Kant on Pure Apperception and Indeterminate Empirical Inner Intuition

2.1. Pure Apperception

Pure apperception is the nonsensible self-consciousness that is essentially related to the I think, an act of self-ascription of mental states (B132).⁵ The I think, portrayed as the “act” of pure apperception (B137), arises from this self-consciousness.⁶ A representation can be “my representation,” i.e., cognitively meaningful for me (B131–132, A116) or conscious (A117n., B133), if and only if it can be accompanied by the I think (B131–132, A116, A117n.).⁷ The I think is a “concept, or rather, if one prefers, [a] judgment” (A341/B399), and contains the “I think,” the “formal proposition” (A354) that expresses pure apperception (A398–399). In contrast, the representation I is not a concept (Prol 4: 334; MAN 4: 543). Pure apperception is intellectual: it arises from understanding (B136, 140) and does not contain any sensible content. Its product, the I think, consists of the I and a representation of thinking, or, equivalently, a representation of spontaneous cognitive mental acts.⁸ Neither rests upon sensibility: the I is empty (A345/B404f., B135, A77/B102), such that it contains no sensible manifold (B135, A355, A381–382), and the representation of thinking is intellectual (B278; NTKrV 23: 39; Anth 7: 141). Pure apperception is transcendental due to its epistemic role: it combines all (intrinsically) separate episodes of consciousness (which include acts of thinking) into one consciousness and, hence, enables all possible functions of understanding (A341–342/B400, A117n, B133–134; Prol 4: 334n.).

Consequently, pure apperception does not count as an experience (i.e., an empirical judgment about spatial or temporal objects); instead, it “precedes” any experience (A354, B276–277).⁹ This is because the I think of pure apperception does not represent the subject as an object located “anywhere or at any time” (NTKrV 23: 39; B407). The I think is the “consciousness of myself in mere thought” (B429), with “mere thought” indicating “merely the logical function and hence the sheer spontaneity” (B428) of mental acts. As a mere thought, the I think is detached from any relation to an inner intuition (B429). It thus represents merely an “object in general” (B429) that lacks any objective reality (B148–149). The “I” in the I think, as a mere thought, does not refer to any object of empirical cognition.

2.2. Inner Sense

The term “inner sense” denotes either a type of self-consciousness or the faculty for this self-consciousness. As used in the first sense, “inner sense” means one’s empirical intuition of one’s inner states (A33/B49, A107, A176/B217; V-Met-K2/Heinze 28: 738).¹⁰ Inner states include, in addition to perceptions and feelings of pleasure and displeasure, thoughts (A357–358) and desires (Refl 17: 366). The form of inner sense is time. As such, inner sense is the empirical consciousness of one’s being in a particular inner state at a particular time.¹¹ Note that the term “inner sense” is approximately synonymous with “empirical apperception” (A107; Anth 7: 134n.; V-Met/Dohna 28: 673; V-Met-K2/Heinze 28: 713).¹²

When grounded on inner sense, the I think becomes distinct from the I think as a mere thought in many aspects:

(Q1) [1] [T]he proposition “I think,” insofar as it says so much as that I exist thinking, is not merely a logical function, but rather determines the subject (which is then at the same time an object) in regard to existence, [2] and this cannot take place without inner sense, whose intuition always makes available the object not as thing in itself but merely as appearance. [3] Thus in this proposition there is already no longer merely spontaneity of thinking, but also receptivity of intuition, i.e., the thinking of my self applied to the empirical intuition of the very same subject. (B429–430, emphasis mine; similarly, B428–429, B430, B277)

Accordingly, the I think as a mere thought can be applied to an empirical intuition produced by inner sense. As explained in fn 10, an inner perception cannot be directly about the self, but is rather about inner states. The inner perception upon which the I think is grounded is an empirical intuition of thinking.¹³ Its object is an appearance (Q1.2). Appearances are “objects of a possible experience” (A238/B298). Thus, the I think, when grounded on inner sense, constitutes an “inner experience” (B276–277, BXLn.),¹⁴ i.e., an empirical cognition about an inner object. The fact that this kind of I think constitutes an inner experience implies that it can determine the corresponding inner intuition (i.e., DEI) through some categories (as Q1.1 above confirms).¹⁵ This kind of I think amounts to an empirical judgment about the subject’s state of thinking. It is no longer an a priori “mere thought,” and the “I think” it contains counts as an “empirical proposition” (B428). As an ordinary empirical judgement that is itself an episode of consciousness, it cannot have any consciousness-unifying function and hence cannot play a transcendental role.

