Becky Tun
07/02/08
Is the distinction between analytic and synthetic truths an unsustainable dogma?
(question from Logic exam paper May 1998)
A brief history of the distinction between the analytic and the synthetic
The distinction between truths that are analytic and truths that are synthetic goes back through many traditions, starting in ideas such as Hume's distinction between relations of ideas and matters of fact, and Leibniz's distinction between truths of reason and truths of fact. Kant defines analytic statements as those which attribute no more to the subject than is contained in the predicate, and whose denials are selfcontradictory, for instance, according to Kant, 'all bodies are extended'. Something synthetic is any assertion that does not have this quality, for instance 'all bodies are heavy'.
In Kant's definition we have two criteria, one psychological (concerning what we perceive to be included in the meaning of an expression) and one logical (concerning the notion of self-contradiction). Whether these two criteria are equivalent, Kant evidently thought that they were, as he says for instance that in order to see that a statement is analytically true 'I have only to extract from [the subject], in accordance with the principle of contradiction, the required predicate.' One of the drawbacks of Kant's definition is that it limits analyticity to propositions that can be expressed in subject-predicate form, which rules out, for instance, conditional statements. Additionally, the notion of self-contradictoriness does not seem to offer much clarification, since it is not clear how it could be defined independently of the idea of analyticity.
Definitions since Kant have centred the distinction around meaning and logic, rather than any psychological criterion. This is reflected in the fact that while Kant talked of 'judgements', modern accounts have preferred to talk of 'sentences' and 'statements'.
Becky Tun
07/02/08
Ayer defines analytic statements as those that are true by definition, as opposed to true by facts of experience. The most recent commonly accepted definition is that of Frege: an analytic statement is either a truth of formal logic or such that it can be turned into one by the substitution of synonyms. Philosophers also sometimes talk of 'true in virtue of meaning alone'.
Analyticity's relationship with the necessary and the a priori
During its history the analytic/synthetic distinction has been closely involved with two other important distinctions: the necessary vs. the contingent, and the a priori vs. the a posteriori. Roughly, if something is necessary, then it could not be /could not have been otherwise, and if something is contingent, then it could be / could have been otherwise. Equally roughly, if something is knowable a priori then you could work it out from your armchair, whereas if it is knowable a posteriori then you would have to have a particular experience or set of experiences in order to come to know it. In the past, the three pairs of terms have been used somewhat interchangeably, because they were thought to carve out the same territory. But since then they have become more and more differentiated as philosophers have edited the boundaries of each distinction.
For Kant, a priori truths were ones that could be known absolutely independently of any particular kind of experience. Our apprehension of the a priori was a mode of knowledge, concerned with a 'realm beyond the senses', where experience could not give guidance or correction. Kant rearranged the boundaries of the three distinctions by claiming to have discovered judgements that were a priori but synthetic. Here is a brief synopsis of Kant's picture of the situation, followed by Ayer's reaction to it, which will set the scene for a discussion of whether the analytic/synthetic distinction (and thereby the related distinctions) is one that actually works.
Buy the full version of these notes or essay plans and more in our Logic and Language Notes.