420 M E M O I K S of fhe 
2, In oral tradition, as a (ingle man is fubjefl to much ca- 
fualtv, Co a company of men cannot be fo eafily fuppofed to join j 
and therefore the credibility of Irf or about H may poflibly be 
judged too high a degree for an oral conveyance, to the diftancc 
of 20 years: But in written tradition, the chances againft the 
truth, or confervation, of a fingle writing are far Jefsf and leveral 
copies may alfo be ealily fuppofed to concur, and thofe fince the 
invention of printing exadly the lame^ as alio feveral diftindl 
fucceffions of luch copies may be as well fuppofed, taken by dif- 
ferent hands, and prelerved in different places or languages ; And 
therefore if oral tradition by any one man or company of men, 
might be fuppofed to be credible after 20 years, at i' of cer- 
tainty, or but f o Of t, a written tradition may be well imagined 
to continue, by the joint copies that may be taken of it for one 
place (like the leveral copies of the fame impreffion) during the 
fpace of icc> if not 200 years ^ and to be then credible at f §f of 
certainty, or at the proportion of 100 to one: And then feeing 
that the fucceffive tranltniffions of this tut of certainty, will not 
diminifli it to a half, untill it palTes the <Je;th hand (for it will be 
near 70 years before the rebate of money, at that intereft will fink 
it to half 3 it is plain that written tradition, if prelerved but by 
a (ingle fuccefiion of copies, will not lofe half of its full certainty, 
until 70 times 100, if not 70 times 200 years are paft^ that is 
7,oco, if not i4,GC0 years: And further, that if it be likewife 
prcferved by concurrent fuccelTions of fuch copies, its credibility 
at that dillance may be even increafed, and grow far more cer- 
tain from the leveral agreeing delivericvS at the end of 70 fuccef- 
fions, than it would be at the very iirft from either of the (ingle 
Lands. 
3. Laflly, in fcating the proportions of credibility for any part 
cr parts of a copy, it may be obferved, that in an original, not 
very long, good odds may be laid, that a copy by a careful hand 
fhall not have fo much as a literal faulty but in one of greater 
length, that there may be greater odds againft any material error, 
and fuch as fhall alter the Icnlc • greater yet, that the fenfe fhall 
not be altered in any conliderable pointy and ftill greater, if 
there be many of thole points, that the error hits not upon fuch 
a (ingle article, as in the third proportion. 
7-d'O 
