In Response To: Dave G
On: Hypotheses vs Guess
Couldn't we use some agreed nomenclature in this area?
DAVE: Also, if the only thing we had to go on was the statistical results,
then I would agree in calling what I proposed a "guess". Also, I partly
agree that the data we have here is so thin that anything we say is often
little more than a guess. However, I think I have enough support in this
case to at least call it a 'hypothesis'.
BRUCE: I think a little better than Dave here seems to of statistics; a good
statistical result, at least in some cases, comes complete with a label as
to its probable accuracy or range. Very grownup. But letting that go, I
approve his hierarchy of certainty labels. In general. I would like to
mention a few more, from what I understand of practice in mathematics. My
concern is that when careful people use careful terms, other less careful
people may then be inclined to dismiss the result as "just an opinion." In
the reputable cases, it is much more than opinion. My picture would go like
1. There are no certainties in history, and no deductive proof in history.
Let's get that straight at the outset. So all statements, even ones well
based in terms of the presently available evidence, are inferences from
evidence, and any inference is always open to a better statement, or to a
different statement based on better evidence. So an individual's failure to
exude confidence may be a mark of discretion, and not a badge of
foolishness. I recommend discretion. Though it does lead to longer
2. There are degrees of reasonableness in different ways of reading the
evidence, and it is not so much the job of history to pick out the right
answer and discard the others, as to choose between more and less adequate
treatments of evidence, or arguments from evidence. The ability to make that
distinction is an essential part of technique in the art of doing history.
As also in astrophysics.
3. A "hypothesis," as I usually encounter it, is a preferred working
conclusion. It is still being subjected to evidential testing, but has
emerged from previous testing as apparently superior to other alternatives.
(Sometimes one can entertain two hypotheses at once; statistics will also
handle that situation, and can help to determine which one of them works
better with the known evidence).
4. The hierarchy terms as I understand and/or recommend them are then:
(a). conclusion. a decision between hypotheses. End of thinking for the time
being. In math, this would be a proved conjecture, sometimes called a
(b). hypothesis. a preferred inference still being tested for its fit with
(c). conjecture. a reasoned guess from evidence, but without full
confirmation from evidence.
(d). speculation. a guess without evidential basis; impressionistic or
(e). refusal to guess. no evidence at all; question undecidable and even
Dave seems to use a two-grade system, where there is nothing between
"hypothesis" (my b) and "guess" (d). This is merely to suggest that there
may be intermediate levels. And more generally that in this area, humanities
folks might with advantage pick up a vocabulary hint from their colleagues
at the numerate end of the park bench.
But there is more than one hint available. Does anybody have a different
hierarchy of labels for more and less certain results, that works for them?
E Bruce Brooks
Warring States Project
University of Massachusetts at Amherst.