Critique of Pure Reason
第 3 章 · 24 分钟

Categories and the transcendental deduction: the question of right

Not where the categories come from, but by what right they apply to objects.

概念地图
Categories and deduction
Not where the categories come from, but by what right they apply to objects.
深色为本章概念,浅色为其他章
陪练模式
不想只是读?让它带你走一遍。
导师把这一章拆成小步,每一步先问你,再讲;你用自己的话答,它先认下对的部分再纠偏。用你自己的 AI,对话只存在这台设备上。
对话只存在本机
Locus01

延伸阅读

Locke · 1690 · the Essay on the origin of ideas
Sets philosophy's main work as tracing where our ideas come from. He argued against innate ideas; the cost is that an answer about origin gets taken for an answer about entitlement — once the origin is traced, by what right the concept applies to objects is still unanswered.
Foundational
Hume · the conclusion about causation
The origin was traced — a habit formed by repeated succession — and the conclusion was that the concept has no objective validity for objects. This is the strongest consequence of running origin and entitlement together, and Kant's immediate reason for prying them apart.
Foundational
Kant · the transcendental deduction · A 1781 / B 1787
Splits the problem explicitly into a question of fact and a question of right, and says the deduction answers only the second. He argued against letting a genetic account stand in for a justification; the cost is that the argument must go through the unity of self-consciousness, and the A version leaned so far toward psychology that B rewrote it.
Turn
Strawson · 1966 · The Bounds of Sense
Holds that the deduction contains a core argument detachable from Kant's transcendental psychology. It tells the reader that this chapter has a skeleton that can be carried away, separable from its eighteenth-century casing.
Revision
机制02

Splitting fact from right

Kant borrows a division from law: how something in fact came about, and by what right a claim is made, are two questions. What answers the second for the pure concepts is the transcendental deduction. This cut is the chapter's most consequential move: Hume's conclusion presses only because an answer of the first kind was used to settle a question of the second kind. Once the two come apart, 'causation is a habit' — even granted — no longer yields 'causation has no objective validity'.

机制One division of labour, through separating origin from entitlement, stops 'we do use it' being read as 'we may use it'.
可迁移性测试
Move it to code compliance: which repository a snippet was copied from is its origin; whether it is licensed for this product is its entitlement. Tracing the origin perfectly still does not grant a licence. The isomorphism breaks in that a licence is a document you can look up, while a category's entitlement can only be argued.
机制03

From 'they must all be mine' to rules

The pivot of the deduction is not an object but a premise that hardly looks like one: every representation I can call mine must be able to be accompanied by an 'I think'. Kant's name for that self-consciousness is apperception. What it does in the argument is turn an extremely weak requirement into an extremely strong constraint: to fall under one and the same 'I', a manifold cannot lie about separately but must be combined by some set of rules, and the categories are that set.

机制A manifold belonging to one and the same self, through having to be combined by one set of rules, makes the categories conditions of experience.
可迁移性测试
Move it to log consolidation: records from several machines can count as one session only if they share fields and an ordering rule; without it they are unrelated entries. The isomorphism breaks in that logging rules are chosen and revisable, while Kant argues his set cannot be swapped.
Derivation04

The skeleton of the deduction

What is being reconstructed is the main line of the B deduction, and what it does and does not establish.

Every representation of mine must be able to be accompanied by 'I think'
A premise Kant chose, and the starting point of the B deduction. It looks feeble and ends up carrying all the weight.
For a manifold to belong to one self-consciousness it must be synthetically combined
A premise Kant chose. Lying about separately does not make one consciousness, and this is never independently proved.
Every way of combining is given by the functions of judgement, and the categories are those functions applied to the manifold of intuition
A premise Kant chose. It treats the judgement table of the logic of the day as a complete list, and is the most fragile line in the chapter.
推导 · 0 / 4
裂缝05

Cracks in this reading

争议地形06
本书主张
The categories hold objectively of every possible object of experience, because a manifold not combined under them could never become the object of one self-consciousness.
另一种看法
One line holds that the deduction proves at most that there must be some rules of combination, never that they must be these twelve; the particular list is a product of logic and science, revisable and in fact revised.
分歧扎在
The disagreement is rooted in whether a validity proof can carry a completeness proof with it, not in whether rules of combination are needed. Both sides grant that a manifold lying about separately is not an experience.
什么证据能裁决
What would settle it is the provenance of the table of judgements: if the twelve forms were taken from the logic textbooks of the day and later logic classifies judgements differently, then completeness has no independent support. This can be rechecked from the history of logic, with no reading of Kant.
The balance tips to the second: modern logic's classification of judgements does not match Kant's table. That defeats the list, not the skeleton — rules are still needed. What is missing is a route to fixing the number of categories that does not go through a table of judgements; in two centuries nobody has supplied one.
Textual test07

The 'deduction answers only the question of right' reading as a testable specification

The reading under test: the deduction's task is confined to the question of right, and its main line does not rest on the psychological origin of the categories.

Edition: A 1781 and B 1787 side by side, the deduction cited by standard A/B pagination with a standard translation alongside
Fixed in advance; B rewrote the deduction, so both editions must be checked.
Decisive passages: the paragraph of the introduction to the deduction that raises the two questions, and the sections of the B main line
Specified in advance, not picked after reading.
What counts as support: the two questions appear paired with a declaration that only the second is answered, and no step of the main line depends on how the categories in fact arise
Both must hold; the distinction stated while the main line still appeals to origin does not count.
What the rival reading predicts: if the deduction were an account of origin, the main line should contain genetic statements about how the mind produces these concepts and lean on them
The two readings differ observably across the sections of the main line.
Pre-registered failure condition: a key step of the main line rests on a genetic statement, or the book never draws the distinction
Either counts as the reading being refuted.
推导 · 0 / 3
接口08
Which slot it hangs on
挂在哪个槽位The conclusion you already hold is probably 'once you know how something came about, you understand it' — a judgement about understanding. This chapter goes at the question it leaves out.
Does this chapter (a) replace 'trace how it came about', or (b) constrain it: origin and entitlement must be asked separately?
慢变量Register two slow variables: how many times you recently defended a practice with 'we have always done it this way', and in how many of those you also stated the conditions under which it applies. Look at the second first.
小结09
本章小结
01The deduction's first cut separates the question of fact from the question of right, which is why Hume's conclusion no longer follows directly.
02The pivot is an extremely weak requirement: whatever is mine must be able to carry an 'I think'.
03'The categories hold of all experience' unpacks the range of 'my experience'; it is not a discovery about the world.
04A validity proof is not a completeness proof; the number twelve rests on a table of judgements that is replaceable.
05'Not combined, so not an object of yours' is consistent with any evidence, so this course books it as unadjudicable.
提取练习 · 合上书,先自己答一遍。
?Is 'every manifold that can become my experience has been combined under the categories' (a) premises merged with definitions or (b) a textual claim refutable by the original?
?Do objections to the table of twelve attack (a) the skeleton of the deduction or (b) the completeness of the list?
Forced choice10

B · Identify the tag

'The categories hold objectively of every possible object of experience' — what kind of step is that?
二选一
Forced choice11

C · Locate the crack

In which situation does this chapter's reading give a confident and wrong answer?
二选一
读到这里,把它记为已读。