Abstract
Even though Kant’s theory of cognition (Erkenntnis) is central to his Critique of Pure Reason, it has rarely been asked what exactly Kant means by the term “cognition”. Against the widespread assumption that cognition (in the most relevant sense of that term) can be identified with knowledge or if not, that knowledge is at least a species of cognition, we argue that the concepts of cognition and knowledge in Kant are not only distinct, but even disjunct. To show this, we first (I) investigate Kant’s explicit characterizations of the nature of cognition. As it turns out, he introduces several different notions that must be carefully distinguished before identifying the one that is central to his project in the first Critique. We then (II) consider the basic features of Kant’s conception of knowledge, indicating both how it involves assent and objective justification and how it relates to our contemporary conception. Next (III), we compare and contrast Kant’s understanding of cognition and his conception of knowledge in a way that allows us to present their fundamental differences and connections. We argue that while cognition, in the most relevant sense, is a species of representation that differs from other representations in that it involves the conceptual determination of a sensibly given object, knowledge (for Kant) is a kind of assent to a judgment that requires consciousness of a sufficient epistemic ground. Finally (IV), by appreciating the differences between cognition and knowledge, we explain several of the implications this conception of cognition has for some of Kant’s main claims in the Critique of Pure Reason as a whole. Among other things, we show how Kant can deny cognition of specific things in themselves while allowing philosophical knowledge about things in themselves in general (e.g. that they exist, are not in space and time, etc.).
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
In Kemp Smith’s defense, “Erkenntnis” is Kant’s translation of “cognitio”, which, in medieval philosophical texts, is commonly translated as “knowledge”.
Henrich (1973).
Cf. Guyer’s and Wood’s translation of the Critique of Pure Reason in the Cambridge Edition of Kant’s writings, who mostly translate ‘Erkenntnis’ as ‘cognition’ (but cf. e.g. B25 and A237/B297, where it is translated as ‘knoweldge’).
For a clear and representative statement of this reading, cf. Grüne (2009, p. 27f).
In the following, we restrict ourselves to theoretical cognition. Kant does introduce (and extensively use) the term “practical cognition”, but his views on the practical case are complex and would require separate treatment. (See Kain (2010) for helpful discussion of practical cognition.) Further, though Kant distinguishes several different types of theoretical cognition—a priori, historical, mathematical, self-cognition—we confine ourselves to the more generic notion of cognition in general.
If analytic judgments qualify as knowledge, they require neither the existence of an object nor the attribution of features to it.
Note that although “Erkenntnis” is a common German word, especially when used as a verb in “erkennen”, it is a technical term in eighteenth-century philosophy insofar as it was used as the German equivalent of “cognitio”, which had been a commonly used term in philosophy since the middle ages and throughout early modern philosophy. Although Kant wrote all his major works in German, most of his terminology is deeply rooted in the latin works of Leibniz-Wolffian tradition. On the use of “cognitio” in early modern philosophy and Kant’s immediate predecessors, cf. Carriero (2013), and Tolley (unpublished manuscript).
The Jäsche Logik has a somewhat uncertain status in Kant’s corpus, because Jäsche was heavily involved in the production of the book in such a way that one cannot be sure that it truly reflects Kant’s position in every respect. However, a very similar list of six degrees of cognition can be found both in Kant’s handwritten Reflexion 2394 (16:343) as well as in the lecture transcript Logik Pölitz (24:539). Interestingly, while the third degree from the Jäsche Logik (“kennen (noscere)”) is missing in the lecture transcript, the fourth degree (“erkennen (cognoscere)”) is missing in the Reflexion (also cf. Logik Dohna-Wundlacken; 24:730). So the seven degrees in the Jäsche-Logik may be the result of Jäsche’s combining the two six-step lists. In any case, the distinction between the degrees of cognition and their ordering seem to be Kant’s own.
Cf. 9:91, where Kant seems to use “cognition” in the Stufenleiter sense by saying that “all cognition, that is, all representations related with consciousness to an object, are either intuitions or concepts” (also cf. 24:752). In the Jäsche Logik, Kant defines the relevant sense of “consciousness” as “a representation that some other representation is in me” (9:33).
