1. Introduction
There is an emerging consensus among interpreters that in his Critical writings Kant ascribes unconscious representations to human minds.Footnote 1 The nature and extent of this commitment over the course of Kant’s philosophical development is however not well understood.
The classification of a representation as unconscious can mean different things. A representation might be unconscious in a weak sense, such that it escapes notice in the course of conscious attention or simply has been forgotten. It might be unconscious in a strong sense, such that it is unavailable to introspection or cannot serve as a direct ingredient in conscious mental states. Each of these characterizations occurs in the writings of Kant’s immediate predecessors and contemporaries. Assertions of his commitment to unconscious representations, however, rarely acknowledge these different characterizations. We can understand the nature of Kant’s view only if we clarify the different ways he thinks one can possess a representation unconsciously.
The extent of Kant’s commitment to unconscious representations is, in part, a function of the role that commitment plays in his thought over the course of his entire philosophical development. Our understanding of this is however limited. Assertions of Kant’s commitment to unconscious representations typically emphasize remarks in the late Anthropologie (1798) and, to a lesser extent, the first Critique. The study of Kant’s commitment prior to those published writings has focused primarily on remarks in the student lecture notes on anthropology that date from 1772 onwards. Ascriptions of unconscious representations in his earliest writings, however, have received virtually no attention. We can understand the extent of Kant’s commitment to unconscious representations only if we determine the prevalence of that commitment within the entirety of his published writings and Nachlass.Footnote 2
Evidence of Kant’s commitment to unconscious representations occurs in writings and marginalia that span four decades. These remarks raise important questions about Kant’s views on a number of issues, including the role of unconscious representations in perception, moral motivation and the cognitive life of non-human animals. An exhaustive study of the nature, extent and significance of Kant’s commitment to unconscious representations is beyond the scope of this article. What follows contributes to this study by documenting and explaining ascriptions of unconscious representations in his early writings.
Section 2 provides necessary context by discussing ascriptions of unconscious representations in post-Leibnizian philosophy.Footnote 3 I briefly describe the technical vocabulary post-Leibnizians used to classify mental content and then explain how some of these terms were used to denote unconscious and conscious representations, respectively. I focus on the writings of two of Kant’s most important predecessors and contemporaries, Alexander Baumgarten (1714–62) and Georg Friedrich Meier (1718–77). Section 3 draws on this material to explain Kant’s use of this technical idiom in early writings from 1747 to 1763. I argue that in his earliest published writings Kant consistently ascribes unconscious representations to human minds and that some of these representations are unconscious in the strong sense that they are neither available to introspection nor direct ingredients in conscious mental states. Section 4 explains how Kant’s view of unconscious mental content expanded during the mid-1760s. I argue that starting in the mid-1760s Kant began to ascribe to the mind not only unconscious representations but also unconscious acts of judging and inferring. Finally, section 5 briefly considers Kant’s principal justification for positing unconscious representations. Student lecture notes from the 1770s and 1780s ascribe to Kant a consistent rationale for positing representations that are unconscious in the strong sense that we can have no direct awareness of them. As a general methodological point, the paper illustrates the importance of understanding post-Leibnizian philosophy for the purpose of interpreting Kant’s writings.Footnote 4
2. Baumgarten and Meier on Unconscious Representations
Post-Leibnizian philosophers employed a set of technical terms and criteria for classifying representations and mental states. The initial statement of this technical vocabulary occurs in the opening paragraphs of Leibniz’s 1684 essay Meditationes de Cognitione, Veritate et Ideis (G IV, 422–6).Footnote 5 There, Leibniz formulated a taxonomy and criteria for classifying ‘notions’, or concepts, in terms of the traditional Cartesian language of obscurity, clarity and distinctness. According to Leibniz’s taxonomy, every concept is either obscure or clear, and every clear concept either confused or distinct.Footnote 6
The immediate influence of Leibniz’s essay appears to have been negligible. Thirty years following its publication Christian Wolff wrote that it had been ignored by those working on the very issues Leibniz had discussed.Footnote 7 What appears to have been decisive for the influence of Leibniz’s taxonomy was its appropriation by Wolff. In the Deutsche Logik (1713) and subsequent writings, Wolff repeats both the taxonomy and the criteria Leibniz had employed for the classification of concepts. Wolff’s use of Leibniz’s taxonomy was itself then appropriated in numerous logic texts of the period, including Ludwig Philipp Thümmig, Institutiones Philosophiae Wolfianae (1725); Johann Christoph Gottsched, Erste Gründe der gesammten Weltweisheit (1733); Friedrich Christian Baumeister, Philosophia Definitiva (1738); Hermann Samuel Reimarus, Vernunftlehre (1756); and Johann Heinrich Lambert, Neues Organon (1764).Footnote 8
There are important interpretative questions surrounding the origin and development of the technical idiom post-Leibnizians used for classifying mental content. These include questions concerning the scope of Leibniz’s original taxonomy, the criteria he and Wolff employed for individuating obscure concepts, and a difficulty in their accounts of the conditions for possessing a clear concept.Footnote 9 A proper account of these issues is well beyond the scope of this paper. In what follows, I explain how two of Kant’s most important predecessors and contemporaries used the post-Leibnizian idiom to express a distinction between unconscious and conscious representations. This, in turn, will allow us to better understand ascriptions of unconscious representations in Kant’s early writings.
One of the most important developments in the reception of the post-Leibnizian idiom is its modification in the writings of Alexander Baumgarten (1714–62). In the Deutsche Logik (1713), Wolff had defined a ‘thought’ (Gedancke) as something like a conscious mental state, or that by means of which we are conscious. He then defined a ‘concept’ (Begriff) as both a constituent ingredient of a thought and a representation of something (§§2–4).Footnote 10 As a consequence, Wolff’s appropriation of Leibniz’s taxonomy in the Deutsche Logik restricted its application to consciously possessed representations. Baumgarten departs from this by using the idiom to classify both unconscious and conscious representations.