2.3. Indeterminate Empirical (Inner) Intuition (IEI)

Some peculiar remarks cannot be accommodated by what has been said thus far:

(Q2) [1] We must not be troubled by the fact that in this proposition [“I think”], which expresses the perception of oneself, I do have an inner experience, [2] and that therefore the rational psychology built on it is never pure but is based in part on an empirical principle. [3] For this inner perception is nothing more than the mere apperception [bloße Apperzeption] I think that makes even all transcendental concepts possible, since in them we say: I think substance, cause, etc. (A342–343/B400–401)

In this passage, Kant speaks of “the mere apperception” that “makes even all transcendental concepts possible.” This is precisely what pure apperception does (A341/B399). Further evidence shows that the term “mere apperception” denotes pure perception (A546/B574, A400). It is thus perplexing that Kant regards pure apperception as inner perception, since pure apperception should be free from any sensible element. In the same way, it is puzzling that the formal proposition of pure apperception, the “I think,” expresses this “perception of oneself.” This is not the only puzzling passage concerning pure apperception in Kant’s Critical writings. In the Critique of Pure Reason (KrV), he characterizes the I think as an “empirical inner representation” (A848/B876, similarly B421). At MAN 4: 543, he maintains that “the thought I…is no concept at all, but only inner perception.” Famously, he describes the I as a “feeling of an existence” (Prol 4: 334n.). In all such passages (henceforth, “target passages”), pure apperception counts as an empirical representation containing or relying on sensation.

A deflationary explanation of the target passages would be that the I think here is a genuine empirical judgment based on the inner perception that is produced by inner sense (DEI). It is thus the type of I think addressed in Q1, rather than I think as a mere thought. Accordingly, Kant’s equation of pure apperception with inner perception in Q2.3 would be ascribed to a sloppy way of formulation: he should have written “this mere apperception which expresses an inner perception,” just as he does in Q2.1 and in Q3.2 below.

I agree that the equation of these two concepts is probably a slip; otherwise, we would have to accept that a representation could simultaneously be a thought and a perception.¹⁶ Nevertheless, a significant difference separates the I think in Q2 from the I think involved in Q1. The I think in Q2 is presupposed by all transcendental concepts (Q2.3; see also A341/B399). Hence, it is also presupposed by the transcendental synthesis of imagination according to these concepts. Since this synthesis affects inner sense (B153–154),¹⁷ the I think in Q2 is presupposed by inner sense and its product, DEI. In contrast, the I think in Q1 is grounded on inner sense and DEI (Q1.2). Because DEI is generated by the self-affection transcendental synthesis of imagination provokes, the I think in Q1 depends on and occurs “after” this synthesis.

Sceptics of a third type of I think might point to a sentence in Q2.1: “in this proposition [‘I think’]…I do have an inner experience.” This suggests that Q2 involves the same kind of I think as Q1, because an inner experience is an empirical judgment that results when understanding is applied to, and hence conceptually determines, an inner perception.¹⁸ Nevertheless, “experience,” employed in a weak, nontechnical sense, could also denote the act or process of having epistemic contact with an object through sensations (B1, BXVII, A11; Anth 7: 128).¹⁹ Therefore, the quotation above could merely mean that the I think is related to experience in this nontechnical sense, rather than that it conceptually determines an inner perception, thus constituting a genuine empirical judgment.²⁰

A contentious footnote in KrV provides decisive evidence for a third type of I think:

(Q3) [1] The “I think” is, as has already been said, an empirical proposition, and contains within itself the proposition “I exist.” … [2] It expresses an indeterminate empirical intuition, i.e., perception²¹ (hence it proves that sensation, which consequently belongs to sensibility, grounds this existential proposition), but it precedes the experience that is to determine the object of perception through the category in regard to time; [3] (a) and here existence is not yet a category, (b) which is not related to an indeterminately given object, (c) but rather to an object of which one has a concept, and about which one wants to know whether or not it is posited outside this concept. [4] An indeterminate perception here signifies only something real, which was given, and indeed only to thinking in general, thus not as appearance, and also not as a thing in itself (a noumenon), but rather as something that in fact exists and is marked as an existing thing in the proposition “I think.” [5] For it is to be noted that if I have called the proposition “I think” an empirical proposition, I would not say by this that the I in this proposition is an empirical representation; for it is rather purely intellectual, because it belongs to thinking in general. [6] Only without any empirical representation, which provides the material for thinking, the act I think would not take place, and the empirical is only the condition of the application, or use, of the pure intellectual faculty. (B422–423n., bracketed numbers mine)

Apparently, Q3 and Q1 address the same type of I think because both concern the empirical proposition “I think” that contains the proposition “I exist,”²² and because in both cases the “I think” expresses, and is grounded on, inner perception representing an act of thinking.²³ Nevertheless, compared with Q1, in crucial respects the I think in Q3 is distinct and manifests some surprising peculiarities that go beyond the conceptual framework of KrV.

First, although both relate to an inner perception, the inner perception in Q1 represents an appearance (Q1.2), whereas that in Q3 represents neither an appearance nor a noumenon (Q3.4). This undermines Kant’s basic distinction between phenomenon and noumenon.²⁴ An empirical intuition, even if indeterminate, should represent an appearance, which constitutes its intentional object (A20/B34).

