This is the first of what I hope to be an ongoing series of posts concerning logic and theories of judgment in the 16th–18th centuries.
To start things off I though I’d comment on something I came across recently in the philosophy blogosphere. Over at the blog New Apps Catarina Dutilh Novaes (CDN) has an interesting post which touches on the normative import of logic for reasoning or thought more generally. In the course of this discussion CDN claims that
the view of logic as having normative import for thought is entirely misguided. It is a relic of Kantian transcendental idealism that most philosophers still hold on to, but usually somewhat uncritically.
What is it for logic to have “normative import”? The idea here, I take it, is that logic has normative import when it functions as the measure of thought and as that to which thought ought to conform. Hence, a thought is found logically wanting if it does not conform to logic (though what particular logic I leave open here) in its internal structure or functional relationship to other thoughts. Call this “normativism” about logic.
In the piece of hers to which CDN links, she argues that
With Kant, logic no longer primarily concerns argumentation; instead, it concerns the inner mental activities of the lonesome thinking subject…According to him, (general) logic deals with “absolutely necessary rules of thought without which there can be no employment whatsoever of the understanding” (Kant 1781/1787/1929, Critique of Pure Reason, p. A52/B76)…thus the ideas that thinking belongs to the realm of normative phenomena and that logic provides the canons for correct thinking are essentially Kantian theses, which are intimately related to his critical project and to transcendental idealism (as argued in MacFarlane 2000).
So, according to CDN (and citing MacFarlane in support), Kant is the originator of normativism about logic, and his normativism is tightly linked to his transcendental idealism, such that insofar as we reject the latter, this gives us reason to reject the former. I believe this conception of Kant mistaken. Here I want to focus on two claims. First, whether Kant thinks of logic as normative in the sense defined. Second, whether Kant’s conception of logic is inextricably tied to his transcendental idealism.
Normative or Constitutive?
For Kant, logic is normative only in the sense that it constitutes the nature of rational thought and only in this sense “contains the utterly necessary rules of thought, without which no use of the understanding takes place” (A52/B76). It has normative import for human subjects only insofar as we take ourselves to be rational thinkers. But this is different from logical normativism.
In order to understand this we need to understand the difference between being constitutive and normative for thinking. Here I rely heavily on Clinton Tolley’s insightful article on Kant on the nature of logical laws.
If logical normativism is to be ascribed to Kant then he must endorse at least the following three claims:
- Thinking subjects can succeed or fail to act (i.e. to think) in accord with logical laws.
- A subject may be thinking even is she fails to act in accord with logical laws.
- Logical laws are normative (i.e. prescriptive or binding) even if no one has or is correctly followed them (though we might add that subjects must be able to follow them).
Tolley uses traffic laws as an example. Drivers can succeed or fail to act in accordance with traffic laws, they still count as driving even when they fail, and the traffic laws are binding even if drivers all fail to heed them. For Kant thinking is importantly different from driving in the respect that a mental act won’t count as thinking if it is not in accordance with logical laws. What are these laws? They include the principle of non-contradiction, and the various principles laid out by Kant’s notorious table of judgment (A70/B95).
This is how Kant distinguishes objective thinking from the mere subjective association of representations. In §22 of his Prolegomena he states this very clearly.
To think, however, is to unite representations in a consciousness. This unification either arises merely relative to the subject and is contingent and subjective, or it occurs without condition and is necessary or objective. The unification of representations in a consciousness is judgment. Therefore, thinking is the same as judging or as relating representations to judgments in general…The logical moments of all judgments are so many possible ways of uniting representations in a consciousness. If, however, the very same moments serve as concepts, they are concepts of the necessary unification of these representations in a consciousness, and so are principles of objectively valid judgments (4:304–5; cf. §19 of the Transcendental Deduction (B141–2)).
Kant argues here that the “logical moments” of judgment/thought are what distinguish these judgments from mere subjective association (cf. B142).
So if Kant endorsed logical normativism he would think that logic specified what it is for a subject to be thinking well (the way ethics specifies what it is for a subject to act well). In contrast, I have suggested that Kant endorse a constitutivist conception of logic. Logic gives the conditions constitutive of thought. One does not count as thinking, according to Kant, if one’s mental states do not stand in logical relations to one another. So (1) and (2) above do not apply here. In this sense Kant disagrees with Locke’s conception of thinking as signifying,
that sort of operation of the Mind about its Ideas, wherein the Mind is active; where it with some degree of voluntary attention, considers any thing (II.ix.1).
Kant thus distinguishes thinking as a particular species of what might be termed mentation. Mentation involves representations and their relations (in the psychologistic idiom of the time), whereas thinking involves representations that stand in logical relations to one another.
As for (3) above, Kant certainly conceives of the laws of thought as independent of the existence or mental activity of any human thinker. So this point holds for his constitutivist view. But this is insufficient to claim that Kant endorses normativism.
So, far from concerning merely the “lonesome inner subject”, Kant’s conception of logic and its relation to thought puts logic front and center as constitutive of the objectivity of thought and the possibility of genuine intersubjective communication and agreement (or disagreement).
Logic and Transcendental Idealism
CDN claims, citing MacFarlane 2000 (dissertation is available here), that Kant’s normativism is tied to his transcendental idealism. If what I said above is correct, then Kant does not really endorse normativism, but rather what I have called “constitutivism” concerning logic’s import for thought.
But even if one doesn’t accept this, I see little reason for linking transcendental idealism and normativism. MacFarlane’s argument, which CDN cites, is concerned with the nature of Kant’s conception of logic as formal rather than as normative, and as MacFarlane himself notes, Kant’s conception of logic does not require transcendental idealism (MacFarlane argues at best only for the converse, see. p. 126).
Moreover, since Kant quite clearly links logic to the laws of the understanding (A52/B76) and the understanding and its activities are intelligible quite independently of sensibility, it is difficult to see how his idealism could really be linked to his logical views. For Kant’s idealism, as he himself points out, stems from his claims concerning space and time as forms of human intuition (A490/B518ff). This means that Kant’s idealist argument is tied to claims concerning sensibility rather than the understanding. Hence it would seem they are independent of any claim concerning logic.
In future posts I intend to discuss more fully Kant’s (and his early modern predecessors) conception of logic and its relations to contemporary work.