Leibniz had endorsed a monadic ontology that defines the essence of a monad in terms of the capacity for representation and posits ‘bare monads’ that represent in the absence of conscious awareness. Leibniz’s philosophy of mind thus posits both unconscious and conscious representations. Baumgarten follows Leibniz in every one of these respects.Footnote 11 In the Metaphysica (1739), he endorses a monadic, representational ontology and posits ‘slumbering monads’ that represent independently of conscious awareness (§§230–4, 394–5; cf. §§400–2).Footnote 12 He defines the ‘soul’ (anima) as a capacity for both representation and conscious awareness, and ascribes to it both unconscious and conscious representations (§§504, 506, 511).
What is distinct in Baumgarten’s account is that he uses the post-Leibnizian idiom to classify these representations. This is distinct for several reasons. First, Leibniz and Wolff formulated taxonomies for the classification of concepts in particular, rather than representations in general.Footnote 13 Second, while in later writings Wolff extends the application of the German cognates for ‘obscure’ and ‘clear’ to other forms of mental content, such as ‘thoughts’ (Gedanken), it is controversial whether he uses ‘obscure’ to denote unconscious mental content.Footnote 14
In contrast to this, Baumgarten uses the idiom to classify both unconscious and conscious representations. Doing so, however, requires fundamental changes to the most basic terms in the taxonomy set forth in the Deutsche Logik and later logic texts. In particular, it requires changes to the connotations of the Latin and German cognates for ‘obscure’ and ‘clear’ so as to incorporate into the taxonomy a distinction between unconscious and conscious representations. Baumgarten implements these changes in the Metaphysica. He uses ‘obscure perception’ (perceptio obscura) to denote an unconscious representation and ‘clear perception’ (perceptio clara) to denote one that is possessed consciously (§§401, 511–12). Corresponding to these changes in the connotations of obscura and clara is the introduction of a different criterion for distinguishing between ‘obscure’ and ‘clear’ representations. In particular, Baumgarten employs a criterion of phenomenal discrimination as a necessary and sufficient condition for the clarity of a representation: one’s representation is clear if and only if one can distinguish that representation (or its referent) from other objects or properties in one’s phenomenal field; otherwise, the representation is obscure (§§510, 520, 528).Footnote 15 This is different from the criterion Leibniz and Wolff formulated for clear concept possession. According to that criterion, one’s concept is clear if and only if one can not only discriminate but also recognize the referent. The distinction between phenomenal discrimination and object recognition is important for explaining cases wherein one is unable to identify the object one discriminates in conscious perception.
Baumgarten’s transformation of the post-Leibnizian idiom is reiterated and emphasized in the writings of his most prominent student and devotee, Georg Friedrich Meier (1718–77). Corresponding to this is a more explicit commitment to unconscious representations, one motivated by a broadly Leibnizian metaphysics.
Meier ascribes unconscious representations to human minds for two reasons. The first concerns the emergence of conscious mental states in a world of objects that lack conscious awareness. In his first published philosophical work, Meier appeals to a Leibnizian principle of continuity and argues that, since there are no leaps in nature, the conscious states of finite beings must emerge from mental states composed of unconscious representations and to which consciousness gradually accrues (Meier Reference Meier1742: §45). It is for this reason that Meier asserts that unconscious representations are explanatorily prior to conscious representations: ‘Our soul first receives obscure representations before these can be unwrapped and made clear’ (Unsere Seele bekomt erst dunckele Vorstellungen, ehe sie dieselben auswickeln und klar machen kan) (§45).Footnote 16 He first posits mental states that are wholly unconscious and then explains the gradual emergence of consciousness by positing a continuum of mental states that include ever increasing numbers of conscious, constituent representations.
A second reason for Meier’s ascription of unconscious representations is his commitment to a monadic, representational ontology. This was arguably implicit in the first edition of Beweis: daß keine Materie dencken können (1742) and is explicit in the expanded, second edition published in 1751. Appealing explicitly to Leibniz, Meier writes that ‘the bodies in this world are composed of monads, or simple substances’ (Meier Reference Meier1751: §§25, 53). He defines the essence of a monad in terms of its capacity for representation and classifies monads according to the different kinds of representations of which they are capable (§58). The simplest monads are unconscious and possess only ‘obscure representations’ (dunckele Vorstellungen). Other monads possess the capacity for representations in ‘a higher degree’ and are ‘thinking monads’ that represent consciously. A variety of monads that differ in their representational capacities is said to be necessary for the best of all possible worlds. In the absence of monads that have ‘merely obscure representations’ there would be a gap in the order of finite objects and a violation of the principle of continuity (§58).
In classifying the different kinds of representations of which monads are capable, Meier follows Baumgarten’s appropriation of Leibniz’s technical idiom. This is particularly true for his account of the representations of human minds. To understand Meier’s account of mental representation, it is important to recognize that he often uses synonymously the terms ‘representation’ (Vorstellung) and ‘cognition’ (Erkenntnis). In his logic texts from 1752, Meier is explicit about this practice: ‘There is no error to fear if we take the representation of a thing and the cognition of that thing to be one and the same’ (Meier Reference Meier1752b: §25).Footnote 17 Once this is recognized, it becomes clear that many of his remarks about the nature of human cognition generalize to the nature of mental representations. The following is one example.