Second, the I think in Q1 can, in some sense, determine the subject according to the category of existence (Q1.1; §4.4.2),²⁵ whereas this is clearly not the case in Q3 (Q3.2, Q3.3).

Third, the I think in Q1 constitutes an inner experience (Q1.2), whereas the I think in Q3 does not, as it “precedes the experience that is to determine the object of perception through the category in regard to time” (Q3.2; henceforth, “precedence claim”).²⁶

All these peculiarities involve a special indeterminacy of the IEI.

The I think in Q3 is instead the same type as that in Q2. ²⁷ In Q3.1, Kant reminds his reader of previous passages where he classifies the “I think” as an empirical proposition, and Q2 fits this characterization.²⁸ Moreover, in the main text where Q3 is provided in a footnote, Kant addresses “the unity of consciousness, which grounds the categories” (B421). Clearly, Kant has in mind here the I think that plays a transcendental role. In this respect, the I think of Q3 is closer to Q2’s than it is to Q1’s, where the I think cannot play a transcendental role (see §2.2), but is instead an introspective representation of an act of thinking (see §4.3).

Consequently, Q3 and Q2 plausibly involve a special kind of inner intuition (i.e., IEI) that is distinct from the one (i.e., DEI) in Q1.^{29, 30} Understanding the IEI’s characteristics and its locus in Kant’s doctrine of mind is decisive for clarifying the target passages. I now evaluate the major interpretations of IEI before offering a novel interpretation.

3.1. Intellectual Intuition

Famously, Rolf-Peter Horstmann construes the IEI as the forerunner and origin of the Fichtean “Tathandlung” (Horstmann 1993: 425). The IEI is an “analytic implicature of the act of thinking” (425). It is not “something that is (like an object), but…something which, in the event of thinking, as it were realizes itself” (425). Consequently, IEI is an intellectual intuition of the self.³¹ I will not engage with this approach extensively since it needs textual support and contradicts Kant’s dichotomy of sensibility and understanding, his explicit rejection of human intellectual intuition (B135), and his repeated characterization of IEI as empirical. I would merely add one point to Rosefeldt’s (2000: 226–227) and Klemme’s (1996: 385–386) thorough criticism of this reading. In Q3.4, Kant asserts that the “indeterminate perception” is “given…only to thinking in general” (B423n.). One could suggest that this shows that IEI is an intuition given to the faculty of thinking, i.e., understanding, rather than to sensibility.³² However, the phrase “thinking in general” is a technical term that refers not to the faculty of thinking but rather to thoughts considered under abstraction from their individual objects.³³ Thinking in general, which includes categories (Br 11: 314; Refl 18: 229), is the form of thinking (Log 9: 13) and is a priori (A382; NTKrV 23: 38). In Q3.5, Kant uses this term again: “for it [the I] is rather purely intellectual, because it belongs to thinking in general” (emphasis mine). If the phrase “thinking in general” referred to understanding, one would infer that empirical thoughts are “purely intellectual” because they belong to understanding.

The phrase “[the IEI’s being] given … only to thinking in general” indeed means that IEI provides empirical materials for the thought I think, which, considered alone, belongs to thinking in general (B399). This reading fits the context of Q3.4 and Q3.5 better—i.e., Q3.6, which explains their meaning.³⁴ In contrast, the “intellectual intuition” reading could hardly accommodate Q3.6.

3.2. Perception of the I Think

IEI has also been seen as a perception of the I think that constitutes the transcendental awareness accompanying all “my” representations, and that enables all acts of understanding (Emundts 2006: 304–306; Kraus 2020: 126–127). Thus, in contrast to DEI, IEI is a perception of the epistemically privileged representation, the I think, rather than a perception of some ordinary representation (Emundts 2006: 304). Admittedly, this reading is favorable for explaining the IEI’s indeterminacy: One could ascribe the IEI’s peculiarity to the special status of the I think. However, because the “I think” expresses and thus presupposes IEI (Q2.1, Q3.2), IEI cannot be a perception of the I think.

