The meaning of evidence

[Sherlock Holmes]

This thread is to discuss the meaning of the term "evidence" particularly with respect to claims made by evolution advocates.

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?

[Bust Nak]

[Replying to Sherlock Holmes in post #1]

This was my objection against your analogy: no matter which twenty pieces you give us, we won't ever see any overlap. Therefore we could not make up a jigsaw where we fill in the blanks and conclude that the whole puzzle has overlapping circles.

I then offered an alternative analogy, supposing that all 20 pieces we have are solid color, which is very possible given several large circles. We could make up a theory where we fill in the blanks and say all the pieces are solid color, color vary only between pieces. We then go on to find more and more solid color pieces, so we tell ourselves our theory is getting stronger and stronger each year as we find more and more "evidence."

Not until much much later, when we finally have a piece that shows multiple color, then we can see the theory was wrong all along.

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?

Yes, I do agree, but this is moot. In the scientific context, there is no room for conflation. When we say evidence we mean evidence, not mere consistence.

[Sherlock Holmes]

[Replying to Sherlock Holmes in post #1]

This was my objection against your analogy: no matter which twenty pieces you give us, we won't ever see any overlap. Therefore we could not make up a jigsaw where we fill in the blanks and conclude that the whole puzzle has overlapping circles.

I then offered an alternative analogy, supposing that all 20 pieces we have are solid color, which is very possible given several large circles. We could make up a theory where we fill in the blanks and say all the pieces are solid color, color vary only between pieces. We then go on to find more and more solid color pieces, so we tell ourselves our theory is getting stronger and stronger each year as we find more and more "evidence."

Not until much much later, when we finally have a piece that shows multiple color, then we can see the theory was wrong all along.

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?

Yes, I do agree, but this is moot. In the scientific context, there is no room for conflation. When we say evidence we mean evidence, not mere consistence.

I stressed in the OP that we are to assume that we already have a reason to believe the puzzle does contain some overlaps, we are expecting that overlaps will be present to some degree if we look for enough pieces.

The fact is that if one expected some overlaps and one never expected overlaps, the few pieces are not enough to decide who's right, each side can argue that the pieces are consistent with either model.

If we have observational data that's highly consistent with two different models, then you tell me - which of the two models are the data evidence of?

[Bust Nak]

I stressed in the OP that we are to assume that we already have a reason to believe the puzzle does contain some overlaps, we are expecting that overlaps will be present to some degree if we look for enough pieces.

Who is this "we" though? In the other thread I pointed out that scientific minded people would just stick to what the pieces shows. If anyone expects there to be overlaps, it would be those who would appeal to external factors such as "I have a note from the puzzle maker that says there are overlaps."

The fact is that if one expected some overlaps and one never expected overlaps, the few pieces are not enough to decide who's right, each side can argue that the pieces are consistent with either model.

They can, but why does that matter when there are better ways to judge who is right beyond mere consistency? The pieces shows no overlaps, who cares that is is consistent with the thesis that there are overlaps, go with the no overlap thesis.

If we have observational data that's highly consistent with two different models, then you tell me - which of the two models are the data evidence of?

Depends on the specifics. Suffice to say we can rank the two models in terms of scientific merit.

[Sherlock Holmes]

I stressed in the OP that we are to assume that we already have a reason to believe the puzzle does contain some overlaps, we are expecting that overlaps will be present to some degree if we look for enough pieces.

Who is this "we" though? In the other thread I pointed out that scientific minded people would just stick to what the pieces shows. If anyone expects there to be overlaps, it would be those who would appeal to external factors such as "I have a note from the puzzle maker that says there are overlaps."

The OP is a hypothetical situation, I wanted people to approach the problem with an existing expectation and to assess the data from that perspective. Of course one can approach it with no expectations too but that wasn't the point of the OP.

The fact is that if one expected some overlaps and one never expected overlaps, the few pieces are not enough to decide who's right, each side can argue that the pieces are consistent with either model.

They can, but why does that matter when there are better ways to judge who is right beyond mere consistency? The pieces shows no overlaps, who cares that is is consistent with the thesis that there are overlaps, go with the no overlap thesis.

There isn't a better way though, that's the point I wanted to come out of this. Other than finding more and more pieces there is nothing more than the few pieces we have. These can be interpreted in either of the two models that one favors.

If we have observational data that's highly consistent with two different models, then you tell me - which of the two models are the data evidence of?

Depends on the specifics. Suffice to say we can rank the two models in terms of scientific merit.

So given the specifics of the jigsaw analogy, which model is best supported by the data (pieces)?

[JoeyKnothead]

I stressed in the OP that we are to assume that we already have a reason to believe the puzzle does contain some overlaps, we are expecting that overlaps will be present to some degree if we look for enough pieces.

