The reason I started this thread is that I often see - what I regard as - a conflation of consistent with and evidence for. If we are to make reasonable inferences and maintain objectivity and avoid making assumption unwittingly then the more precisely we define "evidence" the better I think.
The biggest risk here is to imply that some observation P is evidence for X and only X, rather than evidence for X and Y or Z. Unless we are on our guard we can informally exclude reasonable possibilities Y and Z and so on. Now the observation P might well be evidence for X and only X, but unless that is soundly established we simply can't assume that.
If we mistakenly regard P as evidence for X and only X then we fall into the trap of believing that P can only be observed if X was the cause.
This is exemplified by an analogy I recently put together that I think warrants its own thread, so here it is:
Consider this jigsaw
None of the circles overlap, we can see this when we can see the totality of the jigsaw. But if we already believed for some reason or other, that they must overlap and we only had twenty random pieces and never see the rest, we could make up a jigsaw (theory) where we "fill in the blanks" so to speak and "show" that we sometimes have overlapping circles.
We'd be absolutely right too in saying the twenty pieces were consistent with an image that has overlapping circles, but we'd be dead wrong to say the twenty pieces are evidence of overlapping circles, because as we know, none of the circles actually do overlap.
So do you agree or not, there's a difference between observations that are evidence for some hypothesis vs consistent with some hypothesis and we should always be careful and make this distinction clear in our arguments?
