Which forms of categorical syllogisms are valid? Logicians have disputed the answer for centuries, a dispute that can give us insight into the meaning of validity, the central concept of formal logic. This will be the first of a few posts in which I will briefly discuss the history of syllogistic validity.
It all started with Aristotle, who in his Prior Analytics, Book I, chapters 4-7, detailed sixteen valid forms:
Figure 1: AAA, EAE, AII, EIO
Figure 2: EAE, AEE, EIO, AOO
Figure 3: AAI, EAO, IAI, AII, OAO, EIO
Figure 4: EAO, EIO
If you read Prior Analytics (which is no trivial task), Aristotle presents only the first three figures as figures, omitting any mention of a fourth figure. But in chapter 7 he admits in passing the forms of EAO-4 and EIO-4 as valid, saying,
If A belongs to all or some B, and B belongs to no C … it is necessary that C does not belong to some A.
It is not difficult to see why Aristotle omits AAI-1, EAO-1, AEO-2, and EAO-2. These four forms are his AAA-1, EAE-1, AEE-2, and EAE-2 with the subimplication of the conclusion. Aristotle apparently saw no need to include syllogism forms with particular conclusions when the premises could imply the universal.
Aristotle and Figure 4
It is rather more difficult to understand why Aristotle does not admit the fourth figure, though logicians have argued that it has to do with how he defines a syllogism. We learn from Bertrand Russell, in his Cambridge Essays, that
The fourth figure…was added by Aristotle’s pupil Theophrastus and does not occur in Aristotle’s work, although there is evidence that Aristotle knew of fourth-figure syllogisms.
Theophrastus apparently recognized three more valid forms of figure 4: AAI, AEE, and IAI, bringing the total to 19. These were given Latin names by medieval scholars, but that will be the topic for my next post.