I looked at that puzzle expecting to see some overlaps, but didn't find any, even when looking at individual pieces.

The fact is that if one expected some overlaps and one never expected overlaps, the few pieces are not enough to decide who's right, each side can argue that the pieces are consistent with either model.

I challenge anyone to point to just one piece of that puzzle that shows overlapping colors.

Scieniists form hypotheses based on their assumptions, then devise tests to fret em.

When the assumptions're confirmed, their conclusions're up for peer review. Against other scientists who have em their own assumptions.

When the assumptions ain't confirmed, they revise the assumptions, the testing, or both.

Contrast that with creation institutes that demand all conclusions must fit the creationists' assumptions.

If we have observational data that's highly consistent with two different models, then you tell me - which of the two models are the data evidence of?

The creationist model's only consistent in being supported by poor science and faulty analogies.

[Bust Nak]

The OP is a hypothetical situation, I wanted people to approach the problem with an existing expectation and to assess the data from that perspective. Of course one can approach it with no expectations too but that wasn't the point of the OP.

Okay, then surely finding example after example of non-overlapping pieces, that expectation would change to match the data.

There isn't a better way though...

Sure there is. Fitness is a better way to rank the models.

So given the specifics of the jigsaw analogy, which model is best supported by the data (pieces)?

I told you which one - the model with no overlaps, because none of the pieces shows overlaps.

[Sherlock Holmes]

The OP is a hypothetical situation, I wanted people to approach the problem with an existing expectation and to assess the data from that perspective. Of course one can approach it with no expectations too but that wasn't the point of the OP.

Okay, then surely finding example after example of non-overlapping pieces, that expectation would change to match the data.

There isn't a better way though...

Sure there is. Fitness is a better way to rank the models.

So given the specifics of the jigsaw analogy, which model is best supported by the data (pieces)?

I told you which one - the model with no overlaps, because none of the pieces shows overlaps.

Very well and this is the reason I do not regard the fossil record as evidence of evolution, it is consistent with it in that yes we expect to find fossils, but it is not evidence of continuity, the entire fossil record is discontinuous, even the claims of continuity in certain cases are extremely limited and a stretch to say the least.

If we regard the overlaps in the jigsaw analogy with granular continuity in the fossil record, then we reach the same conclusion - there simply is no evidence of continuity and by extension no evidence of "macro" evolution.

Agree with me or not, this is basically how I see the fossil record and the claims that it is evidence of evolution. To me it looks more or less like the claims the twenty jigsaw pieces are evidence of circles that contain overlaps.

The overlap (evolution) devotee could argue that the overlap fragments were just not preserved well, we can't expect every single piece of the jigsaw to be found, that's ridiculous, etc etc etc

[Bust Nak]

Very well and this is the reason I do not regard the fossil record as evidence of evolution, it is consistent with it in that yes we expect to find fossils, but it is not evidence of continuity, the entire fossil record is discontinuous, even the claims of continuity in certain cases are extremely limited and a stretch to say the least.

And that's why my point was, you shouldn't just look at consistency, you should pay more attention to fitness. It's not because of consistency with evolution that we can use the fossil record as evidence of evolution, but because of its excellent fitness to the prediction of evolution.

If we regard the overlaps in the jigsaw analogy with granular continuity in the fossil record...

Okay, but why would you do that when you can instead equate non-overlaps with granular continuity?

[Sherlock Holmes]

Very well and this is the reason I do not regard the fossil record as evidence of evolution, it is consistent with it in that yes we expect to find fossils, but it is not evidence of continuity, the entire fossil record is discontinuous, even the claims of continuity in certain cases are extremely limited and a stretch to say the least.

And that's why my point was, you shouldn't just look at consistency, you should pay more attention to fitness. It's not because of consistency with evolution that we can use the fossil record as evidence of evolution, but because of its excellent fitness to the prediction of evolution.

If we regard the overlaps in the jigsaw analogy with granular continuity in the fossil record...

Okay, but why would you do that when you can instead equate non-overlaps with granular continuity?

Because that wouldn't demonstrate what I wanted the analogy to demonstrate - namely something absent from the data yet the data being consistent with expectations.

It seems you agree anyway, the fossil record alone is not evidence for evolution, only consistent with evolution.

[Bust Nak]

It seems you agree anyway, the fossil record alone is not evidence for evolution, only consistent with evolution.

No, I absolutely do not agree. This is what I stated, the fossil record is evidence for evolution not BECAUSE it is consistent with evolution. I was referencing the reason. The fossil record is evidence for evolution because it is fitness with evolution's prediction.