The absolutely first division [within human cognition] follows so readily from our investigations thus far that it must occur to every careful reader to divide cognition into that of which we are conscious and that of which we are not conscious. The first kind of cognition is called a clear cognition, and the other an obscure cognition. A clear cognition is always one of which we are conscious and is accordingly a thought (Gedanke) … We are not conscious of an obscure cognition, and so it is not a thought. (1752b: §155)Footnote 18
On Meier’s view, the most fundamental distinction among representations and cognitions is that between those that are unconscious and those that are conscious, respectively. These are the paradigmatic connotations of the terms ‘obscure’ and ‘clear’. As with Baumgarten, Meier’s further classification of clear representations as being either confused or distinct is broadly consistent with Leibniz’s original taxonomy and need not concern us here.
Meier’s claim that ‘obscure’ representations and cognitions are those of which we are ‘not conscious’ requires clarification, for there are several distinct connotations operative in his writings. Two of these concern his claim that the obscurity of cognitions admits of degrees. Meier sometimes uses the term ‘cognition’ to denote a complex aggregation of representations, or constituent marks (Merkmale) (Meier Reference Meier1752b: §158).Footnote 19 A cognition that consists of a mixture of obscure and clear representations is said to be only ‘partially obscure’ (einesteils dunkel), whereas one that consists exclusively of wholly obscure representations is ‘completely obscure’ (ganz dunkel) (1752b: §156; 1752a: §125, AA 16: 318). There is thus a quantitative respect in which representations admit of varying degrees of ‘obscurity’.Footnote 20
In addition to this, degrees of obscurity can vary in a qualitative respect. Some representations are only ‘relatively obscure’ (beziehungsweise dunkel), where this means it is possible for one to become conscious of them (Meier Reference Meier1742: §26; 1752a: §125, AA 16: 316). These representations are unconscious in the weak sense that they merely have been forgotten, are not objects of conscious attention, or simply evade one’s current attempt at recollection. Some of these connotations are operative in Meier’s use of the term ‘obscure’. He writes, ‘If we cannot recall a word, and yet it appears to be on the tip of the tongue (vor dem Munde herumliefe); then we have an obscure representation of it’ (1752a: §124, AA 16: 316). Similarly, Meier appeals to forgetfulness to explain how clear, complex representations can become obscure: ‘If we forget the marks (Merkmale) of a clear cognition, then a clearer cognition can become less clear, until, finally, it is transformed into a completely obscure cognition (ganz dunkele Erkenntnis)’ (1752a: §136, AA 16: 338).
In addition to these ‘relatively obscure’ cognitions and representations, Meier writes of those that are unconscious in the strong sense that they have never been and cannot become ingredients in conscious mental states. An ‘absolutely obscure’ (schlechterdings dunkel) cognition is one example of this. A complex representation all of whose constituent marks are originally obscure is a representation that is ‘so obscure that it is absolutely impossible to make it clear’ (Meier Reference Meier1752a: §125, AA 16: 316; 1752b: §156). The impossibility of making a cognition clear is equivalent to the impossibility of one becoming conscious of it: ‘we confess that we cannot ourselves experience obscure cognition’ (1752b: §159).Footnote 21
Positing representations that are unconscious in the strong sense naturally invites questions about the reasons for positing representations of which we have no direct knowledge. It is a distinguishing feature of Meier’s account that he recognizes the importance of justifying belief in the existence of unobservable representations: ‘how can we be certain that we really possess obscure cognition?’ It is a further distinguishing feature that the reasons he offers are independent of the monadic, representational ontology which motivates his positing such representations. In Deutsche Metaphysik (1720), Wolff suggests very briefly the possibility of inferring the existence of unconscious mental content (§193; cited in Madrid Reference Madrid2012: 196). In contrast to this, Meier offers several arguments to justify belief in the existence of absolutely unconscious representations.
Meier first criticizes the inference from the unobservable character of a phenomenon to its non-existence and then claims we are justified in positing unconscious representations on the basis of their observable effects:
Yes, we concede that we cannot ourselves experience obscure cognition; we can only experience their effects. This confirms completely for us that [obscure cognitions], as their causes, must really be present in the soul. (Ja wir geben zu, daß wir die dunkele Erkenntnis selbst nicht erfahren können, allein wir können ihre Würkungen erfahren, welche uns aufs vollkommenste versichern, daß sie selbst, als die Ursache derselben, würklich in der Seele vorhanden sein müsse) (Meier Reference Meier1752b: §159)
Meier offers two arguments for the claim that we can infer the existence of unconscious representations from their effects. The first proceeds on the moral psychological assumption that every affective state presupposes a representation of something one considers good or bad. Since we sometimes experience emotions of elation or depression without knowing why, Meier infers the representations underlying these emotions must be unconscious. The second reason presupposes the Leibnizian view that our concepts of sensible qualities are complex and composed of constituent representations. Since we always experience these qualities as being phenomenally simple, the representations constitutive of their underlying complexity must be unconscious (Meier Reference Meier1752b: §159).
Let me summarize briefly. Leibniz’s 1684 essay and Wolff’s Deutsche Logik (1713) were important for the formation of the technical vocabulary post-Leibnizians used to classify concepts. In the original version of this vocabulary, the term ‘obscure’ (dunkel, obscura) denoted a concept that is sufficient for conscious discrimination of the referent but not for its recognition. Corresponding to this, the term ‘clear’ (klar, clara) denoted a concept that is sufficient for both conscious discrimination and recognition of the referent.
Both these terms acquire new connotations in the writings of Baumgarten and Meier. In particular, ‘obscure’ is now used to denote representations that are unconscious to varying degrees, and ‘clear’ those that suffice for conscious discrimination of the referent. Meier, in particular, contributed to this usage by offering a more detailed account of unconscious representations and offering reasons for positing the existence of representations that are either inaccessible to introspection or incapable of serving as direct ingredients in conscious mental states.