3.3. Bare Sensation or Feeling

Some commentators maintain that in referencing “indeterminate empirical intuition,” Kant has in mind not an object-related representation but a bare sensation (henceforth, “initial sensation,” following Banham 2011: 121).³⁵ For Kant, sensation is, in contemporary parlance, a qualitative feel, i.e., a subjective experience³⁶ of a certain phenomenal quality. Sensations are either objective or subjective (KU 5: 206). Objective sensations, e.g., “the green color of a meadow,” can constitute a “representation of an object” (KU 5: 206). In contrast, the faculty of pleasure and displeasure generates subjective sensations, e.g., “[the green color’s] agreeableness,” which cannot constitute an objective representation (KU 5: 206). Some advocates of this reading consider initial sensation to be an objective sensation.³⁷ Motivated by Kant’s claim that the I is “a feeling of an existence” (Prol 4: 334), others hold that initial sensation is a subjective sensation.³⁸

Advocates of both views rely on a distinct property of sensation:

[Q4] Sensation in itself is not an objective representation, and in it neither the intuition of space nor that of time is to be encountered, it has, to be sure, no extensive magnitude, but yet it still has…an intensive magnitude. (B208, similarly see Prol 4: 306)

Since sensation in itself does not involve the intuitions of space and time, “the real” of a sensation³⁹ is “not given as a received object” in the sensation (Reuscher 1975: 285), and “does not register as phenomenon” (Banham 2011: 121). This naturally leads to the conclusion that “the real” which is given in sensations is “not a manifold…[and] not subject to the categories” (Reuscher 1975: 285). Therefore, this interpretation can accommodate IEI’s categorial indeterminacy and its character of not representing appearance. Nevertheless, this reading contradicts textual evidence. Objective sensations are certainly “not an intuition” (Prol 4: 306) and feelings are even less so, whereas IEI is declared as an intuition in Q2 and Q3. Moreover, immediately after Q4, Kant says that “apprehension, merely by means of sensation, fills only an instant (if I do not take into consideration the succession of many sensations)” (A167/B209, emphasis mine). Evidently, instants are products of theoretical abstraction. They are only imaginary temporal “boundaries” and are not real (A169/B211). Accordingly, bare sensation, which occupies “only an instant,” must be an imaginary mental item. The phrase “sensation in itself” in Q4 also indicates that it involves a theoretical abstraction, one which seeks to illustrate intensive magnitude. Finally, IEI is “related to an indeterminately given object” and “signifies…something real.” IEI’s object-relatedness, despite being indeterminate, contradicts the description in Q4.

The “subjective sensation” reading suffers from additional problems. Kant did not develop the notion “feeling of an existence” after the Prolegomena. Likewise, the expression “feeling of an existence” does not necessarily designate a radically subjective sensation. Sufficient evidence shows that “feeling” can also denote objective sensation in Kant’s Critical period before he regulates the terms “sensation” and “feeling” at KU 5: 207.⁴⁰ Therefore, “feeling of an existence” could literally signify a sensation of an object (more specifically, its states; see fn 10). The claim that “the I…is nothing more than a feeling of an existence” could accordingly mean that objective sensations of an existence’s states underlie the I. It does not preclude the possibility that such feelings can constitute an indeterminate empirical intuition. Therefore, that claim stands in line with Q2 and Q3.

3.4. Sensory Materials for the Act of Thinking

Many commentators regard Q3.6 as crucial for deciphering the mysteries of IEI: the “I think” is an empirical proposition because it expresses an act of thinking, for which some empirical representations are required to supply materials.⁴¹ In other words, there must be empirical data to provide content for the “that” clause following the “I think.”⁴² Accordingly, the IEI in Q3.2, which is expressed by the “I think” and is responsible for its empirical character, is rendered as a perception that provides sensory content for the “that” clause, rather than as a perception of the act of thinking.⁴³

Although this reading is dominant in the literature, it is untenable in crucial respects. First, a single instance of I think does not necessarily presuppose empirical materials, because the content of the “that” clause could be pure (B3), e.g., it could concern God or logic. For an I think with such a “that” clause, no corresponding IEI can be found. Thus, the suggested reading of Q3.6 can be improved as follows: “the I think is empirical, because instances of I think as a whole depend on empirical representations.” Nevertheless, an empirical representation is defined as one that itself contains or relies on sensations (A20/B34). Additionally, if the inference underlying the modified reading were correct, then logical propositions would be empirical, because logical reasoning as a whole will not take place without empirical data.⁴⁴

The sensory materials required for the “that” clause could be outer or inner perceptions. Therefore, if the third approach were correct, then IEI could be an outer perception. However, the sensation of IEI grounds the “I exist” (Q3.2). What IEI represents is “mark[ed]” [bezeichnet] as an “in fact exist[ing]” thing in the “I think” (Q3.4). What is expressed by the “I think” and explains its empirical character is “the perception of oneself” (Q2.1). All these quotations imply that the sensory content of IEI concerns inner rather than outer objects. IEI can only be an inner perception that represents an act of thinking, as demonstrated in fn 23. Furthermore, if IEI were an outer perception, then it would be inexplicable why an outer perception could manifest the peculiar features mentioned in Q3. How could an outer perception not represent any appearance? How could an outer perception that is expressible in propositional form (i.e., in a “that” clause) be indeterminate? As far as I am aware, no commentator has satisfactorily addressed these questions.⁴⁵

4.1. Revisiting the Nature of the I Think

This section revisits the nature of the I think and demonstrates three propositions: 1) as an actual mental state, the I think is either clear or obscure; 2) as such, it must be grounded on an empirical intuition (i.e., either IEI or DEI), and cannot be a “mere thought” (see §2.1), which is not accompanied by any empirical intuition; and 3) the I think based on IEI is ubiquitous in conscious states. Subsequent sections build on these propositions.

