Skip to main content

Part of the book series: Veröffentlichungen des Instituts Wiener Kreis ((WIENER KREIS,volume 28))


Hermann Cohen’s 1871 classic, Kants Theorie der Erfahrung, had a formative influence, not only on the Marburg school’s reading of Kant, but on their entire conception of philosophy. This influence was further magnified by the substantially revised and expanded second edition of 1885 and the yet further expanded third edition of 1918. Neo-Kantianism was the dominant philosophical movement in Germany in the late nineteenth and twentieth centuries, which means that a work, ostensibly, of Kant scholarship had an influence on the development of German philosophy that few works of secondary literature can claim.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Institutional subscriptions

Similar content being viewed by others


  1. 1.

    E.g. Vaihinger’s careful point-by-point unraveling of Cohen’s reading of the Transcendental Aesthetic in Vaihinger 1887, vol. 2. Even Vaihinger allows his contempt to shine through at moments: “Unverständlich, wie so häufig, Cohen. Th. d. Erf.” (Vaihinger 1892, vol. 2, p. 138). Thanks to Des Hogan for pointing me to that remark.

  2. 2.

    For instance, Heidegger: “die Absicht der Kritik der reinen Vernunft bleibt demnach grundsätzlich verkannt, wenn dieses Werk als ‘Theorie der Erfahrung’ oder gar als Theorie der positive Wissenschaft ausgelegt wird” (Heidegger 1929, pp. 16–17). Cf. Ebbinghaus, J. 1954.

  3. 3.

    Köhnke (1986, pp. 273–275), Brandt (1993, pp. 37–54), Poma (1997, pp. 18, 48–53); and Kuehn (2009, pp. 115–121). Beiser (2014, p. 489) is more balanced, but ultimately is critical of Cohen’s reading.

  4. 4.

    Köhnke (1986, p. 273). Cf. Luft (2015, pp. 43–48).

  5. 5.

    Friedman (1992) and (2013). This, despite the fact that this sentence of Cohen’s could almost function as a summary of Friedman’s decade-long engagement with Kant and Newton: “the transcendental method arose through a reflection upon the Philosophiae naturalis principia mathematica” (KTE 2, p. 67).

  6. 6.

    It should also be noted, that whatever its flaws, contemporary Kant scholarship is deeply indebted to KTE, one of the first works to attempt a rigorous, textually grounded interpretation of Kant’s thought. In doing so, KTE helped create the scholarly tradition that now largely rejects it.

  7. 7.

    The very next sentence reads: “die Kritik der reinen Vernunft ist Kritik der Erfahrung.” I will not attempt to rehabilitate that claim, for, as many scholars have pointed out, the KrV is fundamentally about the possibility of metaphysics, and only secondarily about experience (because the synthetic a priori cognitions of metaphysics are made possible by their relation to experience).

  8. 8.

    “Kants Aufgabe ist also zunächst die Prüfung und Kennzeichnung des Erkenntniswertes und des Gewissheitsgrundes der Netwonschen Naturwissenshcaft, welche er be dem Drohwort der Erfahrung fasste” (KTE 3, p. 93); in the first edition see KTE 1, pp. 206, 208. On the development in Cohen’s reading of Kant see Edel’s Introduction to Werke I.1, pp. 20*, 22–23*.

  9. 9.

    The closely related (though distinct) notion of the infinitesimal plays a key role in Cohen’s philosophical development after 1871 as well as in his mature theory of critical idealism. For reasons of space I cannot here discuss the relation between my invocation of the Bolzano-Weierstrass ε-δ definition of limit and Cohen’s concept of the limit and the infinitesimal. For Cohen’s concept of limits and the infinitesimal method see PIM.

  10. 10.

    An admittedly free rendering of this remark: “Denn in diesem Falle gilt das, was ursprünglich selbst nur Erscheinung ist, z. B. eine Rose, im empirischen Verstande für ein Dinge an sich selbst” (A30/B45).

  11. 11.

