THEODORUS OF CYRENE
205
of proof can be adapted to the cases of U3, V5, &c., if 3, 5 ...
are substituted for 2 in the proof; e.g. we can prove that,
if \/3 is commensurable with 1, then the same number will
be both divisible and not divisible by 3. One suggestion,
therefore, is that Theodorus may have applied this method
to all the cases from V3 to V17. We can put the proof
quite generally thus. Suppose that N is a non-square number
such as 3,5..., and, if possible, let VN = m/n, where m, n
are integers prime to one another.
Therefore m 2 = N .n 2 ;
therefore m 2 is divisible by N, so that m also is a multiple
of N.
Let m — jx.N, (1)
and consequently n 2 — N. fi 2 .
Then in the same way we can prove that n is a multiple
of N.
Let n — v. N (2)
It follows from (1) and (2) that m/n — ¡x/v, where ¡i < m
and v < n; therefore m/n is not in its lowest terms, which
is contrary to the hypothesis.
The objection to this conjecture as to the nature of
Theodorus’s proof is that it is so easy an adaptation of the
traditional proof regarding' V2 that it would hardly be
important enough to mention as a new discovery. Also it
would be quite unnecessary to repeat the proof for every
case up to V17; for it would be clear, long before V17 was
reached, that it is generally applicable. The latter objection
seems to me to have force. The former objection may or may
not; for I do not feel sure that Plato is necessarily attributing
any important new discovery to Theodorus. t The object of
the whole context is to show that a definition by mere
enumeration is no definition; e.g. it is no definition of em-
arijfjiT] to enumerate particular kTnarrj[jLaL (as shoemaking,
carpentering, and the like); this is to put the cart before the
horse, the general definition of eiriarri^-q being logically prior.
Hence it was probably Theaetetus’s generalization of the
procedure of Theodorus which impressed Plato ag bein$
original and important rather than Theodorus’s proofs them
selves.