I will first demonstrate that the I think occurs only in two modes of consciousness. Kant engages with this issue in a footnote:

[Q5] [1] The mere representation I in relation to all others (the collective unity of which it makes possible) is the transcendental consciousness. [2] Now it does not matter here whether this representation be clear (empirical consciousness) or obscure, even whether it be actual [Wirklichkeit]; [3] but the possibility of the logical form of all cognition necessarily rests on the relationship to this apperception as a faculty. (A117n., emphasis mine)

The italicized text indicates that the I think, if it “actually” occurs, is either a clear or an obscure representation. For Kant, a representation is clear if and only if the subject is conscious of having it, i.e., of being in this representational state. A representation is obscure if and only if the subject represents something but is not conscious of being in this representational state (Anth 7: 135ff.).⁴⁶ Such a representation is something present in one’s mind, since it can be accessed and can contribute to cognition despite being an unconscious inner state.⁴⁷ In both modes, the I think counts as “actual.” Labeling something “actual” means that it is an “existence at a determinate time” (A145/B184).

The I think in both modes is actual for two reasons. First, because the I think, whether clear or obscure, is a representation, which is a “modification” (B242, A97) or “determination” of mind (B74, B242; LB 30). As such, the I think is a state of mind, a mental occurrence at a determinate time. Second, the predication of actuality to an object presupposes that this object stands in connection with “some actual perception” (A225/B272). Both modes of the I think meet this requirement. The I think that is clear (or, roughly, conscious) is accompanied and represented by an inner perception (see “empirical consciousness” in Q5.2 and fn 46). The I think that is obscure (or, roughly, unconscious) can become the object of an inner perception and thus become clear if one introspects, i.e., directs attention inward to examine one’s own thoughts.⁴⁸

As an actual mental state, the I think must be an empirical representation, and the proposition “I think” that it contains must be empirically founded. The reason for this is as follows: the spontaneous mental act that the I think represents occurs in time (A103, A99; NTKrV 23: 18).⁴⁹ The I think, whether clear or obscure, accompanies an individual, spontaneous mental act at some moment, and the “I think” contained within the I think refers to this act.⁵⁰ According to the intuition–concept dichotomy (A19/B33), the relation of a representation to an individual object can be established only through empirical intuition (see also NTKrV 23: 26). Therefore, the I think, as an actual mental item, must rest upon an empirical intuition of that spontaneous cognitive act which is responsible for the representation that the I think accompanies. As §§2.2 and 2.3 demonstrate, the empirical intuition is either IEI or DEI. In other words, the I think must be an empirical representation based on either IEI or DEI. This implies that there is no actual occurrence of the I think as mere thought or logical function, as depicted in §2.1. Consequently, part of the puzzle surrounding the target passages (see §2.3) can be clarified.

The conclusions we have drawn so far in this section can also cast light on IEI. There is a kind of I think that is ubiquitous in—i.e., permanently built into—conscious states:⁵¹ the I “accompanies—and indeed with complete identity—all representations at every time in my consciousness” (A362–363; see also A123). The I think is “in all consciousness … one and the same” (B132). This “one and the same” I is “the formal element of consciousness” (Anth 7: 141; A382); this implies that it cannot be separated from consciousness.⁵² Clearly, this ubiquitous I think represents the transcendental condition of consciousness. Such a kind of I think cannot be the I think depicted in Q1, because the latter cannot play any transcendental role (see §2.2). It can only be the I think characterized in Q2. Accordingly, the empirical inner intuition underlying it must be IEI rather than DEI.⁵³ In conclusion, the IEI underlying it is also ubiquitous in consciousness.

One may object that these quotations, as interpreted above, contradict both the apperception principle (see §2.1) and Q5.2, since this principle does not require every cognitively meaningful representation to be actually accompanied by an instance of I think. The mere possibility of such an accompaniment is sufficient.⁵⁴ However, the principle concerns all “my” representations rather than all “my” conscious representations. It does not preclude the possibility that every episode of (attentional) consciousness contains an instance of I think. The statement that “it does not matter … even whether it [the I think] be actual” (Q5.2) indeed alludes to some obscure representations. Such representations are not accompanied by (apperceptive) consciousness and an instance of I think, yet they are cognitively meaningful as they can be accompanied by one such instance indirectly.

4.2. Revisiting Pure Apperception

In this section I argue that pure apperception is content awareness rather than act awareness. Pure apperception, as such, results from a theoretical abstraction that captures the contribution of mere intellect in self-consciousness. It is a mental item shared by all conscious states rather than a self-standing mental state. These considerations, combined with §4.1, clarify the relation between I think, pure apperception, and IEI. This in turn explains the crucial contradiction in Q2, and sets the stage for understanding IEI.

