Right, this is what I said in my first paragraph: science tends to be empirically conservative. As for metrics of success, we're talking about right or wrong here, not successful or unsuccessful. If the scientific consensus in 1900 was P, whereas the current scientific consensus is ¬P, then it's silly to pretend that the previous theory was merely "incomplete". Take plate tectonics as an example. Either the Earth's surface is split into moving plates or it isn't. As it turns out, this hypothesis is correct, and the scientists who denied it well into the 20th century were just flat our wrong.
And you're treating "right" and "wrong" as absolutes, with no in between, which is exactly what Asimov was saying you shouldn't do.
You could apply the same logic you're using for plate tectonics to the shape of the Earth, which is Asimov's main example. If the Earth is a sphere, then it's not flat. If it's an oblate spheroid, then it's not a sphere. If it's slightly pear-shaped, then it isn't an oblate spheroid. So every previous scientific theory about the Earth's shape was just wrong.
The problem with this, as Asimov points out, is that just putting theories into the buckets "right" or "wrong" doesn't tell you how far off they are, predictively, which is what's important. The Earth is not flat--but it's almost flat. It curves about 8 inches per mile; 8 inches per mile is not zero, but it might be indistinguishable from zero if you don't have very good measuring instruments. (And similarly for the increasingly more accurate measurements that have given us more precise knowledge of the exact irregular shape of the Earth.) So treating the Earth is flat is wrong, but it's not very far wrong. It is, however, more wrong than treating the Earth as a sphere, since the error involved is smaller in the latter case.
Similarly, the continents are not motionless, but they're almost motionless. So treating them as motionless is wrong, but it's not very far wrong. You have to have accurate measurements to distinguish the actual motion of the continents from no motion at all. And you have to have even more accurate measurements and more comprehensive data to know that the cause of the motion is that the continents are sitting on plates that are floating in magma.
> the scientists who denied it well into the 20th century were just flat our wrong.
They were wrong, but they were still a lot less wrong than, for example, the Flat Earth Society. Which is Asimov's point.
I’m not actually assuming that ‘right’ or ‘wrong’ are absolutes. I’m just pointing out that some widely accepted scientific theories have turned out to be very clearly at the ‘wrong’ end of the spectrum. If you don’t like the particular example of plate tectonics then just take one of the other examples that I mentioned. E.g., our conception of the structure of matter has undergone several radical changes in the course of scientific history. These changes were fairly conservative in empirical terms, but in terms of the various conceptions of matter that they put forward, they are in some cases radically different. Another example would be the question of whether the universe has a beginning in time. There really isn't any middle ground there. If it did, then the people who said that didn't were wrong, and if it didn't, then the people who said that it did were wrong.
Of course you can always find a germ of truth in a theory if you try hard enough, just as you can find a germ of truth in the book of Genesis if you're determined to defend it. I don't see the point, though. Why bother constructing a false Whig history of scientific progress when the real history is impressive enough?
> our conception of the structure of matter has undergone several radical changes in the course of scientific history
Really? The original conception (or at least the commonly accepted one--even some ancient Greeks hit on the idea of atoms) was that matter was continuous, and it is pretty close to being continuous. Then we started to make observations that indicated that matter was made of atoms. Then we began to get data that indicated that atoms were made of smaller parts. Where is the radical change?
> Another example would be the question of whether the universe has a beginning in time. There really isn't any middle ground there.
Sure there is. What if the concept of "a beginning in time" isn't well-defined, because "time" is not a fundamental property? It's quite possible that both viewpoints--the universe had a beginning in time, and the universe didn't--are useful approximations to something deeper.
You mentioned one of the radical changes: the change from a continuous model to an atomic one. I mentioned another radical change in my post (Rutherford).
As for the claim that matter is close to being continuous, I am not a physicist, but I was under the impression that this was false. I.e., speaking naively, there is far more "empty space" than there is "stuff". I realize that QM makes it difficult to give precise content to that kind of statement, but something like that appears to be true.
> It's quite possible that both viewpoints--the universe had a beginning in time, and the universe didn't--are useful approximations to something deeper.
I can't really answer speculation. However, your only reason for making this speculation, so far as I can see, is your desire to deny that any scientific theory has ever been flat out wrong.
> As for the claim that matter is close to being continuous, I am not a physicist, but I was under the impression that this was false.
We only know this because we have precise enough measurements to detect atoms and their internal structure. Atoms are about a billion times smaller than ordinary macroscopic objects, so it's a very good approximation to treat a cup of water, for example, as a continuous fluid, rather than a bunch of atoms.
> speaking naively, there is far more "empty space" than there is "stuff".
Only on a very naive definition of "empty space". Pop science books and articles often say that "atoms are mostly empty space", but that's because they are thinking of the electrons as little pointlike objects, which they're not. So "atoms are mostly empty space" is wrong, by the definition you've been trying to use, just like "matter is continuous" is wrong.
And before you object that I'm saying "electrons are little pointlike objects" is flat out wrong, electrons are close to being little pointlike objects; you have to make very precise measurements (like the ones that have mapped out the structure of atomic orbitals) to see the quantum "smearing out" of electrons. But in many other contexts, electrons do behave like little pointlike objects.
