Prediction

Ray Hendon rayhendon at worldnet.att.net
Fri Apr 16 16:41:11 UTC 1999


[ moderator re-formatted ]

>"Predictability" is not pronostication.  It is the way we test hypothesis.

>Hope this clears this up some.

>Regards,
>Steve Long

    The test of an hypothesis is a test of the reliability of the prediction
the hypothesis makes.  If the predictive power of the hypothesis is good,
i.e., if it is a reliable predictor, predicting again and again the actual
events,  then we generally accept the hypothesis as valid.  If I am seeking
out manuscripts that were written in and around Rome in the year AD 100,
your hypothesis might be that I should expect to find these documents
written in Latin.  If I look at the documents and find they were indeed in
Latin, and if most other scholars finds the same thing, your hypothesis will
generally be accepted as true.  It accurately and reliably accounts for the
facts we find of that period at that place.

To make any assertion of the past is a prediction as to what one would find
if one went there and looked or listened.  I cannot see, with all due
respect,  that an hypothesis is anything else at its core, than a prediction
that two things are related in a cause-effect manner, and that the
relationship is predictable.

I could have used a less perjorative term like "expected," rather than
"predicted, " although they are the same in a statistical sense.  We are not
saying something should happen,  but is likely, or expected, to happen.  But
I intentionally used "predict" so as no to try and cover up this most
fundamental nature of an hypothesis.

    If I had a recording of every word spoken during Cicero's life, and I
counted the words he spoke and classified each word as belonging to certain
linguistic group, I could then compute what percentage of his words were of
one language and what of another.  Using this kind of modal calculation I
might then advance the hypotesis that I would expect Cicero would speak in
Latin, since the overwhelming body of evidence I possess is of his speaking
this tongue.  Although there is a mystical, somewhat unacceptable
connotation to the words"predict" or "forecast,"   mathematical modeling and
non-mathematical modeling still have a predictive quality that I do not see
can be avoided.

I don't mean this in an argumentative way. I may be just dense on the matter
and fail to see the obvious, but fail to see it ,I do.  The word "expected"
demystifies the statement of predictibility, and puts it in on more or less
a scientific footing, but an hypothesis is still a prognostication.

Sincerely,
Ray Hendon



More information about the Indo-european mailing list