    I have explored some of the same themes in Stang (2015a). An earlier, less adequate, account of universal experience can be found in Stang (2012).

  12. 12.

    Namely, ideas (Locke) or impressions (Hume) of sensation and reflection.

  13. 13.

    The same sentence is repeated nearly word for word in Prol, Ak. 4:292.

  14. 14.

    Another possibility is that there is an ambiguity in the notion of ‘necessity’ involved. However, if we accept the taxonomy of the kinds of necessity argued for in Stang (2011), this does not hold. For of the four kinds of necessity distinguished there—logical, formal, empirical, and noumenal–only two are relevant in this context (formal, empirical) and it is both the case that (a) Locke and Hume would deny that experience contains necessities of either kinds and (b) Kantian experience does acquaint us with such necessities. The conclusion argued for in the main text thus holds: Kant’s claim at B3 involves something other than his ‘official’ notion of experience, a more minimal notion of experience he is willing to deploy to express his (limited) agreement with the empiricists Locke and Hume. For more on Kantian modalities, see Stang (2016a).

  15. 15.

    For instance, at B147 he writes that the categories “dienen nur zur Möglichkeit empirischer Erkenntnis. Diese aber heißt Erfahrung”; at the beginning of the same paragraph he writes: “sich einen Gegenstand denken und einen Gegenstand erkennen, ist also nicht einerlei. Zum Erkenntnisse gehören nämlich zwei Stücke: erstlich der Begriff, dadurch überhaupt ein Gegenstand gedacht wird (die Kategorie), und zweitens die Anschauung, dadurch er gegeben wird” (B146).

  16. 16.

    This corresponds, roughly, to applying an abstract idea to an idea of sensation or reflection (Locke) or applying an abstracted idea to an impression of sensation or reflection (Hume). It would make no sense for Locke or Hume to claim that all of our ideas arise from experience in this Kantian ‘cognitive’ sense (experience #2), for experience is here defined in terms of applying an idea to a sensory impression.

  17. 17.

    A110, A124 ff., A156/B195, B161, B218, A213/B260, and A183/B226.

  18. 18.

    The Analogies of Experience are structured around the distinction between perception (Wahrnehmung) and experience (Erfahrung); see especially B218–219. In thinking about Kant’s technical notion of perception (Wahrnehmung), its difference from experience on the one hand and empirical intuition on the other, as well as its role in the Analogies, I have benefited tremendously from reading Clinton Tolley’s unpublished paper “Kant on the distinction between perception and experience.” My discussion in the rest of this section is deeply indebted to Tolley’s excellent work.

  19. 19.

    Consider that the titles of the Principles of Experience, at least in the A Edition, speak of “Erscheinungen […] ihre Anschauung nach (Axiomen der Anschauung, A162)”, “das Reale, was ein Gegenstand der Empfindung ist” (Antizipationen der Wahrnehmung, A166), and finally “alle Erscheinungen stehen, ihrem Dasein nach, […]” (A176; my emphasis throughout). A160/B199, A178/B210. So we proceed from the principle of the intuition of appearances (Axioms) to the principle of the perception of reality in appearances (Anticipations) to the principle of the experience of their existence (Analogies). I owe appreciation of this point to Tolley “Kant on the distinction.”

  20. 20.

    Kant writes in the A Deduction: “das erste, was uns gegeben wird, ist Erscheinung, welche, wenn sie mit Bewusstsein verbunden ist, Wahrnehmung heißt […] Weil aber jede Erscheinung ein Mannigfaltiges enthält, mithin verschiedene Wahrnehmungen in Gemüte an sich zerstreut und einzeln angetroffen werden, so ist eine Verbindung derselben nötig, welche sie in dem Sinne selbst nicht haben können. Es ist als in uns ein tätiges Vermögen der Synthesis eises Mannigfaltigen, welches wir Einbildungskraft nennen, und deren unmittelbar an den Wahrnehmungen ausgeübte Handlung ich Apprehension nenne. Die Einbildungskraft soll nämlich das Mannigfaltige der Anschauung in ein Bild bringen; welcher muß sie also die Eindrücke in ihre Tätigkeit aufnehmen, d. i. apprehendieren” (A120). Again, thanks to Clinton Tolley for drawing my attention to passages like this.