In some texts, Kant asserts that pure apperception with the formal content “I think” is an immediate intellectual (or pure) consciousness of acts of thinking or synthesis (B133–134, B278, B153; Anth 7: 141–142; V-Met-K2/Heinze 28: 712–713).⁵⁵ This intellectual awareness seems to be present in some significant sense, since the subject can “combine transcendental consciousness” with thoughts (Refl 18: 611; see also B419, A123–124), which are mental occurrences in time. It represents one’s acts of thinking unmediated by an inner empirical intuition, i.e., through the mere thought I think.

A natural way of construing the notion of intellectual consciousness of acts of thinking is to read it as intellectual consciousness of mental acts as mental acts.⁵⁶ In contrast to perceiving an object, we cannot be intellectually aware of a mental act vaguely as “something there.” To be intellectually aware of a mental act as a mental act is to think of it as a mental act, because there is no other intellectual way of representing a mental act as such. On closer examination, nonetheless, this prima facie construal of the notion is problematic, as it cannot be coherently integrated into Kant’s system. First, if one had intellectual act awareness of one’s thoughts—i.e., if one were nonsensibly aware of one’s thoughts as acts—then one would have intellectual awareness of one’s existence.⁵⁷ Kant rejects intellectual awareness of the self’s existence⁵⁸ (directly at B426–427 and indirectly at B420⁵⁹), and treats it as the result of an abstraction (B426–427) because a judgment about the existence of an object must be synthetic and cannot be “reached a priori” (NTKrV 23: 42–43). Second, when one is aware of thinking as a mental activity, this awareness relates to a temporal object because thinking is a process in time (see fn 49). If act awareness represents an individual temporal object, then it must rely on an inner intuition (A19/B33) and cannot be intellectual. Third, intellectual consciousness alone cannot represent a mental act as a mental act, because act awareness as such is too demanding for intellectual consciousness. There are roughly three interpretations of the notion “thinking of a mental act as a mental act”: 1) thinking of a mental act as such, 2) thinking of a mental act as a mental act of certain type (e.g., as a thought about the concept of a triangle), or 3) thinking of a mental act as a particular mental act with some content at some time point (e.g., as a thought about the concept of a triangle at 11: 00 AM Eastern Time). Consider the first, weakest reading. Thinking of a mental act as a mental act requires that the subject can think of it under the concept “mental act.” According to Kant’s theory of concept (Log 9: 58–62, 95), this means that the subject can bring the act under the marks “spontaneous,” “mental,” and “event”. This in turn presupposes that the subject can conceptually distinguish mental acts from nonspontaneous mental events such as sensations, and from nonmental objects such as material things. ⁶⁰ These distinctions are impossible through mere intellectual awareness and must rely on empirical representations. All three readings fall prey to this problem.

The third reading also requires that the subject be able to identify a mental act as a determinate object (i.e., as a particular mental act that has content p at time t) and can distinguish it from other objects. Arguably, if the third reading is followed, the act awareness is even more demanding because representing a mental act as an individual occurrence depends on an identificatory framework: a function which, in this context, only time can serve.⁶¹ The awareness that two instances of a type of thought are different instances of this type presupposes the awareness that these instances are two different events at two different times. Finally, pure apperception is ubiquitous in episodes of (attentional) consciousness (A123; see also A116). If it were truly intellectual act awareness, then Kant would not claim that one is “seldom even conscious” of the synthetical acts themselves (A78/B103, A103–104).

Pure apperception, despite being claimed as the intellectual consciousness of mental acts, is content awareness rather than act awareness.⁶² Three reasons support this view. First, if the “act awareness” reading fails, then pure apperception can only be content awareness. If the subject can still be said to be, in some sense, intellectually conscious of her mental acts without intellectual act awareness, then she must be intellectually conscious of their contents, which are pure concepts or judgments.⁶³ This is because there is simply nothing else in acts of thinking that is cognitively accessible for the subject in an intellectual way. Thus, with the notion of a nonsensible and nontemporal awareness of spontaneous mental acts, Kant means that one is nonsensibly aware of what the mind does by spontaneous acts—that is, that one is nonsensibly aware of their content, i.e., pure concepts or judgments.⁶⁴

Second, the “content awareness” reading offers a better explanation of pure apperception’s role and nature than the “act awareness” reading. Through content awareness, the subject has access to, and hence can manipulate, the content of individual mental acts occurring at different times. Content awareness is consequently presupposed by the unity of consciousness, which is in turn required for intellectual operations such as synthesis, judgment, and inference. In this sense, content awareness fulfills Kant’s description of the transcendental role of pure apperception (A341/B399–400). If pure apperception were intellectual act awareness, then it would be difficult to explain why basic operations of understanding require the highly demanding act awareness rather than content awareness. Additionally, content awareness is necessary for the subject to be aware of representations and to self-ascribe them by means of the I think; hence, content awareness is also presupposed by empirical self-consciousness, and is that which gives rise to individual instances of I think.

