knowledge Condition and the Intelligibility of Error Condition are epistemological constraints. The semantics presupposed by or implicit in an epistemological theory must not preclude the intelligibility either of genuine knowledge or of error: being wrong about how things really are. We must be able to understand both what it is for what there is to appear as it is and for it to appear as it is not. An epistemological theory that does not make both of these intelligible is not adequate to the phenomenon of our efforts to know and understand how things really are.

Approaching epistemology from this semantic direction suggests that behind these epistemological constraints are deeper semantic ones. I think that is in fact the case here. We cannot read these off of Hegel’s extremely telegraphic remarks in the text of the opening paragraphs of the Introduction, but must infer them from the solution he ultimately proposes to the challenges he sets out there. First is what we could (looking over our shoulders at Frege) call the “Mode of Presentation Condition” (MPC). This is the requirement that appearances (senses, representings) must be essentially, and not just accidentally, appearances of some purported realities. One does not count as properly having grasped an appearing unless one grasps it as the appearance of something. When all goes well, grasping the appearance must count as a way of knowing about what it is an appearance of. Appearances must make some reality semantically visible (or otherwise accessible). The claim is not that one ought not to reify appearances, think of them as things, but rather, for instance, adverbially: in terms of being-appeared-to-thus-ly. That is not a silly thought, but it is not the present point. It is that if the epistemological Genuine Knowledge Condition is to be satisfied by a two-stage representational model, representings must be semantic presentations of representeds in a robust sense in which what one has grasped is not a representation unless it is grasped as a representing of some represented. Further along we’ll see how Hegel, following Kant, understands this requirement: taking or treating something in practice as a representing is taking or treating it as subject to normative assessment as to its correctness, in such a way that what thereby counts as represented serves as a standard for assessments of correctness.

A second semantic constraint on epistemological theories that I take to be implicitly in play in Hegel’s understanding of the epistemological GKC is that if the representational relation is to be understood semantically in a way that can support genuine knowledge, it must portray what is represented as exerting rational constraint on representings of it. That is, how it is with what is represented must, when the representation relation is not defective, provide a reason for the representing to be as it is. What we are talking (thinking) about must be able to provide reasons for what we say (think) about it. We can call this the “Rational Constraint Condition” (RCC). Though he does not argue for this constraint in the Introduction, I think in many ways it is the key premise for the argument he does offer. The thought is that the difference between merely responding differentially to the presence or absence of a fact or property and comprehending it, having thoughts that are representationally about it in the sense that if everything goes well counts as knowledge of it, depends on the possibility of that fact or property being able to serve for the knower as a reason for having a belief or making a commitment. The central sort of semantic aboutness depends on being able rationally to take in how things are, in the sense of taking them in as providing reasons for our attitudes, which in virtue of that rational accountability count as being about how those things are.

Hegel learns from Kant to think about representation in normative terms. What is represented exercises a distinctive kind of authority over representings. Representings are responsible to what they represent. What is represented serves as a kind of authoritative normative standard for assessments of the correctness of what count as representings of it (correct or incorrect) just in virtue of being subject to assessments of their correctness in which those representeds provide the standard. The RCC adds that the standard, what is represented, must provide reasons for the assessments. In fact, in the context of Kant’s and Hegel’s views, this is not a further commitment. For neither of them distinguishes between norms (or rules) in general and norms (or rules) that are rational in the sense of being conceptually articulated. All norms are understood as conceptual norms. Norms or rules and concepts are just two ways of talking about the same topic. Conceptual norms are norms that determine what is a reason for what. For a norm to be contentful is for it to have conceptual content: a matter of what it can be a reason for or against and what can be a reason for or against it. This is the only kind of content they acknowledge. The German Idealists are rationalists about norms, in that norms (rules) are contentful exclusively in the sense of being conceptually contentful.