Other readings of the Stufenleiter passage are possible. Thus, although Kant is most naturally read in this way, he does not explicitly claim that all intuitions and all concepts are cognitions (but rather only that all cognitions are either intuitions or concepts). This would allow for concepts that are not cognitions, and ideas might be examples thereof.
Some of these differences seem to derive from a difference in emphasis: Where the classification in the Logik focuses more on acts of cognition (“to perceive something”, “to understand something” etc.), the Stufenleiter is interested in kinds of representations (“perception”, “notion” etc.).
See also B146.
It may seem that existence cannot be required for givenness, since givenness is necessary for cognition and some of Kant’s formulations seems to suggest that what is required for cognition is not the existence, but only the real possibility of the cognized object (cf. e.g. Bxxivf. Fn). However, on our reading real possibility is required not for the givenness-, but for the thought-condition on cognition to be satisfied. Cf. below. For a more detailed argument for requiring that the givenness condition be understood as involving existence (for both empirical and mathematical objects), see Watkins and Willaschek (2017a, b). Note also, however, that the main point of this paper obtains regardless of whether one accepts our specific interpretation of the givenness condition.
Cf. the following passage, where Kant uses “given” first in the passive sense relevant for human beings and then in the active sense pertaining to the divine mind: “a divine understanding, which would not represent given objects, but through whose representation the objects would themselves at the same time be given, or produced” (B145; emphasis added).
Kant indicates that cognition involves more than a concept when he notes that if “an intuition corresponding to the concept could not be given at all, then it would be a thought as far as its form is concerned, but without any object, and by its means no cognition of anything at all would be possible, since, as far as I would know, nothing would be given nor could be given to which my thought could be applied” (B146). That is, it is not simply that we would in fact lack an object, but also that, as far as we could tell, no such object could be given. Kant’s interest in our representations’ “relation to an object” arises in the context of concepts, and thus the thought condition, rather than in that of intuitions, and the givenness condition.
This point is sometimes connected with the claim that objects must be shown to be really possible (Bxxvi), where real possibility is contrasted with logical possibility. Because Kant’s notion of real possibility is complex and does not match up neatly with that of, say, metaphysical possibility, we do not consider how it bears on the thought-condition. Kant’s discussion at, e.g., A96 is relevant, where he explicitly notes that one can think (but not cognize) “objects that are perhaps impossible, or that are perhaps possible in themselves but cannot be given in any experience”. Note also that on this understanding of real possibility if an object is given in intuition and thus exists, one still needs to establish that the object is really possible, for real possibility, as we understand it, is tied to a concept insofar as one must be able to show that the object could also be given in such a way that our concept can apply to it.
For example, Kant sometimes says that the categories must be shown to have “Sinn und Bedeutung”, or sense and significance (or meaning). Though Kant sometimes seems to identify these terms and to take them both to indicate reference, there are other passages (A241/B300) that suggest a subtler set of distinctions. According to these passages, it is one thing for a concept to refer to an object, another to demonstrate that it can refer, and yet another to have a sense of what an object is like or how the object would (have to) appear to us for the concept to apply to it (B149).
Kant adds a further specification to the thought-condition, namely that thought must contain positive content to contribute to cognition: “[i]t is not yet a genuine cognition if I merely indicate what the intuition of the object is not, without being able to say what is then contained in it; for then I have not represented the possibility of an object for my pure concept of the understanding at all” (B149).
This shows that the relation between a representation and its object that is constitutive of a cognition is not one of adequate representation or “agreement” (A58/B83).
Kant denies only that a cognition could be “wholly false” (24:93), presumably because this would make it impossible for the representation to stand in a suitable relation to its object.
Cf. Kant’s discussions of the relation between truth and cognition at B115 and A293/B350, neither of which decides the issue. The mere fact that “erkennen” in German is a success verb and thus appears to imply truth does not help either since Kant does allow for false cognition.
However, only singular cognitions directly represent individual objects as such; general cognitions, which satisfy the givenness-condition only in a more relaxed sense, do so only indirectly, by having singular cognitions as their instances.