> I can't really answer speculation.
It isn't just my speculation; it's a serious theoretical proposal in cosmology. Google "eternal inflation".
> your only reason for making this speculation, so far as I can see, is your desire to deny that any scientific theory has ever been flat out wrong.
Quite the contrary; cosmologists consider such proposals seriously because neither simplistic model--"the universe had a beginning in time" or "the universe did not have a beginning in time" as flat statements with no wiggle room--works well.
The problem you are having here is that you continue to think that we humans can get something absolutely right. But every single one of your "absolutely right" statements turns out to have hidden qualifications when we dig deeper. And your "flat out wrong" statements turn out to be workable approximations in many contexts.
>so it's a very good approximation to treat a cup of water, for example, as a continuous fluid, rather than a bunch of atoms.
It's a good approximation empirically. I've said a few times now that scientific progress is indeed quite conservative in empirical terms. However, we know now that water is very unlike a continuous fluid in terms of its actual constitution.
> So "atoms are mostly empty space" is wrong, by the definition you've been trying to use, just like "matter is continuous" is wrong.
Yes I know, that's why I mentioned that QM makes things more complicated. However, it's still roughly true that atoms are mostly empty space, in the sense that they are very far from being of uniform density.
>Google "eternal inflation".
If you're referring to theories which attempt to extend eternal inflation in to the past, then these are attempted variations on steady-state theories. See e.g. this paper: http://arxiv.org/abs/astro-ph/0111191 . It gives a useful overview of the history. (As to whether its specific result is correct, I make no judgment.)
>The problem you are having here is that you continue to think that we humans can get something absolutely right
I'm not sure where you're getting that from. I suppose that we can get some things absolutely right (e.g. some mathematical truths), but that's not central to anything I've said.
What puzzles me is why you're trawling through all these examples trying to show that scientists have never really been wrong about anything. Is it really your view that there has never been a period in scientific history when the scientific consensus on a given topic was simply mistaken?
But that's the whole point: if a theory is a good approximation empirically, then how can you say it's "flat out wrong"? Doesn't that obfuscate the very important distinction between ideas that are good approximations, but not quite right, vs. ideas that aren't even good approximations? (See below for more on this.)
> Is it really your view that there has never been a period in scientific history when the scientific consensus on a given topic was simply mistaken?
In the sense in which you are using the term "simply mistaken", I think there have been such views, but I think they're a lot rarer than you appear to think. (At least if we restrict ourselves to views that can be taken as "scientific" in some sense.)
An example of a view which I think was indeed simply mistaken is the phlogiston theory of combustion. It was simply mistaken because it made no predictions whatsoever; it was not empirical. And when a better theory of combustion was found, there was no sense in which the phlogiston theory could be viewed as an approximation to it.
>Doesn't that obfuscate the very important distinction between ideas that are good approximations, but not quite right, vs. ideas that aren't even good approximations?
No, it just adds an additional distinction between idea that are good approximations but not quite right and ideas that are good approximations but completely wrong.
There's no contradiction. It makes approximately the right predictions, but it's completely wrong as a statement of how things actually are.
A very clear example of such a hypothesis would be one concerning identity. E.g., who was it who committed the murder? The hypothesis that Joe Smith did it might make lots and lots of correct predictions and yet still be completely wrong (because it was actually someone else).
> It makes approximately the right predictions, but it's completely wrong as a statement of how things actually are.
You're assuming that we know "how things actually are". If we know that, how do we know it? If your answer is "because we have a theory that makes correct predictions", how do you know our best current theory gets everything correct? It has to be literally "everything" in order to justify your statement that we know "how things actually are"; anything less than perfection from our current theories does not justify your claim.
It should be obvious that we do not know that our current theories are perfectly correct. All we really know is that they get more predictions correct than our previous theories. But that still leaves room for our best current theories to make incorrect predictions--we just don't know which ones they are (yet). And if they make any incorrect predictions, then there could be some other theory that makes more correct predictions, but models "how things actually are" very differently from our current theories. After all, that has already happened, when we came up with our best current theories. If it happened once, it can happen again. And if it happens again, then all your claims about "how things actually are" are, by your own definition, wrong.
>You're assuming that we know "how things actually are".
No, I'm only assuming that we sometimes know how things actually aren't. In other words, we can sometimes be pretty sure that a particular theory is wrong even if we have no confidence that our current best theory is correct.
I think you're going to have a very tough time defending this claim. You might have better luck, though, with a claim that ends up doing much the same work for the subject under discussion: right and wrong do exist, and every theory we humans have come up with so far is wrong by that purist standard. But some human theories are less wrong than others, and it's the "degree of wrongness" that matters, practically speaking.
If it's only possible to deny that scientific theories are ever wrong by denying the very distinction between right and wrong, then I think that underlines my point. Freedom from being wrong at the price of never being right doesn't seem like a very good bargain.
As to the question of whether pragmatist theories of truth are viable, that is a separate one, but I tend to think that the standard objections are quite decisive.
