.comment-link {margin-left:.6em;}

Prior Knowledge

Wednesday, August 17, 2005

Kant's analytic/synthetic distinction

Kant's analytic/synthetic distinction differs from the common distinction that passes under the same name. Importantly, it seems that Kant's distinction survives the attacks levelled by Quine at the analytic/synthetic distinction in his 'Two Dogmas' paper. (I here follow Richard Smyth's explanation of Kant's distinction in Smyth's book 'Forms of Intuition'. Smyth does not compare the distinction to the contemporary one and neither does he relate Quine to Kant.)

The common use of the terms 'analytic' and 'synthetic' is to refer to a distinction between types of truths. Analytic truths are logical truths and may also include truths of meaning or of convention (though Quine disputes this latter inclusion). Synthetic truths are factual truths.

For Kant, the distinction is between types of justification rather than types of truth. He is concerned with our epistemic right to a given judgment or belief. We have an epistemic right to a judgment if, given the best existing evidence, we cannot be faulted on epistemic grounds for holding that judgment to be true. Epistemic right is an objective rather than a subjective notion in the sense that we are concerned not with the evidence that a particular individual happens to possess, but by the best possible grounds that can be produced from existing evidence.

Consider the class of judgments to which we have a right given available evidence. Within this class of judgments we can distinguish different ways in which the right is secured. The a priori/a posteriori and analytic/synthetic distinctions capture these differences.

Some judgments are such that their justification requires an appeal to experience. An obvious example is - 'The cup in front of me is red'. These are a posteriori judgments. All a posteriori judgments are also asynthetic. All judgments in the class under consideration that are not a posteriori are a priori.

A priori judgments can be split into analytic and synthetic ones. Analytic judgments are justified solely by virtue of logical relations among other judgments that do not themselves need justification from experience.

The analytic judgment 'All bachelors are unmarried' is justified by merely logical relations among a further set of judgments that includes the definition of 'bachelor'. At this step the following Quinean objection can be raised. Quine points out that all statements, even definitions, are made true by the world. The definition of 'bachelor' is made true by facts about our liguistic community and therefore essentially involves the empirical world just as the truth of 'the cup is red' is essentially involving of the world.

But Quine's point that there is no non-natural truth-maker and that the only truth-maker is the empirical world is irrelevant to Kant's distinction. For, Kant's distinction is between types of justification and not types of truth. The fact that analytic judgments may appeal to meanings that originate in the world does not mean that we must appeal to some particular set of experiences in justifying them. The question of origin is separated from that of justification. Though both the 'bachelor' statement and the 'cup' statement are made true by the world, there is a crucial difference in how we justify the two. In particular, justifying the latter- but not the former - requires an appeal to experience.

Subtracting the class of analytic judgments from the class of a priori ones leaves the synthetic a priori ones. What justifies synthetic a priori judgments? Kant suggests that they are justified by the presuppositions of all knowledge. These include Kant's 'forms of intuition'. Unlike analytic judgments, synthetic a priori ones are non-trivial and informative. Since they rely on the presuppositions of all knowledge, they can inform us a little about what shape our knowledge of the world will take.

Links to this post:

Create a Link


  • Nice one.

    But, don't we need to know what things mean before we are justified in believing p?

    Surely we cannot be justified in believing something that we cannot understand. Therefore, we are required to understand the terms of the sentence before we can understand the proposition. If understanding definitions of terms (x) are a necessary component for acquiring justification for p, then Quine's argument still holds.

    We will always require the linguistic framework in order to believe a proposition.

    8/18/2005 11:17:00 AM  

  • I still find it intuitively plausible to suppose that we can distinguish the justification from needing to know the meaning of words used in the justification. Surely we can distinguish the philosophical or scientific justification (an epistemic notion) from all that is needed in the process of justifying (which would include concerns irrelevant to epistemology). The process includes not just needing to know the meanings of words but also eating enough to ensure that I don't starve before I can carry out the justifying research. It even seems in some sense necessary for my justifying act that the earth not be so close to the sun that all life on earth be extinguished. Yet it seems a stretch to count these as relevant to the justification of some proposition.

    8/25/2005 03:11:00 PM  

  • It seems to me T that you are using a slight variation on 'justified'. The reply to Quine proposed by sagars has no reference to justifing our belief of a statement without knowing some things e.g. words etc , that would be a posteriori, rather it is justification in light of what we know.

    I will explain myself further.

    What makes 'all bachelors are unmarried' justifiable a priori (and Analytic) is not the syntax or the meanings of the words, it is what the subject of 'bachelor' is and what the predicate 'unmarried' is.

    Unmarried is contained within bachelor. If I told this statement to a child that did not understand what bachelor meant, it would not take away from it being justified a priori , the child would jjust have no idea what I was saying.

    Yes for you to personally have justification for this statement you must have prior understanding of the english language, but your understanding does not affect the justification of the statement in itself.

    Another point to raise is the consequences of your statement. Your statement, if true, would eliminate all a priori knowledge.

    You may say that this is fine, we cannot know things untill we investigate them or experience them, but then you are saying that 3+2=5 only after we experience it.

    A linguistic framework is only necessary for your own justification of a , not the justification of the statement itself, in itself.

    5/23/2006 02:26:00 PM  

Add a comment

<< Home