Kant sometimes speaks of assent (“Fürwahrhalten”) as a kind of judging (e.g. 9:66), which makes sense insofar as assent is directed at a judgment. But the converse does not hold, since Kant explains judgment as a kind of complex conscious representation and never mentions assent. It is one thing consciously to represent some objective state of affairs (such as a ball’s being red), it is another to take it to be true that this state of affairs obtains (that the ball is red). This does not mean, however, that cognition consists in the mere entertaining of a thought. Rather, in the basic case, it is the awareness of the existence of an objects and some of its features.
“Judgment” (like Urteil) is ambiguous between the act of judging and the content judged. We do not want to deny that Kant sometimes uses “judgment” in the former sense, which may be assimilated to that of “assent(ing)”. Where Kant identifies judgment and cognition, however, he typically uses judgment in the latter sense (for a kind of representation).
We take this to be a claim not about cognition in general, but, as the context makes clear, about philosophical cognition in particular.
Where the English translation has Kant speaking of an “unknown object” (A479/B507), “unknown” is actually a translation of “unbekannt”, which is a cognate not of “knowledge” (Wissen), but rather of “cognition” (Erkenntnis).
Matters are different when it comes to the Metaphysical Foundations of Natural Science, which focuses on the question of scientific knowledge.
Moreover, while Kant expands on his accounts of opinion and belief in the course of the Canon, knowledge does not receive separate treatment, but is mentioned only in passing after the quoted passage. In fact, Kant’s main focus in Sect. 3 of the Canon is on belief (as the attitude appropriate towards God’s existence and the immortality of the soul). Thus, the only paragraph in the whole Critique that explicitly addresses the concept of knowledge does so in order to clarify, not the concept of knowledge, but that of belief. This underlines our previous observation that knowledge is not Kant’s central concern in the first Critique.
For a similar reading, cf. Pasternack (2011, p. 202).
For a different reading, cf. Chignell (2007, p. 33), who claims (against Kant’s insistence that knowledge requires certainty) that objectively sufficient grounds are fallible (42). This claim may rest on confusing fallibility in the sense that we can mistakenly take ourselves to know something (and thus to have objectively sufficient grounds) with the (for Kant incoherent) assumption that knowledge itself (and the grounds it is based on) could be fallible (also cf. next paragraph).
That certainty implies truth is supported by the fact that Meier, on whose textbook Kant based his logic lectures, defines certainty as “consciousness of the truth of a cognition” (1752, § 29). Kant accepts this definition “for the time being” in Logic Blomberg: (24:57). Also cf. Logic Blomberg: “the certainty of a cognition, on the other hand, rests on its objective truth” (24:143).
Cf. McDowell (2011). Our fallibility, according to McDowell, resides in our cognitive capacities (which can be misapplied), not in the individual case (which is either true or false and, if true and based on the right kind of reason, knowledge).
Cf. Engstrom (2009, p. 108 ff.), for a similar account of error in Kant.
Despite Gettier, most current epistemologists seem to accept some kind of tripartite definition of knowledge (true belief plus X).
For discussion of significant differences between these two notions, see Chignell (2007, p. 37).
In what follows, unless otherwise indicated, ‘cognition’ means ‘cognition in the narrow sense’. That cognition in the wider sense differs from knowledge should be obvious.
Since cognition requires the involvement of concepts whose objective reality must be established (e.g., in the Transcendental Deduction for the categories), cognition may require some kind of (“philosophical”) justification, but it does not require the kind of (“epistemic”) justification that is at issue with knowledge.
If Kant’s account of an objectively sufficient ground could be specified further, this claim might need to be revised accordingly. It is, to our mind, striking that Kant does not explicitly relate his account of objectively sufficient grounds to his account of cognition.
According to Kant, analytic judgments do not contain “determinations” (9:111), that is, do not “determine” their objects (also cf. 20:268: “Determining means judging synthetically”). Kant does at times speak of “analytic cognition”, but we are inclined to read those passages as referring to cognition in the broader sense.
It is true that the special kind of objective justification that Kant requires for knowledge involves apodictic certainty, i.e., consciousness of necessity, which might seem to bring cognition closer to knowledge after all. However, the kind of consciousness involved in apodictic certainty is different from the kind of consciousness that one has in cognition. For in the case of cognition, we must be conscious of the object, including its existence and features, whereas in the case of knowledge, we must be conscious of the (objective) grounds that justify our assent to a judgment, and of their adequacy as objective grounds. In short, though consciousness must be present in both cases, what one must be aware of in each case is quite different.