The Rational Constraint Condition accordingly fills in the sense of “representation” or “aboutness” on which the Mode of Presentation Condition depends. And these two semantic conditions provide the crucial criteria of adequacy for satisfying the two epistemological conditions: the Genuine Knowledge Condition and the Intelligibility of Error Condition. For the intelligibility of genuine knowledge of or error about how things really are turns on the rational normative constraint those realities exert on what count as appearances or representings of those realities just insofar as they are subject to normative assessments of correctness and incorrectness (knowledge or error) in which those realities serve as the standard, in the sense of providing reasons for those assessments.

Supposing that these four conditions represent the relevant criteria of adequacy for epistemological theories (and their implicit semantics), what is the argument against two-stage representational theories that are committed to a strong difference of intelligibility between representings and representeds (appearance and reality)? Why can’t theories of this form satisfy the criteria of adequacy? It is characteristic of two-stage theories, not just Descartes’s but also those of Kant and Frege, that they incorporate a distinction between two ways of knowing or understanding things. Some things are known (only) representationally: by being represented. Other things—at least some representings, according to the regress argument—are known nonrepresentationally: in some way other than by being represented. If we are interested in investigating cognitive faculties in the context of theories like this, we are interested in the representation relation. For cognitive faculties are the instrument or medium that produces representings of the real. But then we must ask: Is the representational relation, the relation between representings and what they represent, itself something that is known representationally or nonrepresentationally? If it is itself something that is knowable or intelligible only by being represented, it seems that we are embarked on a vicious Bradleyan regress (a successor to the one Descartes was worried about). The epistemological enterprise is not intelligible unless we can make sense of the relation between representations of representational relations (what they are for representers) and those representational relations themselves, and then representations of those relations, and so on. Until we have grasped all of that infinite chain of representings of representings of representings …, we are not in a position to understand the representational relation, and hence not the “instrument or medium” of representation. Semantic skepticism—skepticism about what it is so much as to purport to represent something—must then be the result. This argument is essentially the Cartesian regress-of-representation argument for nonrepresentational knowledge of representings, applied now not just to the representings, but to the representational relations they stand in to what they represent.

So if epistemology, and so knowledge, is to be intelligible, it seems that within this sort of framework we must embrace the other horn of the dilemma, and take it that the representation relation is something that can itself be known or understood nonrepresentationally—that in this respect it belongs in a box with the representations or appearances themselves. Responding this way to the dilemma concerning our understanding of the representational relation is, in effect, acknowledging the Mode of Presentation Condition. For it is saying that part of our nonrepresentational understanding of appearances (representings) must be understanding them as appearances (representings) of something. Their representational properties, their “of”-ness, their relation to what they at least purport to represent, must be intelligible in the same sense in which the representings themselves are.

The Rational Constraint Condition says that for appearances to be intelligible as appearances, representings, modes of presentation, of something they must be intelligible as rationally constrained by what they then count as representing. This means that what is represented must be intelligible as providing reasons for assessments of correctness and incorrectness of appearances or representings. Reasons are things that can be thought or said: cited as reasons, for instance, for an assessment of a representing as correct or incorrect, as amounting to knowledge or error. That is to say that what provides reasons for such assessments must itself, no less than the assessments, be in conceptual form. Giving reasons for undertaking a commitment (for instance, to an assessment of correctness or incorrectness) is endorsing a sample piece of reasoning, an inference, in which the premises provide good reasons for the commitment. It is to exhibit premises, the endorsement of which entitles one to the conclusion. So the reasons, no less than what they are reasons for, must be conceptually articulated.

Put another way, appearances are to be intelligible, graspable, in the sense that they are conceptually articulated. Understanding the judgment that things are thus and so requires knowing what concepts are being applied, and understanding those concepts. One does that only insofar as one practically masters their role in reasoning: what their applicability provides reasons for and against, and the applicability of what other concepts would provide reasons for or against their applicability. If the relation between appearances and the realities they are appearances of—what they represent, how they represent things as being (“thus and so”)—is to be intelligible in the same sense that the appearances themselves are (so that a regress of representation is avoided), this must be because that relation itself is a conceptual relation: a relation among concepts or concept-applications, a relation between things that are conceptually articulated.

