A perennial theme of my philosophical peregrinations is the difference between (and relation between) science and metaphysics. This bears directly on the arguments made by creationists and design proponents.
Design proponents often try to distinguish themselves from both creationists and Darwinists by arguing that they alone are faithful to empiricism — “following the evidence wherever it leads” — whereas both creationists and Darwinists interpret the evidence through the lens of some a priori conceptual framework, a metaphysics. (I take it to be false, and importantly false, that one can only hold metaphysics in a dogmatic fashion, and that empiricism is the enemy of metaphysics — though of course empiricism is the enemy of dogmatism, if one’s empiricism does not itself become dogmatic.)
Crucial to the design proponent move here — and one can see something comparable in creationism — is the following line of thought (call the “Teach the Controversy” argument):
(1) good scientific methodology requires that we distinguish between observable entities and posited entities;
(2) If the same observations can be explained by two different sets of posited entities, observation cannot tell which ones to prefer;
(3) so all posits are equally ‘metaphysical’ or ‘speculative’;
(4) and the liberal state should be neutral with regard to metaphysics (subject to certain caveats);
(5) so it is an unjust exercise of state power (and contrary to the tenets of political liberalism) for public schools (including state-supported universities) to refuse to allow a public space for different metaphysical doctrines.
Now, I accept (4), and I also accept (1) and (2). So why don’t I accept (5)? It’s because I deny (3) — that all posits are equally metaphysical or speculative.
The line from science to metaphysics is not crossed when one goes from observables to posits — and it is a deep flaw of any empiricism which holds this. (Holding this view leads to an instrumentalist philosophy of science, which makes scientific progress a mystery — as we’ve discussed here elsewhere — and it also leads to phenomenalism and, if left unchecked, to either Berkleyian idealism or Humean agnosticism.)
For it is true that introducing posits is a central and ineliminable feature of scientific reasoning — what Peirce nicely called “the abductive leap”. But merely introducing posits is, while necessary, also insufficient for scientific purposes. We need also some way of testing the positing, wherein we say, “we observe x, y, z, and we posit that there is some unobserved entity A which is causing x, y, and z. But if A were the case, then we would also expect to observe t, u, and v. So let us look for them!” If we do observe t, u, and v, then that bolsters our confidence in the existence of A; if not, it weakens that confidence.
So it is not the case that all posits are equal — some can be tested, and some can’t be, and some posits fare better on some tests than on others, and so on. And this means that the real distinction between science and metaphysics isn’t based on what is observed vs what is posited, but between testable posits and untestable posits. And if this is correct, then the empiricist epistemology taken for granted by creationists and design proponents is fundamentally flawed.
[nods] Yep, sounds about right to me. The way I’ve always expressed it is, Okay, it’s all down to presuppositions. Fine. Are all presuppositions equally valid? Is there any way to decide whether or not Presupposition X is more valid than Presupposition Y?
http://en.wikipedia.org/wiki/Omphalos_hypothesis
My main issue is creationists like Sal try to morph evidence and interpretation to agree with a worldview, rather than forming a worldview through evidence and interpretation.
So you’re proposing an extension to Rumsfeldism then? Knowable unknowns and unknowable unknowns?
As I see it, it’s got to be a feedback-loop (what philosophers call “dialectical relation”) between worldview and evidence — we bring a worldview to the evidence and re-evaluate worldview in light of evidence. In my preferred lingo:
rationalism: worldview without evidence
empiricism: evidence without worldview
pragmatism: feedback-loop between evidence and worldview.
Nice! That quip by Rumsfeld always reminds me of Zizek’s rejoinder, that psychoanalysis concerns the unknowns knowns — the things that you know, but don’t know that you know.
But actually, I was after something different from either Rumsfeld or Zizek — I was just trying to distinguish between testable posits and untestable posits. Despite my occasional flirtation with positivism, I actually have tremendous respect for metaphysics. (Lately I’ve gotten very interested in speculative realism, which I think can be brought into productive conversation with pragmatism, even though pragmatists often present themselves as hostile towards metaphysics.)
It’s just that the untestable posits of metaphysics, whatever might be said in their defense — and I think it’s a great deal! — still won’t amount to the kind of testable posits that play an ineliminable role in causal explanation, and that the basic epistemological ‘sin’ of creationism (and ID) is to conflate these two different kinds of posits.