  21. 21.

    I mean ‘appearing’ here in the empirical sense, in which it is used to mark the difference between the real empirical properties of objects in space and time (phenomena) and how they appear to perceivers’ sense organs at a given time under given conditions. The transcendental sense of ‘appearing’ marks the difference between the real empirical properties of objects in space and time (phenomena) and their unknowable inner constitution (noumena). In this paper I am trying to remain as neutral as possible on the second distinction, which has received the lion’s share of attention from commentators on Kant’s idealism. For the empirical appearance/thing in itself distinction, and its difference from the transcendental distinction, see A29–30/B45 and A45–46/B62–63.

  22. 22.

    By ‘existence’ he cannot mean the causal efficacy of the object, as he seems to in other contexts (e.g. Ak. 4:468), for that would render the conclusion of the second and third Analogies nearly trivial: to represent an object as causally efficacious we must represent is using the categories <cause-effect> and <community>.

  23. 23.

    Kant makes this point about experience in sense #3 (synthetic unity of perceptions) at B218, but it follows immediately from his definition of experience (#2 on my reading) as empirical cognition at B147 and the discussions of intuition and perception in the Axioms and the Anticipations.

  24. 24.

    A192/B237 and A190/B235, respectively.

  25. 25.


  26. 26.

    “Denn überhaupt würden wir, nach Gesetzen der Sinnlichkeit und dem Kontext unserer Wahrnehmung, in einer Erfahrung auch auf die unmittelbare empirische Anschauung derselben stoßen, wenn unsere Sinnen feiner wären, deren Grobheit die Form möglicher Erfahrung überhaupt nichts angeht” (A226/B273).

  27. 27.


  28. 28.

    “Wo also Wahrnehmung und deren Anhang nach empirischen Gesetzen hinreicht, dahin reicht auch unsere Erkenntnis vom Dasein der Dinge” (A226/B273).

  29. 29.

    One also finds the idea of a single ‘universal’ experience at A230/B283, A493/B521, A495/B524, and A582/B610.

  30. 30.

    The material in this section dovetails with the discussion in Stang (2016b), where I use similar arguments to critique the interpretation of Allais (2015).

  31. 31.

    This holds true regardless of which specific notion of experience from §1 we have in mind.

  32. 32.

    These notions are defined as follows in Stang (2011) (cf. Stang 2016, Sect. 7.3):

    Formal possibility It is formally possible that p just in case it is compatible with our forms of experience that p.

    Formal necessity It is formally necessary that p just in case it is incompatible with our forms of experience that not-p.

  33. 33.

    These formulae have to be distinguished because existence is not a real predicate; we cannot assimilate the fact there is an object with property F to the fact that this object has some other property, e.g. the property of existing. Consequently, existential facts need to be formulated in general terms, as in Ex (e.g. there is an x such that Fx). The left-hand side of (3) should not have the form ‘appearance a exists’ because existence is not a predicate some objects have and others lack—in other words, if the sentence ‘Appearance a exists’ is well-formed (the singular term refers) then it is true. See Stang (2015b) and (2016a) for further discussion.

  34. 34.

    See Allais (2015, p. 47) and Langton (1998, pp. 144–145) for discussion.

  35. 35.

    “Eben um deswillen sind auch die Grundsätze der Modalität nichts weiter, als Erklärungen der Begriffe der Möglichkeit, Wirklichkeit und Notwendigkeit in ihrem empirischen Gebrauche” (A219/B266).

  36. 36.

    The Postulate of possibility (quoted in the main text) does give, for instance, merely epistemic conditions under which we can obtain justification for judgments about possibility. It tells us what possibility is for empirical objects (see previous note).

  37. 37.

