>At 04:19 9/6/97 EDT, Paul S. Dixon wrote:
>>It does follow. The parameters or factors of this statistical
>>probability problem are given only as: pre-copulative anarthrous
>>predicate nominatives in the Gospel of John. Now, it was determined
>that 94% of those are qualitative. So, given a pre-copulative anathrous
>>predicate nominative in the Gospel of John we can say the statistical
>>probability that it is anarthrous is quite high, about 94%
>>(technically, we should not include the sample as part of this here;
thus 49 of >>52, but still 94%).
>>
>>Sure, if you consider other factors, then the probability might
>>change. But, this does not affect the validity of my findings.
>
>I do not wish to dispute the correctness of your observations of the
>data; however, the validity of your conclusion is the same as the
argument
>that since most humans in the world have dark hair (e.g. China, India,
>Africa, Latin America), it is therefore likely that Swedes have dark
hair.
No, no. You are mixing apples and oranges here and getting orpples. No
wonder you find my conclusions fruity, and a rather strange fruit at
that. But, they do not mix. Let me explain.
You see, apples ... , I'm sorry, statistical probability is different
from the kind of probability you and Bayes are talking about.
Statistical probability is very well defined and calculated. It deals
with a finite set of possible outcomes, and a finite subset of desireable
outcomes. The statistical probability, then, of the occurrence of a
desireable outcome selected at random is determined by dividing the
finite subset of possible desireable outcomes by the finite set of
possible outcomes. For example, in a deck of cards (52 possible
outcomes), the statistical probability of selecting a heart drawn at
random is determined by dividing 13 (set of possible desireable outcomes)
by 52 (set of possible outcomes). This is the meaning of what I did in my
thesis. The set of possible outcomes was 53 (I threw in an extra card,
kidding), i.e., 53 pre-copulative anarthrous predicate nominatives in
John's Gospel. Of that total, I found that 50 were qualitative; hence
50/53 = 94%. This means if we select any of the 53 at random, then the
statistical probability it would be qualitative is 94%. This follows.
And, if we select QEOS in Jn 1:1c, as a sample, then it follows that the
statistical probability that it is qualitative is 94%.
What you and Bayes are talking about is something altogether different,
for you are looking at or gathering additional information as you go, and
the probability will and does change, as you do so. For example, in Jn
1:1c you might factor in other considerations, such as, the probable
nuance of QEOS when: 1) it occurs with the articular construction in the
same context, 2) double nuancy occurs (which I happen to reject, except
for double-entendres), 3) or any number of other considerations.
But, you are remiss if you think my conclusions are faulty because I have
not factored in these considerations, or, if you think (as your
counter-example seems to suggest) that I have erroneously introduced an
extraneous desireable outcome. I have not.
The only qualification for applicability of my statistics is that a
precopulative anarthrous predicate nominative in John's Gospel be
selected. QEOS in Jn 1:1c is a worthy example.
snip
Paul Dixon