KN:
How do you know if the tested posit is always true ? What time frame would you accord to the proof ? Earth was believed to be flat and it was tested to be flat, Earth was tested to be the center of solar system. More recently, Blackhole was believed to exist and proved to exist. Now Hawkings and Lawrence claim there is no blackhole. In science there is no permanent tested posit. New evidence changes the theory. So all posits – whether tested or not remains speculation for ever.
I’m not KN, but I’ll give my answers to some of your points. There’s no guarantee that KN will agree.
Often, we do not have any account of truth, such as would allow us to determine whether a posit is true. This is why I prefer to say that scientific theories are neither true nor false.
I’m not at all sure that is correct. But then I’m not an historian.
Science is not about truth, it is about method. It was assumed that the earth was flat, as part of measuring methodology. And, as long as the method worked, that was taken to support a flat earth. (Note that this is hypothetical, because I am not an historian).
As measurement methods were refined, discrepancies showed up, leading to the view that the earth was something like a sphere. The greek geometers were already measuring the diameter.
In any case, it is about method, and we judge science by how well its methods work. I take that testing of methods to be the kind of testing that KN is talking about, though I could be wrong about that.
Again, I’m not sure it’s right to say that was ever tested. Rather, the assumption of geocentrism was used as the basis for measuring methods, and those methods worked.
Incidentally, if you ever get a speeding ticket for driving too fast, the ticket is going to list your driving speed in geocentric terms. We still rely heavily on that system for terrestrial measurement. But, as astronomy improved, the heliocentric system allowed us to make better sense of the solar system.
As we find better methods, we adopt them. The old methods still work, but the new ones work even better. So we change our posits to suit what works best.
Compare with ID. For ID, there is no methodology that is in regular use for solving important problems.
I am not sure if that line of argument works to refute ID arguments
If I understand you correctly, examples of such unobserved posits woudl be:
– for ID, the designer and the mechanism used to implement the design.
– for evolution, unobserved transitional forms and the mechanisms of macro evolution
Now ID proponents can argue that such a designer is the better inference to the best explanation. They do so both by positive arguments for the designer, eg CSI and the “non-existence” of junk DNA, and by negative arguments against the macro evolutionary mechanism.
So I think that means any refutation of ID must involve arguments about specifics of science and specifics of philosophy of science: is CSI meaningful? is there junk DNA? what is a good explanation in science? what constitutes mechanism in a scientific explanation?
And of course one sees such arguments in this forum and many others.
What I find amazing is the energy of a small group of people here and throughout the blogosphere on both sides who seem willing to go through the details of these arguments repeatedly with no hope of convincing the person they are arguing with.
Why do these people do that? That would be an interesting topic for a sociologist of science and society.
Although there is lots to criticize in Rumsfeld’s legacy, I always thought he got a bum rap in the reporting of his use of this phrase.
“Unknown unknowns” and “known unknowns” are both standard terms used by project management methodology for risk management, and Rumsfeld was just referring to that standard usage.
I might be out on a limb here, but I think the difference between your “observables and posits”, which correspond, I think, to my “data and models” isn’t one of category but of level-of-analysis.
I hold that all we ever have are models – including our observations – but that what we are literally “given” (“data”) at one level analysis, are models at a lower level, and that the models we make of those data become data at a higher level.
Take fMRI brain imaging (my field!)
For the physicists who design the scan sequences, they model their data, which are measurements of changing magnetic flux, by means of a fourier transform, into a measure of signal intensity from a particular source in the brain. But what I am, literally “given” is a 4D set of image intensities (3 spatial dimensions, one time dimension). It’s their model, but my data. And those data become my dependent variables in my model, in which I attempt to predict signal change by knowing what the person in the scanner was seeing, hearing, or doing (e.g. pressing a button, correctly or incorrectly) at the time. And if my model fits my data, I can publish a brain imaging, showing which parts of the brain show activity predicted by my model.
Someone then puts all those models into a meta-analysis. Mine, and other people’s models, now becomes someone else’s data. And so on.
Actually I’ve omitted a few nestings there, but I hope my point is clear. We do not have direct access (IMO) to reality – all we have is our observations which are themselves models that we confirm or inform by making prediction. This is how perception works, and it is also how measurement instruments work. All measurements are in some sense proxy measurements of Reality.