    The possibility of the skeptical scenario should not be confused with the fact that objects we experience are the appearances of things we can never know, or that perhaps the very objects we experience also have inner natures we can never know (depending on how one interprets Kant’s transcendental idealism). The skeptical scenario concerns how these things in themselves appear to us, not how they are in themselves. In the skeptical scenario, the existence of these appearances and their empirical properties (how they appear to us) could be in principle inaccessible to our natural-scientific experience.

  38. 38.

    E.g. “alles ist wirklich, was mit einer Wahrnhemung nach Gesetzen des empirischen Fortgangs in einem Kontext steht. Sie sind also alsdenn wirklich, wenn sie mit meinem wirklichen Bewußtsein in einem empirischen Zusammenhange stehen, ob sie gleich darum nicht an sich, d.i. außer diesem Fortschritt der Erfahrung, wirklich sind” (A493/B521).

  39. 39.

    Cf. Stang (2011) and Stang (2016a, Sect. 9.3).

  40. 40.

    I am guided here by the interpretation of Friedman (1992); his interpretation is somewhat modified in Friedman (2013). On this point, I agree more with Friedman’s earlier interpretation, though I do not have space here to explain my disagreements with the finer points of the meticulous analyses in Friedman (2013).

  41. 41.

    MAdN, Ak. 4:559.

  42. 42.

    MAdN, Ak. 4:555–556.

  43. 43.

    See the discussion of ‘alternative’ judgments at Ak. 4:556 and the footnote at 559–560.

  44. 44.

    This relies on reading MAdN as a reconstruction, within the contours of Kantian transcendental philosophy, of Newton’s project in the Principia. On this point see Friedman (1992, pp. 159–164) and the much more extensive discussion in Friedman (2013, esp. pp. 531–561).

  45. 45.

    Cf. Friedman (1992, pp. 47–48), and Friedman (2013, pp. 156–158, 534–536).

  46. 46.

    One of the most important passages in the “Transcendental idealism as key” section is ambiguous on what kind of modal/subjunctive analysis Kant wants to employ: “Es ist im Ausgange ganz einerlei, ob ich sage, ich könne im empirisichen Fortgange im Raume auf Sterne treffen, die hundertmal weiter entfernt sind, als die äußersten, die ich sehe: oder ob ich sage, es sind vielleicht deren im Weltraume anzutreffen, wenn sie gleich niemals ein Mensch wahrgenommen hat, oder wahrgenommen wird” (A496/B524). I have argued though that it is very much not “einerlei” whether we say that there actually are such distant stars in virtue of the fact that we could experience them (too weak), or in virtue of the fact that they are there to be experienced (this is circular), or in virtue of the fact that if were in the right conditions we might experience them (too strong). I have argued that Kant should have said explicitly that we would experience them “im empirisichen Fortgange im Raume” if that “Fortgang” continued far enough.

  47. 47.


  48. 48.

    Readers concerned that this depends upon the transcendental illusion of assuming that the complete space of possible properties is ‘given’ should see Stang (2012) for further discussion, as well as Stang (2016a, pp. 290–292).

  49. 49.

    In KrV Kant argues that all objects of experience überhaupt have quantitatively determinate extensive magnitudes (Axioms of Intuition, A161/B202) and intensive magnitudes (Anticipations of Perception, A165/B207). This is what makes possible the application of mathematics to objects of experience in general. In MAdN he argues for the more specific claim that all bodies (outer objects composed of matter, the movable in space) have quantitatively determinate motions (Phoronomy), fill quantitatively determinate regions of space (Dynamics), and exert quantitatively determinate forces upon one another (Mechanics). This is what makes possible the application of mathematics to bodies. An extensive magnitude is one, the representation of which requires representing its parts; an intensive magnitude is a degree of reality, not possessed in virtue of having a multiplicity of parts. Kant’s categorization of various magnitudes (e.g. velocity, mass, force) as either intensive or extensive is complicated and I cannot go into the details here.

  50. 50.

