Date: Tue, 23 Jun 1998 01:44:56 -0400 From: Mike Grondin To: crosstalk@info.harpercollins.com Subject: Re: The Thomas/Q Hypothesis I hope that it doesn't lessen the value of the discussion under this thread on the relative priority of Thomas vis-a-vis Matt & Luke to say that it probably proceeded from a misunderstanding. In my earlier attempt to improve the wording of the "Thomas/Q Hypothesis", it was certainly not my intention to rule out in advance either that (a) Thomas preceded Matt & Luke, or that (b) Matt & Luke preceded Thomas. What I was trying to do was in fact the exact opposite - to remove from the hypothesis any suggestion whatsoever of either (a) or (b). My concern was with fitting the hypothesis as closely as possible to the evidence, not with trying to sneak my own views in through the back door, so to speak. What was the source of the misunderstanding? I had proposed as an adequate statement of the hypothesis the following: > Those sayings in Thomas that correspond to Q are usually more like > the Lukan form than the Matthean form (where this is understood > not to imply that Luke & Matthew preceded Thomas). At this point, I seem to have made a small mistake in presentation. The "where" clause probably shouldn't have been presented as part of the hypothesis proper. I agree with Mark's intuition that it's simply unnecessary. The hypothesis proper already "does not imply ...". But Mark seems to have misunderstood the wording of the "where" clause, as in his comment: > The possibility that Luke and Matthew precede Thomas is, of course, > a quite reputable position and one that we should not write out of > the equation before doing this kind of primary research. The "where" clause does not in fact rule out this position. The cause of the misunderstanding may lie in confusing (1) "this does not imply that A preceded B" with (2) "this implies that A did not precede B". In any case, I hope that Mark will believe that I did not intend to rule out any position whatsoever on the chronological priority issue, nor did I take the "where" clause as so doing. But now I want to address Mark's comments about scientific hypotheses in general, and this one in particular. First, Mark's position: > Even without that qualifying statement, I still prefer the original > statement, which seems to me more red-blooded and more amenable to > testing. To make clearer what Mark is saying here, compare the two alternatives: > (H1) Luke's form of Q sayings is generally supported by Thomas > against Matthew. > (H2) Those sayings in Thomas that correspond to Q are usually more > like the Lukan form than the Matthean form. What Mark says about H1 is this: > This is a useful, red-blooded statement, which makes a forceful > claim. As such it is useful, I would say, for testing. Now I certainly have no objection to verbal robustness, in its place. But is a hypothesis the proper place for it? Mark suggests that it is, because, in his view, there is some connection between robustness and testing. Although it isn't clear to me what that connection might be, perhaps it may be useful to contrast my own views of what constitutes an adequate scientific hypothesis. It seems to me that the primary concern in stating a scientific hypothesis is to see to it that the hypothesis is neither "under-determined" nor "over-determined" by the evidence. That is to say, the hypothesis should neither "over-reach" nor "under-reach" the evidence. Which is to say that the primary criterion of the adequacy of a hypothesis is the extent to which it "fits" the evidence - not any intrinsic quality which it may possess in itself. The "evidence" in this case is a set of judgments based on specific wording laid out by Steve Davies, viz: > There can be four possible conclusions for each example: > 1. This one is a spurious example (not-Q or not related to Thomas). > 2. Matthew's version is more like Thomas than Luke's. > 3. Luke's is more like Thomas than Matthew's. > 4. There are elements more like Thomas in both to a roughly > equal degree... although even so one might be able to say > [that] Thomas is more like Mt or Lk than the contrary. Statements 1 & 4 can be safely ignored; the hypothesis must capture the difference between 2 & 3. But the exact wording is important! Statements 2 & 3 use the phrase "more like" (which occurs in H2), not the phrase "is supported by" (which occurs in H1). What's the difference? Among other things, by not using the same basic wording in the hypothesis that was used in the above statements to gather evidence, there is the real danger that, if statements 2 & 3 had been worded otherwise, we might have gotten different results. Which means that H1 is a poor fit for our evidence. But what of the claim that the more robust the hypothesis, the more amenable to testing? Whether or not this claim is true in general seems to me to be rather beside the point. The question at this point is *not* "Which hypothesis is more robust?", but rather "Which hypothesis best fits the evidence?" And again, the question at this point is *not* "Which hypothesis is more amenable to testing?", but rather "Which hypothesis best fits what was in fact tested?" And it seems to me that H2 is to be preferred over H1 on these grounds. Putting it a slightly different way: We are not now in a position of choosing between alternative hypotheses prior to conducting an investigation (which is the point at which Mark's criteria would be applicable). The investigations have already proceeded along certain lines, and we now need to see to it (not having done so originally) that the hypothesis adequately conveys all the evidence and nothing but. Mike G. ____________________________________ The Codex II Student Resource Center http://www.Geocities.com/Athens/9068