What to count (was: Hair-splitting...)

Paul S. Dixon dixonps at juno.com
Thu Jul 15 11:22:49 EDT 1999



On Thu, 15 Jul 1999 06:55:22 -0400 "Carl W. Conrad"
<cwconrad at artsci.wustl.edu> writes:

>At 12:10 PM -0400 7/14/99, Paul S. Dixon wrote:
>>I appreciate what you say.  If we follow the scientific approach
carefully,
>>then we will need to be aware of such considerations before drawing
>>conclusions based on a numbers count alone.  Would you not agree,
>>however, that in some cases a numbers count alone could yield
>>significant and telling information?   If we are trying to determine,
>>for example, whether the aorist tense in the indicative mood is
>>increasingly assuming the functions of the perfect and pluperfect
>>in the same mood, shouldn't we expect to see this reflected in
>>the data?  If not, then could you be a bit more specific as to why
>>not in this particular situation?  Are you suggesting the change
>>in meanings of verbs could skew the data?  If so, how?
>>
>>To put it another way, if the data does not suggest this hypothesis,
>>then what possible basis would we have for suspecting it?
>
>With all due respect, Paul, you seem to be beating a dead horse on this
>issue, repeatedly insisting that somehow a question like this OUGHT to 
>be able to be resolved by perusal of raw statistics alone in the face of
>assertions by Daniel and myself that the data for which the statistics
have
>been gathered need to be analyzed before the conclusions can be drawn. 

Carl, if you read me more carefully, you will discover that I have 
not insisted that this question ought to be able to be resolved by
a perusal of raw statistics.  In my above paragraph, for example,
I asked whether a numbers count alone COULD yield significant
and telling information.

I am only cautioning us against creating a false dichotomy: either
always drawing conclusions on the basis of raw data alone, or never
doing so.  Just because we should not always do it, does not mean
we can never do it, nor that we never should.  I sensed this was
the reason being given as to why a raw data count in this instance
wouldn't do.  It just doesn't cut it.  But, you amplify below, so
I cheerfully progress.

>So let me reiterate a couple of the points that I've tried to state
>previously, pointing out why ignoring these points might lead to false
>conclusions from the raw data.
>
>(1) I think that the data for perfect and pluperfect need to be 
>examined separately in comparison with data for the aorist, and I think
also 
>that it is important to get data for instances of each particular verb
used in
>these three tenses; to the extent that the same verb is found in both
the
>aorist and pluperfect tenses where we can show that the aorist form is
>being understood as indicating time anterior to that of the main verb, 
>we cannot make a simple assertion that the aorist has replaced the 
>pluperfect:
>we must say rather that the language is in flux and that many writers
>prefer the aorist for indicating anterior time while some continue to
use
>the pluperfect in similar instances. And I frankly think--but have to
wait
>for analysis of significant data to be sure--that this is the case:
namely,
>that the pluperfect is obsolescent in the period of the GNT but not 
>yet obsolete.

But, if this is so, unless an opposite trend simultaneously exists
among some verbs or authors, we should expect to find some 
statistical reflection within the raw data itself.  Even if it is so only
among certain authors or verbs, why wouldn't the over all stats
reflect it?  If A=B and C>D, then (A+C)>(B+D).

>
>(2) When I offered numbers for pluperfects in the GNT texts by groups 
>I tried to indicate the number of instances in each group of such verbs 
>as OIDA/HiDEIN, hESTHKA/hESTHKEIN, EIWQA/EIWQEIN for precisely the
>reason that although these verbs have perfect/pluperfect morphological
forms, >they are semantically equivalent to present and imperfect
forms--their >MEANINGS refer to present and continuing past time. For
that reason, >instances of
>those verbs in the perfect and pluperfect have to be discounted when
>reading the statistics and we need to pay more attention to both the
>numbers and the kinds of verbs that one can and does find in BOTH the
>aorist AND in the perfect/pluperfect morphology.

Yes, these verbs have perfect/pluperfect morphological forms and 
are semantically equivalent to present and imperfect forms, but
exactly how would their inclusion in the data skew anything?
Their forms don't change, do they?  If not, then A=B, and we
aren't really concerned about that anyhow.  It is only the C and
D part of the data, where the perfect/pluperfect form is changing
to the aorist, that we are concerned about.  If so, then their 
inclusion would only make the outcome less noticeable, but
it should still be noticeable.  Again, if A=B and C>D, then 
(A+B)>(C+D).

>
>The bottom line is that raw statistical data are not going to yield
>meaningful conclusions unless what's being measured is homogeneous 
>units;
>where what's being measured is NOT homogeneous, distinctions of kind 
>are going to have a bearing upon how the statistics are read. I don't 
>think there's anything very profound about this, but it is what people
who 
>don't think statistics "lie" or can be misused fail to realize.

Yes, sometimes raw data statistics are mis-interpreted; but, just
because someone appeals to the raw data does not mean he
is necessarily mis-interpreting.

The horse is not quite dead yet.  A few more whacks ought to do it.

Paul Dixon

___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.



More information about the B-Greek mailing list