The conclusion is that if the Rational Constraint Condition must be satisfied in order to satisfy the Genuine Knowledge Condition and the Intelligibility of Error Condition (if the RCC really is a semantically necessary condition on satisfying these epistemological criteria of adequacy)—perhaps because it is a necessary condition of satisfying the Mode of Presentation Condition, which itself is a necessary semantic condition on satisfying the epistemological GKC and the IEC—then those conditions cannot be satisfied by a two-stage representational theory that is committed to the strong differential intelligibility of representing and represented. If not only representings, but the representation relation must be intelligible in a sense that requires their conceptual articulation, then both ends of the representation relation must be conceptually contentful. Only in that way is it intelligible how what is represented can exert rational constraint on representings, in the sense of providing reasons for assessments of their correctness or incorrectness.

III. A Nonpsychological Conception of the Conceptual

I have been working to find structure beneath what appears on the telegraphic, metaphor-laden surface of the text of the opening paragraphs of Hegel’s Introduction. I claim so far only to have sketched a potentially colorable argument. Further exploration is required in particular of the reasons for accepting the RCC, which this exposition reveals as the principal load-bearing premise. A key component of that enterprise would be clarifying the concepts of conceptual articulation and conceptual content—what the RCC says must characterize both representing and represented, which commitment to a representational theory with a strong difference of intelligibility denies. It will help to begin on this latter task by looking at what sorts of theories might be thought to be available, once the strong difference of intelligibility of appearance and reality has been denied—that is, once one is committed to not excavating a gulf of intelligibility between representings and what they represent.

One place to begin is with Frege’s proposed definition in “The Thought”: “a fact is a thought that is true.”⁴ Thoughts for Frege are the senses of declarative sentences. They are claims, in the sense of claimable contents, rather than claimings. A fact, he is saying, is not something that corresponds to or is represented by such a sense. It just is such a sense; one that is true. Facts are a subset of claimables, senses, representings, cognitive appearings. I take it, though, that Frege retains the two-stage representational model for the relation between senses and their referents—for thoughts and truth-values. And this matters for what he thinks senses are: modes of presentation of referents. But as far as the relations between thoughts and facts are concerned, he does not appeal to that model. Again, Wittgenstein says: “When we say, and mean, that such-and-such is the case, we—and our meaning—do not stop anywhere short of the fact; but we mean: this-is-so.”⁵ In these cases, the content of what we say, our meaning, is the fact. Such an approach is sometimes talked about under the title of an “identity theory of truth.”⁶ It is sometimes attributed, under that rubric, to John McDowell.⁷

On such an approach, there is no principled gulf of intelligibility between appearance and reality (mind and world), because when all goes well the appearances inherit their content from the realities they are appearances of. Thoughts (in the sense of thinkings) can share their content with the true thoughts (in the sense of thinkables) that are the facts they represent. (As indicated earlier, this is not the way Frege would put things. For him, facts are a kind of representing, not in the first instance of representeds.) Representings are distinct from representeds, so the two-stage representational model is still endorsed. But they are understood as two forms in which one content can be manifested.

What is most striking about views of this stripe is that they are committed to the claim, as McDowell puts it in Mind and World, that “the conceptual has no outer boundary.” What is thinkable is identified with what is conceptually contentful. But the objective facts, no less than the subjective thinkings and claiming about them, are understood as themselves already in conceptual shape. The early Wittgenstein also thought of things this way. “The world is everything that is the case, the totality of facts.” And what is the case can be said of it. Facts are essentially, and not just accidentally, things that can be stated. Views with these consequences provide a very friendly environment in which to satisfy the Rational Constraint Condition and so (in the context of a suitable Kantian normative understanding of aboutness) the Mode of Presentation Condition on understandings of the relations between cognitive appearances and the realities of which they are appearances.