Third, the “content awareness” reading better accommodates textual evidence. In terms of content awareness rather than act awareness, pure apperception is described as “a consciousness of what the human being does” (Anth 7: 161, emphasis mine). In addition, the fact that Kant, in passages cited above, refers to pure apperception as one’s consciousness of acts of thinking or synthesis does not contradict but rather reinforces this reading. Let us first inspect the most significant such passage:

[The thought that these representations given in intuition all together belong to me] still presupposes the possibility of [the consciousness of the synthesis of the representations], i.e., only because I can comprehend their manifold in a consciousness do I call them all together my representations.” (B134, emphasis mine)

Remarkably, the second, explanatory part of this passage does not appeal to act awareness. It suggests that “the consciousness of the synthesis” is consciousness of the synthesized content rather than consciousness of the synthetical act. A similar case, occurring just a few lines before at B133, further confirms this reading. Moreover, thinking is nothing but operating concepts and judgments (A68–69/B93–94, A126), and these are the contents of thought. Therefore, content awareness, although not an awareness of an act of thinking as such, is an awareness of the essence of such an act, and can count as consciousness of acts in a broad sense. Finally, Kant occasionally employs terms related to mental items, such as “representation” and “concept,” liberally: They can refer either to a mental state or act of some kind, or to its content.⁶⁵ Therefore, it is reasonable to suppose that a similar ambiguity exists in how the terms “act,” “thinking,” and “synthesis” are used in the notion of “consciousness of (acts of) thinking or synthesis.”

It can be further argued that pure apperception is not a self-standing, distinct representation. The concept of “pure apperception” results from a theoretical abstraction: for a given conscious state, one focuses on that nonsensible, formal aspect which constitutes its transcendental conditions and abstracts from all its sensible components.⁶⁶ The rationale for this relates to the originality of pure apperception, a privileged status that it enjoys against ordinary representations. All one’s (ordinary) representations—i.e., representations that convey a cognitive value or content for the subject—must be such that the I think can accompany them (B131–132). In contrast, pure apperception is “original” and does not require another layer of I think (B132). Clearly, pure apperception is not a self-standing, ordinary representation.

Strong textual evidence supports this reading. Pure apperception constitutes “the formal element” and empirical apperception “the material of consciousness” (Anth 7: 141–142).⁶⁷ This suggests that a conscious inner state cannot consist of pure apperception alone. Pure apperception grounds the categories in terms of judgments such as “I think substance, cause, etc.” (Q2.3). No such judgments can be detected by introspection. The proposition “I think” does not specify the type of spontaneous acts that it self-ascribes, whereas actual instances of self-ascription can occur with expressions such as “I perceive,”⁶⁸ “I infer,” “I assume,” and “I will.” Thus, it is no wonder that the “I think” is claimed as an idealized “general expression” (B138, A398) or “formal proposition” (A354), i.e., as an abstraction from actual instances of self-ascription. Furthermore, Kant occasionally speaks of “the I” rather than the “I think” when discussing pure apperception (A117n., B68, B278; Anth 7: 141). This indicates that he is not bound to the “I think.” Finally, whereas empirical apperception belongs to psychology (Anth 7: 141, 134n.) or physiology (Anth 7: 398), transcendental apperception belongs to logic (Refl 20: 270; Refl 15: 85; Anth 7: 141; A398). The logical viewpoint relies on an abstraction “from every object” (A398). The concept of “pure apperception” results from considering apperception separately from, among other things, the object of empirical apperception. This object is a particular inner state (Anth 7: 141–142). A highly illuminating remark reveals that such an abstraction indeed underlies Q2, a crucial passage for IEI:

[Q6] For inner experience in general and its possibility, or perception in general and its relation to another perception, without any particular distinction or empirical determination being given in it, cannot be regarded as empirical cognition, but must be regarded as cognition of the empirical in general, and belongs to the investigation of the possibility of every experience, which is of course transcendental. (A343/B401, emphasis mine)

Note that Q6 immediately follows Q2, which contains the mysterious claim that pure apperception expresses a “perception of oneself.” Q6 evidently addresses this perception.