The early Kant was well acquainted with the writings of Baumgarten and Meier. He used Meier’s Vernunftlehre (1752b) as the official text for his lectures on logic during his first semester as a docent, and Meier’s Auszug aus der Vernunftlehre (1752a) every semester thereafter. In his lectures on metaphysics and ethics, Kant regularly used Baumgarten’s Metaphysica (1739), Initia philosophiae practica (1740) and Ethica philosophica (1751).Footnote 22 He lectured on the writings of Baumgarten and Meier nearly every semester of an academic career that spanned four decades. In early and later writings, Kant uses the Latin and German cognates for ‘obscure’ and ‘clear’ with the distinct connotations Baumgarten and Meier had assigned those terms. This includes Meier’s distinctions between ‘relatively obscure’ (beziehungsweise dunkel) and ‘absolutely obscure’ (schlectherdings dunkel) representations, and between ‘partially obscure’ (einesteils dunkel) and ‘completely obscure’ (ganz dunkel) representations. This is one important reason for this article’s focus on Baumgarten and Meier.
In what follows, I argue that in his earliest writings Kant uses this vocabulary to ascribe unconscious representations to the mind. It is important to recognize that the writings on which I focus were published between 1747 and 1763. This is significant because, while the publication of Leibniz’s Nouveaux Essais was important for Kant’s philosophical development, that work did not appear until 1765 in the edition published by Raspe.Footnote 23 As a consequence, the account of petite perceptions set forth in that work is less useful for interpreting Kant’s earliest view of unconscious representations than might normally be expected. A further consequence is that the published writings of Baumgarten and Meier acquire greater interpretative significance because of this.
3. Unconscious Representations in Kant’s Early Writings
Two related commitments define Kant’s early philosophy of mind. Each of these was common among his post-Leibnizian predecessors and contemporaries. The first is a representationalist view of mental content. The second is the partitioning of the mind into higher and lower cognitive faculties. Kant characterizes each of these commitments in ways that imply a commitment to unconscious representations.
Representationalism about mental content is the view that the content of the mind consists paradigmatically, if not exclusively, of intentional representations. A representation is intentional if and only if it is a representation of something. An early expression of Kant’s commitment to representationalism occurs in Gedanken von der wahren Schätzung der lebendigen Kräfte (1747): ‘Now, the entire internal state of the soul is nothing other than the aggregation of all of its representations and concepts, and, insofar as this internal state is related to something external, it is called the status repraesentativus universi’ (1: 21).Footnote 24 Kant is here endorsing a claim central to the post-Leibnizian literature. In the Metaphysica (1739), Baumgarten had defined the soul as a power for representing the universe (vis repraesentativa universi) and characterized its content in terms of the totality of its representations (perceptio totalis) (§§513–14). In the minds of rational human agents, this totality consists of a variety of obscure, confused and distinct representations. Such variety is implicit in Kant’s reference to the ‘representations and concepts’ of the soul’s internal state. In the post-Leibnizian literature, the Latin and German terms for ‘representation’ (repraesentatio, Vorstellung) were typically used to denote representations in general, and those for ‘concept’ (notio, conceptus, Begriff) to denote representations that are possessed consciously.Footnote 25 Interpreted with reference to this terminological distinction, Kant’s claim is that the content of the mind consists exclusively of representations, some but not all of which are possessed consciously.
Kant’s commitment to this form of post-Leibnizian representationalism is reiterated fifteen years later and in similar terms. In Begriff der Negativen Grössen (1763), he writes: ‘There is something great and, I should think, very true in this thought of Herr Leibniz: The soul occupies itself with the entire universe by means of its power for representation, even though only an infinitely small number of these representations are clear’ (2: 199).Footnote 26 The reference to ‘clear’ representations implies a distinction between ‘clear’ and ‘obscure’ representations, the latter of which are said to constitute the preponderance of the representations in the soul. Kant makes a similar claim in a Reflexion which Adickes dates to this period: ‘The field of obscure perceptions is the most expansive’ (Obscurarum perceptionum campus est amplissimus) (R176, 15: 64). The claim that the preponderance of the mind’s representations are unconscious is one that was emphasized by Kant’s predecessors. In the Metaphysica (1739), Baumgarten had written that the totality of obscure representations constitutes ‘the foundation of the soul’ (fundus animae) (§511). Meier is similarly committed to this view and asserts in his logic texts that obscure representations are explanatorily prior to those that are clear. Kant emphasizes this latter claim repeatedly in the margins of his copies of Baumgarten’s Metaphysica and Meier’s Auszug.Footnote 27
In these early passages, Kant asserts the content of the mind consists exclusively of representations and implies that most of these are ‘obscure’ (dunkel). In the writings of Meier, we saw that the term dunkel is used to denote mental content that is unconscious to varying degrees, and that the term admits of several distinct but related connotations. Understanding the role of ‘obscure’ representations in Kant’s early writings requires clarifying the intended connotation of that term. We can begin to do this by considering the second, aforementioned commitment definitive of Kant’s early philosophy of mind.
Among Kant’s predecessors and contemporaries, it was common practice to partition the mind into lower and higher cognitive faculties. The ascription of unconscious representations to the lower faculty was common within post-Leibnizian philosophy of mind. In the writings of Baumgarten and Meier, the lower faculty was the site of both ‘obscure’ and ‘clear but confused’ representations, and the higher faculty those that are ‘distinct’. As a consequence, they ascribed both unconscious and conscious representations to the lower faculty.
Kant follows Baumgarten and Meier but also departs from them in a fundamental respect. Like his predecessors, he partitions the mind into lower and higher faculties, and ascribes unconscious representations to the lower faculty. He departs from his predecessors, however, in ascribing conscious representations exclusively to the higher faculty.