The defensibility and plausibility of this sort of approach depend principally on the details of the understanding of the metaconcept of the conceptual (conceptual contentfulness, conceptual articulation) in terms of which it is explicated.⁸ For on some such conceptions, it is extremely implausible and indefensible. For instance, if one’s understanding of concepts is ultimately psychological, then the idea that thoughts (thinkings, believings) and facts might have the same conceptual content would seem to have undesirable consequences. If one thinks that what is in the first instance conceptually contentful is beliefs and thoughts, and that other things, such as visual and auditory sign designs (marks and noises) can count as conceptually contentful only at one remove, by being expressions of beliefs and thoughts, then the claim that the facts those beliefs and thoughts (and derivatively, marks and noises) express (when all goes well) are themselves conceptually contentful threatens to make the existence of those facts (including ones that will never be expressed or represented) objectionably dependent on the existence of thinkings and believings.⁹ The same unfortunate sort of implication results from conjoining the RCC version of the MPC with Davidson’s claim that “[o]nly a belief can justify a belief.” Berkeley claims that the only things we can be understood to represent by our thoughts are other thoughts (the thoughts of God). Some of the British Idealists thought that the reality that appeared to us in thought and belief consisted of the thought of the Absolute—and thought they had learned that lesson from Hegel. More recently, Derrida (using Ferdinand de Saussure’s conceptually pre-Kantian and pre-Fregean terminology) offers a picture of a world consisting only of signifiers, with the only things available to be signified being further signifiers. At this point, things have gone badly wrong. All these arguments involve ignoring what Sellars calls “the notorious ‘ing’ / ‘ed’ ambiguity,” which turns on the distinction between thoughts and beliefs in the sense of acts of thinking and believing and thoughts and beliefs in the sense of what is thought or believed, or what is thinkable or believable. If Hegel’s opening argument has to be filled in in a way that has this sort of idealism as its consequence, we ought to exploit it by modus tollens, not modus ponens.

In fact, though, Hegel’s idea is that the criteria of adequacy for accounts of the relations between appearance and reality that underlie his argument can be satisfied without untoward consequences in the context of quite a different, wholly nonpsychological conception of conceptual contentfulness. The kind of idealism that requires a “world-thinker” on the objective side, no less than a finite thinker on the subjective side, is indeed a reductio. But what it should lead us to reject is not the claim that two-stage representational theories must avoid making strong distinctions of intelligibility between representings and representeds (because they cannot then satisfy the RCC and MPC, and so not the GKC and IEC either) but the psychological conception of conceptual articulation (and hence intelligibility) with which they have been conjoined.

Hegel gets his concept of conceptual content from thinking about Kant’s theory of judgment, and taking on board his understanding of concepts as functions of judgment. Kant understands judging in normative and pragmatic terms. On the normative side, he understands judging as committing oneself, taking responsibility for something, endorsing the judged content. On the pragmatic side, he understands these normative doings in practical terms: as a matter of what one is committed or responsible for doing. What one is responsible for doing is integrating the endorsed content into a constellation of other commitments that exhibits the kind of unity distinctive of apperception. Doing that (“synthesizing” the unity) is extruding from the dynamically evolving unity commitments that are materially incompatible with the new commitment, and extracting and endorsing, so adding, commitments that are its material consequences. Judging that p is committing oneself to integrating p with what one is already committed to, synthesizing a new constellation of commitments that exhibits the rational unity characteristic of apperception. From Hegel’s point of view, that extrusion or expulsion of incompatible commitments and extraction of and expansion according to consequential commitments is the inhalation and exhalation, the breathing rhythm by which a rational subject lives and develops.

Synthesizing a normative subject, which must exhibit the synthetic unity distinctive of apperception, is a rational process because if one judgment is materially incompatible with another, it serves as a reason against endorsing the other, and if one judgment has another as a material inferential consequence, it serves as a reason for endorsing the other. Understanding the activity of judging in terms of synthesis-by-integration into a rational unity of apperception requires that judgeable contents stand to one another in relations of material incompatibility and consequence. For it is such relations that normatively constrain the apperceptive process of synthesis, determining what counts as a proper or successful fulfilling of the judging subject’s integrative task responsibility, or commitment. Concepts, as functions of judgment, determine what counts as a reason for or against their applicability, and what their applicability counts as a reason for or against. Because this is true of all concepts, not just formal or logical ones, the incompatibility and inferential consequence relations the concepts determine must in general be understood as material (that is, having to do with nonlogical content of the concepts), not just logical (having to do with their logical form).¹⁰