    The continuity of the extensive magnitude of objects (between any two extensive magnitudes an intermediate magnitude is possible) follows from the continuity of space (A169/B211). The continuity of intensive magnitude is more complicated; Kant argues for it in the Anticipations of Perception (B208). The Anticipations of Perception, and the idea of continuity and intensive magnitude, became deeply significant for Cohen’s Kant interpretation and his systematic philosophy, starting with the publication of PIM in 1883 and continuing with KTE 2 in 1885; see his discussion of Kant’s theory of intensive magnitudes in PIM, 105–113, and KTE 3, 538–562, as well as the discussion of these themes in his systematic work LRE, 418–420, 462–464.

  51. 51.

    This is a desirable feature of the construction because it allows for a middle way between the extremes of an object being wholly indeterminate with respect to a property or having a fully determinate value with respect to that property.

  52. 52.

    Cohen’s anti-psychologism is complicated in KTE 1 by his (atavistic?) commitment to Herbartian psychology (KTE 1, pp. 122, 128); in KTE 2 and KTE 3 he is more clearly opposed to any psychological interpretation of Erfahrung. See Beiser 2010, p. 487 (as well as Edel’s Introduction in KTE 3, p. 19*) on the Herbartian strand in KTE 1.

  53. 53.

    “Kants Aufgabe ist also zunächst die Prüfung und Kennzeichnung des Erkenntniswertes und des Gewissheitsgrundes der Netwonschen Naturwissenshcaft, welche er be dem Drohwort der Erfahrung fasste” (KTE 3, p. 93).

  54. 54.

    See the references to Friedman in note 44.

  55. 55.

    MAdN, Ak. 4:472.

  56. 56.

    Ak. 4:298. I cannot here go into detail on the my interpretation of the relation among the different senses of experience distinguished in §2 and the Prolegomena distinction between judgments of perception and judgments of experience.

  57. 57.

    What Cohen calls “das Faktum der Wissenschaft” (KTE 1, pp. 41, 94).

  58. 58.

    Various scholars have made the point that Cohen effectively conflates the “analytic” method of the Prolegomena with the “synthetic” method of the KrV; see Beiser (2014, p. 489) and Luft (2015, pp. 224–225). In Cohen’s defense, he is aware of this distinction and explicitly argues that Kant in fact adopts an analytic or regressive method in the KrV.

  59. 59.

    One might think that we can address the sense in which we ‘construct’ nature in experience by appeal to Kant’s theory of geometric construction in pure intuition, but there are two barriers to this. First, Kant distinguishes between objects of pure intuition (mathematical objects) and concrete causally efficacious objects in space and time, the objects that constitute nature, the objects of experience (mathematical natural science). Construction of the latter, whatever it is, is going to be quite different than construction of the latter. Secondly, the whole notion of mathematical construction in pure intuition is itself a complex one in Kant, so we might be in danger of explain obscurum per obscurius.

  60. 60.

    Ak. 4:472.

  61. 61.

    My argument here, even some of my formulations, are clearly indebted to the work of Michael Friedman here. However, given my explicit construction of universal experience and my more ‘idealist’ reading of how phenomena depend on universal experience (see §§ 2–3) I claim something stronger than Friedman: what it is for an object to have a certain motion is for it to be represented as having that motion in a universal experience whose stages are constructed using Newtonian laws. I do not have the space here to explore this deviation from Friedman’s less ‘metaphysical’ reading.

  62. 62.

    For this tripartite classification of interpretations of transcendental idealism see the classic paper Ameriks (1982).

  63. 63.

    In a parallel discussion at A45/B63, Kant distinguishes between a rainbow as an appearance in the empirical sense (in which it is colored) and as a thing in itself in the empirical sense (in which it is a collection of colorless raindrops). He identifies the rainbow ‘in itself’ as the rainbow as it is represented in “allgemeine Erfahrung.”

  64. 64.

    I would like to thank Christian Damböck and the audience at a conference at the University of Vienna in the Fall of 2014 for their questions and feedback on an earlier draft of this paper. Special thanks also to Ian Drummond for excellent copy-editing.