In the Nova dilucidatio (1755), Kant discusses the role of unconscious representation in human agency:
But the case can arise where the grounds that incline the will in one of two directions completely escape consciousness, but where, nonetheless, one alternative is chosen in preference to the other. In that case, the thing is reduced from a higher faculty of the mind to a lower one, and the mind is directed in one direction or another by the preponderance of an obscure representation (repraesentationis obscurae). (1: 401)
He repeats this claim several pages later: ‘For all that is brought about by such irresolute uncertainty is that the thing [i.e. the determining ground] is reduced from the higher faculty to the lower one, from a representation that is conjoined with consciousness to representations that are obscure’ (1: 406).
On Kant’s view, every human action presupposes an antecedently determining ground (ratio antecedenter determinans). To say that a ground, or reason (ratio), is antecedently determining is to say it is the reason why a subject is determined with regard to one predicate rather than another (1: 391–2).Footnote 28 In some cases, an action is chosen on the basis of conscious deliberation, in which case the agent is conscious of the determining ground. In other cases, however, an agent chooses an action amid pronounced uncertainty and in complete ignorance of the grounds that incline her will in competing directions. In these latter cases, Kant writes, the determining ground that inclines the will in one direction rather than another is an ‘obscure representation’ in the lower faculty and one that ‘completely escapes consciousness’ (conscientiam plane fugiant).
In the passages above, the connotation of repraesentatio obscura is that of a representation that is unconscious in a strong sense, and that in two respects. In the first of these, the representation is unconscious in the sense that it is not available to introspection. This is evident from the nature of the phenomenon with which Kant is concerned, namely, one’s inability to know the determining ground of one’s action amid an experience of ‘irresolute uncertainty’ (ancipiti tali dubitatione). It is however conceivable that a representation might be possessed consciously and yet not be available as an object for introspection. Such a representation might perhaps figure in tacit awareness but elude one’s capacity for conscious attention. For this reason, it is important to recognize the second, stronger respect in which the determining ground is said to be unconscious. This is implicit in the contrast between ‘a representation that is conjoined with consciousness’ and ‘representations that are obscure’. If a representation is not ‘conjoined with consciousness’, then it is unconscious in the strong sense that it does not serve as a direct ingredient in conscious mental states, including those constitutive of tacit awareness. Since the representation in question is the determining ground, the representation plays that causally efficacious role simultaneously with its not being a direct ingredient of conscious awareness. This is a significantly stronger sense of what it means for a representation to be unconscious. This would explain why Kant thinks the representation is not available to introspection.Footnote 29
In the passages quoted above, Kant identifies consciousness with the higher faculty and implies that representations in the lower faculty are exclusively unconscious. This is implicit in both passages but particularly evident in the second. The determining ground is ‘reduced from a higher faculty to a lower one, from a representation conjoined with consciousness to representations that are obscure’ (a superiori facultate ad inferiorem, a repräsentatione cum conscientia coniuncta ad obscuras redeat). Kant’s explanation of what it means for a determining ground to be transferred from the higher to the lower faculty makes sense only if he ascribes all and only obscure representations to the lower faculty. If he ascribed ‘clear but confused’ representations to the lower faculty, then reduction of the determining ground to the lower faculty would not necessarily have the implication he describes. The implication is that, whereas a determining ground in the higher faculty is ‘a representation conjoined with consciousness’, a determining ground in the lower faculty is ‘obscure’. Kant’s remark implicitly excludes the possibility that a determining ground in the lower faculty might be a clear but confused representation. The reason why is that he identifies first-order consciousness exclusively with the higher faculty and ascribes all and only obscure representations to the lower faculty.Footnote 30
Kant’s identification of consciousness with the higher faculty is reiterated seven years later in Die falsche Spitzfindigkeit der vier syllogistischen Figuren (1762). In the philosophically substantial concluding remark, he advances three distinct claims. First, possessing a distinct concept is explanatorily dependent on an act of judgement. Second, the capacities for judging and inferring are one and the same. Third and finally, the higher cognitive faculty rests exclusively on the capacity for judgement (2: 58–9).Footnote 31
In his explanation of the third claim, Kant is explicit that the capacity for judgement is a necessary and sufficient condition for possessing the higher faculty. What has not been recognized is that he identifies the capacity for judgement with the capacity for conscious awareness. This becomes apparent once we explicate his early concept of ‘inner sense’.
In the course of elaborating on the nature of the capacity on which the higher faculty rests, Kant writes:
If one is able to understand what kind of secret power it is through which judgement becomes possible, one will have untied the knots. It is my present view that this power or ability is nothing other than the capacity for inner sense, that is, the ability to make one’s representations objects of one’s thoughts … The entire higher cognitive faculty rests on this capacity. (2: 60)
The notion of ‘inner sense’ is one that Kant characterizes in different ways over the course of his philosophical development. In writings from 1770 onwards, his use of the term often implicates a capacity for self-consciousness, as it does for Locke (Longuenesse Reference Longuenesse1998: 233ff.; Kitcher Reference Kitcher2011: ch. 2). The passage above however predates those remarks and uses ‘inner sense’ with a different connotation. Among Kant’s predecessors outside the post-Leibnizian tradition, Christian Thomasius, Andreas Rüdiger and Christian August Crusius all defined ‘inner sense’ as a capacity for first-order conscious awareness (Thiel Reference Thiel2011: 345–9). Within the post-Leibnizian tradition, Baumgarten repeats this definition in the Metaphysica: ‘Sense represents either a state of my soul – INNER SENSE – or a state of my body – OUTER SENSE. Thus, SENSATION is either INTERNAL through inner sense (consciousness in the strict sense), or EXTERNAL, which becomes actual through outer sense’ (§535). The definition of inner sense in terms of consciousness is repeated in Kant’s marginalia. Amid some Reflexionen which Adickes dates to the years 1752–6, Kant discusses the nature of consciousness and explicitly identifies it with inner sense: ‘Das Bewust seyn ist sensus internus’ (R1680, 16: 80).Footnote 32 There and in another Reflexion from the same period, Kant denies consciousness to animals that lack inner sense (R4230, 17: 469). In the student lecture notes contemporaneous with the 1762 essay on logic, Johann Gottfried Herder ascribes to Kant the identification of inner sense with consciousness: ‘Consciousness – inner sense – is thus the distinguishing feature of a rational being’ (Das Bewusstsein – der innere Sinn – ist also der character distinctus der Vernunftwesen) (Metaphysik Herder, 28: 117).