I have introduced the idea of conceptual content as articulated by relations of material incompatibility and consequence in Kantian terms of the norms such contents impose on the process of judgment as rational integration: their providing standards for the normative assessment of such integration as correct or successful, settling what one has committed oneself to do or made oneself responsible for doing in endorsing a judgeable content. But I also said that Hegel’s notion of conceptual content is not a psychological one. One could mean by that claim that what articulates conceptual content is normative relations, a matter of what one ought to do, rather than something that can be read immediately off of what one actually does or is disposed to do. That distinction is indeed of the essence for Kant (and for Hegel). But in Hegel’s hands this approach to conceptual content shows itself to be nonpsychological in a much more robust sense. He has the idea that conceptual contentfulness in this sense characterizes not only the process of thinking on the subjective side of the intentional nexus, but also what is thought about, on the objective side.

For objective properties, and so the facts concerning which objects exhibit which properties, also stand in relations of material incompatibility and consequence. Natural science, paradigmatically Newton’s physics, reveals objective properties and facts as standing to one another in lawful relations of exclusion and consequence. That two bodies subject to no other forces collide is materially (nonlogically, because of laws of nature) incompatible with their velocities not changing. That the acceleration of a massive object is changed has as a material consequence (lawfully necessitates) that a force has been applied to it. In the first case, the two ways the world could be do not just contrast with one another (differ). It is impossible—so Newtonian physics, not logic, tells us, hence physically impossible—that both should be facts. And in the second case it is physically necessary—a matter of the laws of physics—that if a fact of the first kind were to obtain, so would a fact of the second kind.

It follows that if by “conceptual” we mean, with Hegel, “standing in relations of material incompatibility and consequence,” then the objective facts and properties natural science reveals as physical reality are themselves in conceptual shape. Modal realism, the claim that some states of affairs necessitate others and make others impossible, the acknowledgment of laws of nature, entails conceptual realism: the claim that the way the world objectively is, in itself, is conceptually articulated. This is a nonpsychological conception of the conceptual in a robust sense, because having conceptual content, standing in relations of material incompatibility and consequence, does not require anyone to think or believe anything. If Newton’s laws are true, then they held before there were thinkers, and would hold even if there never were thinkers. The facts governed by those laws—for instance, early collisions of particles—stood in lawful relations of relative impossibility and necessity to other possible facts, and hence on this conception of the conceptual had conceptual content, quite independently of whether any subjective processes of thinking had gone on, were going on, or ever would go on (in this, or any other possible world).

As I am using the term, a “psychological” theory of the conceptual understands concepts as something like mental particulars, or aspects of mental particulars: as essentially features of psychological or intentional states, paradigmatically thinkings and believings. Hegel’s nonpsychological understanding of the conceptual, as a matter of standing in relations of nonlogical incompatibility and consequence allows for psychological and intentional states and episodes to count as conceptually contentful, but does not restrict the applicability of conceptual predicates to such states and episodes. It is important to keep this point firmly in mind when considering his conceptual realism. For the result of conjoining conceptual realism about the objective world with a psychological understanding of the conceptual is a kind of Berkeleyan idealism, according to which objective facts require a world-thinker whose thinkings they are.¹¹ This is emphatically not Hegel’s thought (nor is it Frege’s, Wittgenstein’s, or McDowell’s)—although his use of the term “Weltgeist” (which appears three times in the Phenomenology) has misled some (including some of his admirers, such as Royce, and even Bradley) on this point. I say something later about how else we might understand his remarks in the Preface about the necessity of “construing Substance also as Subject.”