Indeed, even a cursory acquaintance with current literature devoted to contemporary philosophical topics reveals that the issues Kant is concerned with are still very much lively topics of debate, as is on display in, for example, Burge (2010, 2013). How experience can afford us a direct awareness of objects and their features is a central theme in McDowell (1994, 2009).
Although Kant does occasionally use some variant of the phrase ‘analytic cognition’ (cf. A8/B12), he does not, as far as we are aware, ever use the phrase analytic knowledge, so this point is a somewhat speculative suggestion. Still, one might think that it is hard to see how, given their nature, analytic propositions could be false, though one might well have concerns about whether they have the kind of semantic content that is necessary for them to refer to objects of a given domain and thus qualify as cognition.
This is not to exclude that the empirical world might be grounded in some but not all of the things in themselves. It excludes only that we could know which they are. In this sense, our knowledge of the intelligible ground of experience (the “transcendental object”; cf. A109) is purely general.
Both authors contributed equally to this paper. For helpful comments on earlier versions of this paper we thank audiences at the Deutsche Kongress für Philosophie 2014 in Münster, Germany, a conference on Kant and Knowledge and Cognition at the University of Wisconsin, Madison, and at a meeting of the History of Philosophy Roundtable at University of California, San Diego as well as referees for Synthese.
References
Allison, H. (2004). Kant’s transcendental idealism (2nd ed.). New Haven/London: Yale University Press.
Burge, T. (2010). The origins of objectivity. Oxford: Oxford University Press.
Burge, T. (2013). Cognition through understanding: Self-knowledge, interlocution, reasoning, reflection: philosophical essays (Vol. 3). Oxford: Oxford University Press.
Carriero, J. (2013). Epistemology Past and Present. Proceedings of the Aristotelian Society, 113, 175–200.
Chignell, A. (2007). Kant’s concepts of justification. Nous, 41, 33–63.
Chignell, A. (2014). Modal motives for noumenal ignorance: Knowledge, cognition, coherence. Kant-Studien, 105, 573–97.
Engstrom, S. (2009). The form of practical knowledge. Cambridge, MA: Harvard University Press.
Grüne, S. (2009). Blinde Anschauung. Zur Rolle von Begriffen in Kants Theorie sinnlicher Synthesis. Frankfurt am Main: Klostermann.
Guyer, P. (1987). Kant and the claims of knowledge. Cambridge: Cambridge University Press.
Henrich, D. (1973). Die Beweisstruktur von Kants Transzendentaler Deduktion. In G. Prauss (Ed.), Kant. Zur Deutung seiner Theorie von Erkennen und Handeln (pp. 93–104). Köln: Kiepenheuer und Witsch.
Kain, P. (2010). Practical cognition, intuition, and the fact of reason. In J. Krueger & B. Lipscomb (Eds.), Kant’s moral metaphysics (pp. 211–230). Berlin: de Gruyter.
McDowell, J. (1994). Mind and world. Cambridge: Harvard University Press.
McDowell, J. (2009). Having the world in view: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press.
McDowell, J. (2011). Tyler Burge on disjunctivism. Philosophical Explorations, 13, 243–255.
Meier, G.F. (1752). Auszug aus der Vernunftlehre. Halle: Johann Justinius Gebauer.
Pasternack, L. (2011). Kant’s Doctrinal Belief in God. In O. Thorndike (Ed.), Rethinking Kant (Vol. 3). Cambridge: Cambridge Scholars Publishing.
Schafer, K. (2013, November). Kant’s conception of cognition and our knowledge of things-in-themselves (unpublished manuscript).
Tolley, C. (2013, January 28). Kant’s use of Erkenntnis in its historical context (unpublished manuscript).
Watkins, E., & Willaschek, M. (2017a). Kant’s account of cognition. Journal of the History of Philosophy, 55, 83–112.
Watkins, E., & Willaschek, M. (2017b). Givenness and cognition: Reply to Grüne and Chignell. Journal of the History of Philosophy, 55, 143–152.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Willaschek, M., Watkins, E. Kant on cognition and knowledge. Synthese 197, 3195–3213 (2020). https://doi.org/10.1007/s11229-017-1624-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-017-1624-4