Kant’s claim that inner sense is ‘the ability to make one’s representations objects of one’s thoughts’ is thus best understood as the claim that inner sense is simply the capacity for possessing representations consciously. Since he thinks this capacity is necessary and sufficient for the capacity for judgement, he can claim without inconsistency both that ‘the entire higher cognitive faculty rests on this capacity [for inner sense]’ and that ‘the higher cognitive faculty rests absolutely on the capacity for judgement’ (2: 59). These remarks, in turn, reiterate Kant’s earlier identification of consciousness with the higher faculty in the Nova dilucidatio.Footnote 33
If consciousness is identical with the higher faculty, then the representations in the lower faculty are exclusively unconscious. Earlier, we saw that in the Nova dilucidatio Kant uses repraesentatio obscura to denote representations that are unconscious in the strong sense that they are inaccessible to introspection and employed independently of their being ‘conjoined with consciousness’. A remark written shortly after the 1762 essay on logic suggests that he continues to hold this view. Though it was not published until 1764, the Preisschrift (Untersuchung über die Deutlichkeit der Grundsätze der natürlichen Theologie und der Moral) was completed in December of 1762 (Henrich Reference Henrich1967: 11). There, Kant writes, ‘Obscure representations (dunkele Vorstellungen) are those of which one is not conscious’ (2: 290). The intended connotation of the definiens is clarified to some extent by the context of the remark. Kant criticizes ambiguous uses of technical terms in philosophical analysis. He then illustrates the point by describing a fallacious argument some philosophers use to justify positing unconscious representations:
Most philosophers give as examples of obscure concepts those that we have in deep sleep. Obscure representations are those of which one is not conscious. Now, some experiences show that we also possess representations in deep sleep, and since we are not conscious of these, it follows that they were obscure. Here, consciousness has two meanings. Either one is not conscious that one has a representation, or one is not conscious that one has had a representation. The first denotes the obscurity of the representation as it is in the soul; the second shows nothing other than that one does not remember it. Now, the example provided merely shows that there could be representations of which one is not conscious when one is awake; from this however it does not at all follow that the representations which occur during sleep were not clearly conscious. (2: 290)
Kant’s discussion of the example is compressed and involves some confusion. Properly speaking, the ambiguity that is exploited by the philosophers in question concerns not ‘two meanings’ of ‘consciousness’ but rather two different connotations of what it means to not be conscious of something. What is important, however, is the implication of Kant’s criticism. He defines ‘obscure representation’ as an unconscious representation and then criticizes those who justify this claim by appealing to an insufficiently strong notion of what it means for a representation to be unconscious. He claims the obscurity of a representation ‘as it is in the soul’ is unconscious in a sense stronger than that of a representation that one merely ‘does not remember’.Footnote 34 Earlier, we saw that Meier had appealed to the latter as an example of a representation that is only ‘relatively obscure’ (beziehungsweise dunkel), and that he contrasted this with representations that are obscure in an ‘absolute’ sense (schlechterdings dunkel). The passage above suggests Kant thinks the paradigmatic connotation of ‘obscure representation’ is that of one that is unconscious in a sense stronger than that of being merely ‘relatively obscure’.
The criticism advanced in the Preisschrift is repeated four years later in Träume eines Geistersehers (1766) (2: 338n.). There Kant continues to ascribe unconscious representations to the mind but denies the representations that occur in deep sleep are an example of these. What is noteworthy about this later text, however, is the more expansive account of unconscious mental content it implies.
4. Unconscious Acts of Judging and Inferring
In the 1762 essay on logic, Kant had explicitly identified the capacity for judgement and reasoning with the capacity for conscious awareness. Remarks in Träume eines Geistersehers (1766) imply both a significant change in view and a more expansive commitment to unconscious mental content. In the course of discussing dubious metaphysical concepts, Kant offers the following explanation of their origin:
In the course of experience, many concepts originate in secret and obscure inferences and subsequently propagate themselves among other concepts without consciousness of either the experience or the inference on the basis of which the concept was constructed. One can call such concepts surreptitious. There are many of these, some of which are nothing but delusions of the imagination, whereas others are true, insofar as obscure inferences are not always erroneous. (2: 320n.)
Kant asserts the mind performs ‘secret and obscure inferences’ (geheime und dunkele Schlüsse) of which it is not aware. The claim that a surreptitious concept is acquired ‘without consciousness of either the experience or the inference’ on which that concept is based implies that the inferences in question are unconscious in a sense stronger than that of representations one merely does not remember. It is an open question whether an inference that is unconscious in this sense might later be accessible to conscious awareness through recollection. The point at present is that Kant ascribes to the mind not only unconscious representations but also unconscious acts of judging and inferring. This, in turn, entails a departure from his earlier identification of the capacity for judgement and reasoning with the capacity for consciousness.Footnote 35 What explains this change in Kant’s view?