  • Allais L (2015) Manifest reality: Kant’s idealism and his realism. Oxford University Press, Oxford

    Book  Google Scholar 

  • Ameriks K (1982) Recent work on Kant’s theoretical philosophy. Am Philos Q 19(1):1–24

    Google Scholar 

  • Beiser F (2002) German idealism: the struggle against subjectivism 1781–1801. Harvard University Press, Cambridge (Mass.)

    Google Scholar 

  • Beiser F (2014) The genesis of Neo-Kantianism, 1796–1880. Oxford University Press, Oxford

    Book  Google Scholar 

  • Brandt R (1993) Hermeneutik und Seinslehre bei Hermann Cohen. In: Brandt R, Orlik F (eds) Philosophisches Denken-Politisiches Wirken: Hermann Cohen-Kolloquium Marburg 1992. Olms, Hildesheim, pp 37–54

    Google Scholar 

  • Friedman M (1992) Kant and the exact sciences. Harvard University Press, Cambridge (Mass.)

    Google Scholar 

  • Friedman M (2013) Kant’s construction of nature: a reading of the metaphysical foundations of natural science. Cambridge University Press, Cambridge

    Book  Google Scholar 

  • Heidegger M (1929) Kant und das Problem der Metaphysik. Vittorio Klostermann, Frankfurt a. M.

    Google Scholar 

  • Köhnke KC (1986) Entstehung und Aufstieg des Neukantianismus. Suhrkamp, Frankfurt a. M.

    Google Scholar 

  • Kuehn M (2009) Interpreting Kant correctly: on the Kant of the Neo-Kantians. In: Makkreel R, Luft S (eds) Neo-Kantianism in contemporary philosophy. Indiana University Press, Bloomington, pp 113–131

    Google Scholar 

  • Langton R (1998) Kantian humility. Oxford Univeristy Press, Oxford

    Google Scholar 

  • Luft S (2015) The philosophy of the Marburg school: from the critique of scientific cognition to the philosophy of culture. In: Warren N de, Staiti A (eds) New approaches to Neo-Kantianism. Cambridge University Press, Cambridge, p 221–239

    Chapter  Google Scholar 

  • Poma A (1997) The critical philosophy of Hermann Cohen. SUNY Press, Albany

    Google Scholar 

  • Stang N (2011) Did Kant conflate the necessary and the a priori? Noûs 45(3):443–471

    Google Scholar 

  • Stang N (2012) Kant on complete determination and infinite judgment. Br J Hist Philos 20(6):1117–1139

    Article  Google Scholar 

  • Stang N (2015a) Who’s afraid of double affection? Philosophers’ Imprint 15(8):1–28

    Google Scholar 

  • Stang N (2015b) Kant’s argument that existence is not a determination. Res 91(3):583–626

    Google Scholar 

  • Stang N (2016a) Kant’s modal metaphysics. Oxford University Press, Oxford

    Book  Google Scholar 

  • Stang N (2016b) Appearances and things in themselves: actuality and identity. Kant Rev 21(2):283–292

    Article  Google Scholar 

  • Vaihinger H (1892) Commentar zu Kants Kritik der reinen Vernunft, Vol 2. Spemann, Stuttgart

    Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Nicholas F. Stang .

Editor information

Editors and Affiliations

Rights and permissions

Reprints and permissions

Copyright information

© 2018 Springer International Publishing AG

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Stang, N.F. (2018). Hermann Cohen and Kant’s Concept of Experience. In: Damböck, C. (eds) Philosophie und Wissenschaft bei Hermann Cohen/Philosophy and Science in Hermann Cohen. Veröffentlichungen des Instituts Wiener Kreis, vol 28. Springer, Cham.

Download citation

  • DOI:

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-319-58022-7

  • Online ISBN: 978-3-319-58023-4

  • eBook Packages: Social Science and Law (German Language)

Publish with us

Policies and ethics