In light of my argumentation thus far, the notion of apperception underlying Q2 and Q6 can be outlined as follows: The actual, and, in conscious states, ubiquitous act of self-ascription I think (§4.1) is motivated by a particular perception (IEI) of the cognitive act involved in generating the content that is expressed in the “that” clause following “I think.” The I think refers, via IEI, to this particular cognitive act, and constitutes a kind of act awareness. Thusly, the I think self-ascribes the relevant representational state and is, hence, an empirical representation. To determine the transcendental condition and structure of self-consciousness, the instances of I think and episodes of consciousness into which these instances are built could be considered under abstraction. That is, one abstracts from their sensible details (e.g., their temporal features), the special character of the spontaneous act that the I think represents, and from the properties of the inner perception provoking the I think.⁶⁹ What remains with respect to self-consciousness are pure apperception (intellectual content awareness that gives rise to the instances of the self-ascriptive I think), the self-ascriptive act of I think in general, and IEI in general. Due to their general character, they belong to transcendental investigation. Thus, the apparent contradiction in Kant’s claim that pure apperception expresses an inner perception (Q2) can be ascribed to a confusion caused by this theoretical abstraction. What he actually has in mind is that the instances of I think, from which the notion of pure apperception is abstracted, express an inner perception.

4.3. Indeterminate Empirical Intuition (IEI)

In this section I address the characteristics of IEI and its locus in the structure of self-consciousness, and will demonstrate that IEI is ubiquitous in nonintrospective conscious states. As such, it is obscure, since attention is externally focused. As introspection occurs rarely compared to outer experiences, conscious states (especially in the context of KrV) can be considered as generally containing an obscure IEI that gives rise to an obscure I think. Textual analysis indicates that Q2 and Q3 concern nonintrospective conscious states, and hence concern the IEI as an obscure representation.

Upon first inspection, IEI is always present in conscious states. There are two reasons for this. First, as argued in §4.1, the I think grounded on IEI is ubiquitous in conscious states. Second, as mentioned above, IEI is effectuated by affection through synthetical acts that are constitutive of experience. Since such acts occur constantly in conscious experience, IEI must be ubiquitous in conscious states.

This claim requires further specification and qualification concerning different modes of conscious states. In nonintrospective conscious states, attention is externally directed, i.e., toward objects in the material world. IEI is obscure insofar as attention is externally directed. This is because attention makes the representations inside its focus clear but leaves those that are outside its focus and inside the sensory field obscure.⁷⁰ Therefore, in nonintrospective conscious states, IEI stays outside the focus of attention and hence remains obscure (hence “nonfocal obscure representation”).⁷¹ IEI can become a clear, empirical inner intuition only through introspection, i.e., when one attends to one’s own inner states. Additionally, in introspection, what IEI represents becomes conceptually determined, because understanding determines inner sense in attention (B157n.). Thus, in introspection, IEI becomes DEI, and the obscure I think that accompanies it becomes a clear, conscious one. As introspection occurs rarely compared to conscious outer experiences—after all, steering attention inward requires voluntary effort (V-Met-L1/Pölitz 28: 227)—conscious states can still be said to generally contain an obscure I think that is motivated by an obscure IEI. This is especially the case in the context of KrV, which mainly concerns empirical cognitions of the external world.⁷²

On closer scrutiny, Q2 and Q3 address precisely the nonintrospective scenario. Their immediate contexts (i.e., A343/B401 and B421–423) show that the I think that Kant has in mind there cannot be the clear or conscious I think occurring in introspective consciousness. The I think in Q2 “makes even all transcendental concepts possible” (A343/B401). It occurs in propositions such as “I think substance, cause” (A343/B401, emphasis mine). The I think in Q3 is the representation to which the “unity of consciousness” is anchored (B421); it is a transcendental condition for the unified conscious experience in general. These expressions indicate that the intentional object of the conscious state that contains this type of I think is not the subject.⁷³ Especially considering the context of KrV, which mainly concerns empirical outer cognitions, Kant cannot be addressing introspective consciousness as a default when he speaks of the I think. Thus, the I think in Q2 and Q3 is not the conscious I think that is involved in introspection, an unusual type of consciousness; rather, Kant is referring to the ordinary mode of I think, the obscure self-thought that occurs in all episodes of nonintrospective consciousness. In this mode, the corresponding IEI constitutes an obscure empirical self-consciousness.

A comparison with Q1 can further reinforce this observation. In Q1, Kant deals with “[t]he proposition ‘I think’, insofar as it says so much as that I exist thinking” (B429, emphasis mine). The subordinate clause indicates that what concerns Kant is the empirical I think that represents the determination of the subject’s existence, i.e., its inner state of thinking. Moreover, the text following this quotation says that the subject is “at the same time an object” (Q1.1, or B429). Clearly, this I think occurs in a conscious representational state whose intentional object is the subject and its determination. Therefore, in Q1, Kant has in focus the conscious I think in introspection. When one is introspecting one’s thoughts and contemplating one’s own existence, the self-perception underlying this I think is not obscure. It stands in the focus of attention, and hence is determined with respect to (some) categories (fn 15).

4.4. Solving the Puzzle of IEI

I now demonstrate how the suggested reading explains all obscurities around the target passage Q3, and resolves long-standing controversies regarding IEI.