Hegel thinks that underlying this point about the conceptual character of objective reality is a deeper one. For he thinks that the idea of determinateness itself is to be understood in terms of standing in relations of incompatibility and consequence to other things that are determinate in the same sense. He endorses Spinoza’s principle Omnis determinatio est negatio. For something to be determinate is for it to be one way rather than another. This thought is incorporated in the twentieth-century concept of information (due to Shannon¹²), which understands it in terms of the partition each bit establishes between how things are (according to the information) and how they are not. Everyone would agree, I take it, that if a property does not contrast with any properties, if it is not even different from any of them, then it is indeterminate. To know that an object had such a property would be to know nothing about it. Beginning already in the Perception chapter of the Phenomenology, Hegel argues that determinateness requires more than mere difference from other things. It requires what he calls “exclusive” [ausschließend] difference, and not mere or “indifferent” [gleichgültig] difference. Square and circular are exclusively different properties, because possession by a plane figure of the one excludes, rules out, or is materially incompatible with possession of the other. Square and green are merely or indifferently different, in that though they are distinct properties, possession of the one does not preclude possession of the other. An essential part of the determinate content of a property—what makes it the property it is, and not some other one—is the relations of material (nonlogical) modally robust incompatibility it stands in to other determinate properties (for instance, shapes to other shapes, and colors to other colors). We can make sense of the idea of merely different properties, such as square and green, only in a context in which they come in families of shapes and colors whose members are exclusively different from one another.

An important argument for understanding determinateness Hegel’s way, in terms of exclusive difference or material incompatibility (one pursued in the Perception chapter), is that it is required to underwrite an essential aspect of the structural difference between the fundamental ontological categories of object and property (particular and universal). Aristotle had already pointed out a structural asymmetry between these categories. It makes sense to think of each property as coming with a converse, in the sense of a property that is exhibited by all and only the objects that do not exhibit the index property. Has a mass greater than five grams is a property that has a converse in this sense. But it does not make sense to think of objects as coming with converses, in the analogous sense of an object that exhibits all and only the properties that are not exhibited by the index object. This is precisely because some of those properties will be incompatible with one another, and so cannot be exhibited by a single object. The number 9 has the properties of being a number, not being prime, being odd, and not being evenly divisible by 5. If it had a converse, that object would have to have the properties of not being a number, being prime, being even, and being divisible by 5. But nothing can have all of those properties.

It follows that a world that is categorially determinate, in that it includes determinate properties (and relations) and objects (distinguishable by their properties and relations), so facts (about which objects exhibit which properties and stand in which relations) must be determinate in Hegel’s sense: the properties must stand to one another in relations of material incompatibility. If they do that, they will also stand to one another in relations of material consequence, because a property P will have the property Q as a consequence if everything incompatible with Q is incompatible with P. So being a bear has being a vertebrate as a consequence, because everything incompatible with being a vertebrate—for instance, being a prime number—is incompatible with being a bear.

Because Hegel understands being conceptually contentful as standing to other such items in relations of material incompatibility and consequence, to take the objective world to be minimally determinate, in the sense of consisting of facts about what objects have what properties (and stand in what relations to each other), is to take it to be conceptually structured. For him, only conceptual realists are entitled to think of objective reality as so much as determinate. (Modal realism comes for free. We didn’t need Newtonian physics to get to conceptual realism in this sense; the barest Aristotelian metaphysics is already enough.) This conception of the conceptual is nonpsychological in a very strong sense.

IV. Alethic Modal and Deontic Normative Material Incompatibility

In this sense, there is no problem seeing both sides of the appearance / reality distinction as conceptually structured.¹³ So we are not on that account obliged to excavate a gulf of intelligibility between them. For the same reason, the principal obstacle to satisfying the Rational Constraint Condition, and therefore the Mode of Presentation Condition, is removed. (Though I haven’t said anything positive about how they might be satisfied, either.) That means in turn that the semantic presuppositions that I have been reading Hegel as taking to make it impossible to satisfy the epistemological criteria of adequacy expressed by the Genuine Knowledge Condition and the Intelligibility of Error Condition—namely, the strong difference of intelligibility of representings and represented—can also be avoided. Access to all of these desirable consequences is to be opened up by the nonpsychological structural understanding of the conceptual, and so of intelligibility, in terms of relations of material incompatibility and (so) consequence.