A comprehensive answer to this question is beyond the scope of this essay. I will however briefly suggest one possible explanation. Several scholars have noted that during the mid-to-late 1760s Kant began to classify judgements as being either ‘analytic’ or ‘synthetic’. When explaining this distinction in the first Critique and later writings, he emphasizes that analytic judgements make explicit relations among concepts that are thought only implicitly. It is worth noting that Kant characterizes implicit and explicit judgements in phenomenal terms. In an analytic judgement, he writes, ‘the predicate B belongs to subject A as something which is contained in the concept A in a concealed manner (versteckter Weise)’ (A6/B10; emphasis added). In the judgement ‘All bodies are extended’, I arrive at the concept of extension by analysing the concept of body, where this means ‘I only become conscious of the manifold which I always think [in the concept of body] to find the predicate therein’ (A7/B11; emphasis added). In the late Preisschrift (on Progress in Metaphysics) from 1791, Kant offers the following definition: ‘Judgements are analytic, namely, if their predicate represents clearly (explicitly) only that which is thought in the concept of the subject, albeit obscurely (implicitly), e.g. “Every body is extended”’ (Urtheile sind nämlich analytisch, wenn ihr Prädicat nur dasjenige klar (explicite) vorstellt, was in dem Begriffe des Subjects, obzwar dunkel (implicite), gedacht war; z.B. ein jeder Körper ist ausgedehnt) (AA 20: 322). Kant here characterizes implicit and explicit judgements using terms he employs elsewhere to distinguish between unconscious and conscious representations. Further work is required to determine whether he is, in fact, using those terms with the technical connotations he appropriated from Baumgarten and Meier. If he is, however, this would explain why he stopped identifying the capacity for consciousness with the capacity for judgement during the mid-1760s. The suggestion is that reflecting on the explicative function of analytic judgements led Kant to reflect more extensively on the nature of implicit judgements, and that he developed an account of the latter within his existing commitment to unconscious representations.Footnote 36
Kant’s claim that ‘there are many’ surreptitious concepts suggests he thinks unconscious inferences occur with some frequency. The passage quoted above from the 1766 essay is indeterminate on this point. Remarks elsewhere however suggest that this is in fact Kant’s view. In a lengthy Reflexion that Adickes dates to the late 1760s, Kant writes, ‘Every act of the understanding and reason can occur in obscurity’ (Alle actus des Verstandes und Vernunft könen in der Dunkelheit geschehen); and, more strikingly, ‘Most of what the understanding undertakes occurs in obscurity’ (Das Meiste geschieht vom Verstande in der Dunkelheit) (R177, 15: 65).
Importantly, Kant wrote these remarks in the margins of his copy of Baumgarten’s Metaphysica and, more specifically, in the chapter devoted to empirical psychology. This chapter was the official text for his lectures on anthropology beginning in 1772. It is thus not surprising that some of the earliest student lecture notes attribute to Kant claims very similar to those in the Reflexion just cited:
The obscure acts of our entire cognitive power constitute the greatest part of the state of the soul. The fewest cognitions are illuminated through consciousness. … The understanding is most efficacious among obscure representations, and all clear representations are for the most part the result of long and obscure acts of reflection. (Reflexionen) (Anthropologie Collins [1772–3], 25: 20)
Preliminary judgements belong to the obscure representations. Before a person effects a determinate judgement, he has already effected in advance a preliminary judgement in obscurity. (Anthropologie Friedländer [1775–6], 25: 481)
As the first sentences of these remarks suggest, the judgements and inferences in question are ‘obscure’ in the sense that they occur independently of conscious awareness. There is reason to think this connotation is implicit in the final remark as well.
5. Kant’s Justification for Positing Unconscious Representations
Anthropologie Friedländer is noteworthy in that it is the earliest set of student lecture notes that contains a distinct chapter devoted exclusively to the existence and nature of ‘obscure’ representations (25: 479–82). Over the course of three pages, the term ‘obscure’ is used in ways that imply several distinct connotations. These include a characterization of the contents of memory as being obscure, as well as the claim that some obscure representations can become clear (25: 480–1). The definition offered at the outset of the chapter, however, is that of a representation that is unconscious in the strong sense that it is inaccessible to introspection:
Obscure representations contain the secret spring of that which occurs in the clear (im Lichten) and so must receive consideration. Obscure representations are those of which one is not conscious. How then can one observe them? Not immediately; however, I can infer that there are representations in me of which I am not conscious. (Anthropologie Friedländer [1775–6], 25: 479)
This remark is both corroborated by a Reflexion that Adickes dates to this period and repeated in subsequent sets of student lecture notes throughout the 1780s. A somewhat fragmentary remark in the Collegenentwürfe (1773–7?) characterizes ‘obscure representations’ as those that one has and of which ‘one is in no way conscious’ (man sich seiner gar nicht bewust ist) (15: 665). In subsequent student lecture notes, one finds variations of the paragraph quoted above from Anthropologie Friedländer.
There have been disputes in philosophy as to whether there are obscure representations of which we are in no way conscious. Various philosophers say obscure representations are of a kind such that we do not know that we have them. But how can we claim that we know something of representations of which we are not conscious? Locke says this, but he is wrong; for we certainly do know these. As long as they are obscure, we are not conscious of them; they lie in immediate sensation, but through inferences we can certainly come to know (hervorbringen) that they are there. (Menschenkunde [1781–2?], 25: 867)
I am not conscious of my obscure representations; how then do I know that I have them [?] I do not know them immediately; I can however infer that I have them on the basis of their effects. (Anthropologie Mrongovius [1784–5], 25: 1221)
Obscure representations are those of which we are not immediately conscious; rather, through their effects … Many dispute the existence of obscure representations and say: How can one be convinced of the existence of obscure representations if we are not conscious of them? To this one can respond: it is simply not necessary that one is conscious of it through sensation if one can come to consciousness of it through inferences. (Es ist eben nicht nöthig daβ man durch Empfindung sich bewuβt ist, wenn man durch Schlüsse zum Bewuβtseyn kommen kann.) (Anthropologie Busolt [1788–9], 25: 1439–40)
These passages ascribe to Kant an enduring commitment to representations that are unconscious in the strong sense that they are inaccessible through introspection. The fact that they are inaccessible through introspection is why knowledge of their existence is problematic.