Hegel’s term for what I have been calling “material incompatibility” is “determinate negation” [bestimmte Negation].¹⁴ His term for what I have been calling “material consequence” is “mediation” [Vermittlung]—after the role of the middle term in classical syllogistic inference.¹⁵ The first is the more fundamental concept for Hegel—perhaps in part because, as I argue in the previous section, wherever there are relations of incompatibility, there will also be relations of consequence. Hegel often contrasts determinate negation (material incompatibility) with “formal” or “abstract” negation (logical inconsistency): square is a (not the) determinate negation of circular, whereas not-circular is the (not a) formal negation of it. (These are Aristotelian contraries, rather than contradictories.) We are in a position to see that the choice of the term “determinate” to mark this difference is motivated by Hegel’s view that it is just relations of determinate negation in virtue of which anything is determinate at all. This is as true of thoughts as it is of things—of discursive commitments on the side of subjective cognitive activity no less than of facts on the side of the objective reality the subject knows of and acts on. That is why, though the conception is at base nonpsychological, Hegel’s metaconcept of the conceptual does apply to psychological states and processes. Thinkings and believings, too, count as determinately, and so conceptually contentful, in virtue of standing to other possible thinkings and believings in relations of material incompatibility and consequence.

But are subjective commitments conceptually contentful in the same sense that objective facts are—even given Hegel’s definition? When we say that being pure copper and being an electrical insulator are materially incompatible we mean that it is (physically, not logically) impossible that one and the same object, at one and the same time, has both properties. But when we say that the commitments to a’s being pure copper and a’s being an electrical insulator are materially incompatible, we do not mean that it is impossible for one and the same subject, at one and the same time, to undertake both commitments. We mean rather that one ought not to do so. That “ought” has the practical significance that violating it means that one is subject to adverse normative assessment, that any subject with two commitments that are materially incompatible in this sense is obliged to do something, to relinquish (or modify) at least one of them, so as to repair the inappropriate situation. But it is entirely possible for a subject to find itself in this inappropriate normative situation. There is a similar disparity on the side of consequences. If conducting electricity is objectively a consequence of being pure copper, then it is necessary that any object that has the one property (at a time) has the other (at that time). But if one acknowledges a commitment to some object’s being pure copper, it is still possible that one not acknowledge commitment to that object’s conducting electricity. It is just that one ought to.

This is to say that the relations of material incompatibility and consequence in virtue of which objective facts and properties are determinate are alethic modal relations: a matter of what is conditionally impossible and necessary. The relations of material incompatibility and consequence in virtue of which the commitments undertaken and predicates applied by discursive subjects are determinate are deontic normative relations: a matter of what one is conditionally entitled and committed to. We may think of these as alethic and deontic modalities, if we like, but they are still very different modalities. Hegel is writing downstream from Kant’s use of “necessity” [Notwendigkeit] as a genus covering both cases. “Notwendig” for Kant means “according to a rule.” He can accordingly see “natural necessity” and “practical necessity” as species of one genus. (They correspond to different uses of the English “must.”) Nonetheless, these are very different modalities, substantially different senses of “necessary” (or “must”). The worry accordingly arises that two quite distinct phenomena are being run together, and that the attempted assimilation consists of nothing more than the indiscriminate use of the same verbal label “conceptual.”

One of the metacommitments for which I claimed Kant’s authority is that to be intelligible (in a successor-sense to Descartes’s) is to be conceptually structured or—what on this broadly structuralist-functionalist account of content amounts to the same thing—conceptually contentful. Once again following Kant, Hegel understands understanding (and so intelligibility) in ultimately pragmatic terms: as a matter of what one must be able practically to do to count as exercising such understanding. What one must do in order to count thereby as grasping or understanding the conceptual content of a discursive commitment one has undertaken (or is considering undertaking) is be sensitive in practice to the normative obligations it involves. That means acknowledging commitments that are its consequences, and rejecting those that are incompatible with it. This is, in one sense, immediate intelligibility of commitments, in that it is commitments that one acknowledges, and so has in the first instance attitudes toward. In another sense, of course, this sort of intelligibility is not at all immediate, because it is mediated by the relations to all the other possible commitments, whose relations of material incompatibility and consequence articulate the content acknowledged.

What about the intelligibility of objective states of affairs, which are conceptually contentful in virtue of the alethic modal connections of incompatibility and consequence they stand in to other such states of affairs, rather than the deontic normative relations that articulate the conceptual content of discursive commitments (which are “immediately” intelligible in that practical sense I’ve just been talking about)? The key point is that what one needs to do in order thereby to count as practically taking or treating two objective states of affairs (or properties) as alethically incompatible is to acknowledge that if one finds oneself with both the corresponding commitments, one is deontically obliged to reject or reform at least one of them. And what one needs to do in order thereby to count as practically taking or treating one objective state of affairs as a necessary (lawful) consequence of another is to acknowledge the corresponding commitment to one as a consequence of the corresponding commitment to the other. Here “corresponding” commitments are those whose deontic normative conceptual relations track the alethic modal conceptual relations of the objective states of affairs. Isomorphism between deontic normative conceptual relations of incompatibility and consequence among commitments and alethic modal relations of incompatibility and consequence among states of affairs determines how one takes things objectively to be. Practically acquiring and altering one’s commitments in accordance with a certain set of deontic norms of incompatibility and consequence is taking the objective alethic modal relations articulating the conceptual content of states of affairs to be the isomorphic ones.¹⁶

Because of these relations, normatively acknowledging a commitment with a certain conceptual content is taking it that things objectively are thus and so—that is, it is taking a certain fact to obtain. And that is to say that in immediately grasping the deontic normative conceptual content of a commitment, one is grasping it as the appearance of a fact whose content is articulated by the corresponding (isomorphic) alethic modal relations of incompatibility and consequence. This is how the Mode of Presentation Condition is satisfied in this sort of two-stage representational model while eschewing a strong distinction of intelligibility. The Rational Constraint Condition is satisfied, because if the subject is asked why—that is, for what reason—one is obliged to give up a commitment to Q(a) upon acknowledging a commitment to P(a) (something we express explicitly by the use of deontic normative vocabulary), the canonical form of a responsive answer is: Because it is impossible for anything to exhibit both properties P and Q (something expressible explicitly by the use of alethic modal vocabulary).¹⁷ And similarly for consequential relations among commitments.

The Genuine Knowledge Condition is satisfied on this model. For taking thoughts to stand in representational relations to facts implicitly involves commitment to the possibility of an isomorphism between subjective normative relations of incompatibility and consequence and objective modal relations of incompatibility and consequence, and the model does not semantically preclude such an isomorphism from holding objectively—at least locally and temporarily.¹⁸ The model also makes sense of the possibility of error (it satisfies the Intelligibility of Error Condition). For, following Kant, it construes the representation relation in normative terms. In manipulating (acquiring and rejecting) commitments according to a definite set of conceptual norms (deontic relations of incompatibility and consequence), one is committing oneself to the objective modal facts (alethic relations of incompatibility and consequence) being a certain way—as well as to the ground-level empirical determinate facts they articulate being as one takes them to be. So the model also says what must be the case for that isomorphism relation to fail to hold in fact. Then one has gotten the facts wrong—perhaps including the facts about what concepts articulate the objective world.

In this chapter I have aimed to do six things:

To demarcate explicitly the exact range of epistemological theories, epitomized by those of Descartes and Kant, that fall within the target area of Hegel’s criticism.
To set out clearly the objection that he is making to theories of that kind.
To formulate Hegel’s criteria of adequacy for a theory that would not be subject to that objection he is implicitly putting in play.
To introduce the nonpsychological conception of the conceptual that will form the backbone of Hegel’s response (even though it is not officially introduced in the Introduction itself, but must wait for the opening chapters of Consciousness).
To sketch the general outlines of an epistemological and semantic approach based on that conception of the conceptual.
To indicate how such an approach might satisfy the criteria of adequacy for a theory that is not subject to Hegel’s objection.

The next chapter looks more closely at the account of representation that I take Hegel to construct out of the elements put in play by this discussion.

Chapter 1

Conceptual Realism and the Semantic Possibility of Knowledge

I. Classical Representational Epistemology

II. Genuine Knowledge and Rational Constraint

III. A Nonpsychological Conception of the Conceptual

IV. Alethic Modal and Deontic Normative Material Incompatibility