Of equal significance is the justification Kant offers for positing unconscious representations. Earlier, we saw that Meier recognized the need to justify ascriptions of unconscious representations of which we have no direct awareness. His principal justification is that we can infer the existence of unconscious representations from their conscious effects (1752b: §159). It is a striking feature of the passages quoted above that they consistently ascribe to Kant the same justification for positing unconscious representations.
Properly speaking, the student lecture notes are not part of Kant’s published and unpublished writings, nor are they comparable to anything like verbatim reports from those who attended Kant’s lectures. Their significance as textual evidence is best restricted to cases where they corroborate claims discernible in his writings. For this reason, it is important that Kant repeats many of the claims quoted above from the student lecture notes in the late Anthropologie in pragmatischer Hinsicht (1798). In the opening paragraph of the chapter entitled ‘On the representations that we have without being conscious of them’, Kant writes:
To have representations and yet to not be conscious of them – therein appears to lie a contradiction; for how can we know that we have them when we are not conscious of them? Locke already made this objection and for that reason denied the existence of representations of this kind. However, we can certainly be mediately conscious (mittelbar bewuβt) that we have a representation even if we are simultaneously not immediately conscious of it. Such representations are called obscure; the remaining ones are clear and, when their clarity also extends to the constituent representations of the whole and their relation to it, they are distinct representations, either of thinking or of intuition. (7: 135)
One knowledgeable of the post-Leibnizian idiom might readily notice the traditional taxonomic distinctions implicit in Kant’s use of the terms ‘obscure’, ‘clear’ and ‘distinct’. One familiar with the connotations that dunkel and klar possessed in the writings of Baumgarten, Meier and the early Kant would have good reason to interpret Kant’s late use of those terms as denoting unconscious and conscious representations, respectively. One acquainted with the post-Leibnizian text Kant first used in his lectures on logic might notice he is here repeating Meier’s principal justification for positing the existence of representations unconscious in the strong sense that they are not accessible through introspection. This is the implicit claim that we can be ‘mediately conscious’ of unconscious representations by inferring their existence from observable effects.Footnote 37
6. Conclusion
Interpreted with reference to the writings of Baumgarten and Meier, Kant’s published writings and Nachlass exhibit a sustained commitment to the existence of unconscious mental content. In his earliest published work, Kant ascribes to the mind representations that are unconscious in the strong sense that they are inaccessible to introspection. In writings and marginalia from the mid-1760s and later, he extends his commitment to unconscious mental content by positing unconscious acts of judging and reasoning. Paired with remarks in the student lecture notes and late Anthropologie (1798), this material suggests a substantial commitment to unconscious representations is an enduring feature of Kant’s philosophy of mind for more than four decades.
Kant’s extensive commitment to unconscious representations raises important questions about their place in his accounts of consciousness, empirical cognition and moral motivation. The same is true of his commitment to unconscious acts of judgement and inference, and their relation to his account of the explicative function of analytic judgements.
Readers sceptical of Kant’s enduring commitment to unconscious representations sometimes appeal to either of two passages in the first Critique. The first is the famous claim in the B-Deduction that the ‘I think’ must be able to accompany all of one’s representations (B131–2). Other interpreters have offered readings of that passage that are consistent with Kant’s commitment to unconscious representations, and so I direct interested readers to that literature.Footnote 38 A second passage to which the sceptics sometimes appeal has received only limited commentary. In the second edition of the first Critique, Kant writes, ‘consciousness at every moment has a degree that can always be diminished’ (das Bewuβtsein hat jederzeit einen Grad, der immer noch vermindert werden kann) (B414). He then elaborates on this claim in a lengthy footnote that includes the following remark: ‘Clarity (Klarheit) is not, as the logicians say, consciousness of a representation; for a certain degree of consciousness, which however does not suffice for memory, must itself be found in some obscure representations (dunkelen Vorstellungen)’ (B414n.).
Elsewhere, I offer a detailed reading of these passages. Here, I restrict myself to two remarks. First, Kant writes that ‘a certain degree of consciousness … must itself be found in some obscure representations (manchen dunkelen Vorstellungen)’ (emphasis added). The qualification is subtle but important. Kant does not claim a degree of consciousness is present in every ‘obscure’ representation; he only claims it is present in ‘some’. This is wholly consistent with his positing representations that are unconscious in the strong sense that we have no direct awareness of them.
The second point concerns how exactly Kant conceives the restricted class of obscure representations that admit of ‘a certain degree of consciousness’. Earlier, we saw Meier had claimed the obscurity of cognitions admits of degrees. He asserted a cognition is ‘partially obscure’ (einesteils dunkel) if it contains a mixture of obscure and clear representations, or constituent marks (Merkmale). He then posits a continuum of obscure cognitions that differ in terms of the number of clear marks each contains (1752b: §156; 1752a: §125, AA 16: 318).
Kant’s remark is best interpreted with reference to Meier’s account of ‘partially obscure cognitions’. A complex, obscure representation admits of ‘a certain degree of consciousness’ if one or more of its constituent representations is clear. If Kant also holds there is a continuum of ‘partially obscure representations’ that differ in the number of clear marks each contains, this would explain why he thinks consciousness admits of degrees. It would also suggest Baumgarten and Meier were more influential for Kant’s later thinking about the nature of consciousness than has been previously recognized. I pursue this topic elsewhere.