But because our models are highly predictive (or at least if they are) then we can infer that we’ve got some kind of handle on what is Really Going On.
That’s not accurate. You should read the article. What they say is “there are no black holes—in the sense of regimes from which light can’t escape to infinity. There are, however, apparent horizons which persist for a period of time.”
The discussion is about the properties of black holes, not about their existence. The actual objects under discussion have been observed, unlike any deity.
That is exactly why science is an infinitely better “way of knowing” than theology. It is also why intelligent design creationism isn’t science.
That is a poor concept of science, or if you want a very limited epistemology.
What we do and can reproduce in the lab is the reality. In the lab conditions reality works in that way and we can expect that we can make it work in the same way in similar conditions. We can make new hypothesis, that you call “predictions” and test them in the lab.
This statement is in contrast with your premises. If you only have models what you are only constructing is more coherents models but that do not allow you say nothing about reality, you have only models.
Yes, that’s what I said, Blas: we only have models.
And if those models are highly predictive, they are very useful. And the fact that highly predictive models are possible suggests an underlying Reality – that we are not brains-in-vats.
No your models arn´t highly predictive. They are only coherents. When your model fails a prediction you adjust them not to increase predictability but only to make it coherent.
Coldcoffee gives us a nice example of the kind of error I had in mind. He/she writes,
The Earth was believed to be flat, until it was measured by Eratosthenes and found to be a sphere. The Ptolemaic model of the solar system yielded quite usable calculations until the Copernican model was devised and found to yield better predictions. (Importantly, the Copernican model didn’t quite work until Kepler introduced elliptical orbits, and of course it took Galileo to show that the Copernican model was actually true of the solar system and not just a better instrument for making predictions.)
And in the 20th century, the Newtonian model was subjected to two major challenges — general relativity and quantum mechanics — which are, unfortunately, incompatible with each other so far as we know at present.
One cannot know that any particular posit will always be accepted, but that does not make the posit mere “speculation”. Not all posits are epistemically equal, because some posits are confirmed by testing, some posits are disconfirmed by testing, and some posits are untestable. The last category — the untestable posits — are the ones that are “speculative” in my sense.
I take some slight issue with this; I would say that one model yields better predictions than competing models because it is a more accurate map of reality. I don’t see this “because” as crossing the line from science to metaphysics — contra constructive empiricists — because I see it as explicating the metaphysics that science needs in order to make sense at all.
So there is a real distinction I would insist upon between “scientific metaphysics” — the metaphysics that science needs to make sense at all — and genuinely “speculative metaphysics” — the metaphysics that tells us whether there is anything in reality that corresponds to our values, ideals, and hopes. I don’t regard the latter as useless or dispensable — it answers to a deep human intellectual and spiritual need — but confusing it with scientific metaphysics is a serious mistake.
‘Now Hawkings and Lawrence claim there is no blackhole.”
This is naughty. They don’t deny the presence of gravitational singularities, just some of their properties.
I do not particularly care if stuff is real or made of math. That does not seem answerable.
I perceive science as asking how stuff works. How can we organize our perceptions? It’s a bit like house cleaning and filing. There is no true way to organize and file, but there are differences in usefulness and efficiency.
Not really, Aristotle was well aware of the spherical earth, and the Pythagoreans before him also told of the spherical earth. I write this not to disagree on the whole, since I think your point is well-made, I’m just noting details of an example.
As far as I know, the moment and circumstances of the recognition of the spherical earth is lost to history, and, of course, no modern science ever tested the earth to be flat–it was almost certainly the default assumption. No one seems to know what gave away the spherical earth. The disappearance of ships over the curvature of the earth, eclipses and the sense that earth’s shadow appears on the moon? It seems strange that such a momentous discovery goes unremarked, even in the sketchy writings left from that time, but I suspect that the recognition of the spherical earth spread slowly, and was old hat to those in the know, so why write about it, except incidentally, like the Pythagoreans did?
So much for the trivia…
Glen Davidson
That is a right way to see the science no far from mine. I would say that if science is only models as Lizzie says, there many coherents models that can respond how stuff works, they all intersect the experimental data and differ in the extrapolations.
PS there is no “perfect” coherents models, all model is imperfect has gaps and unresolved questions. More work for scientists!
There are ways of comparing Models. You seem very keen to make all models equal. they aren’t.
Then too, pedants sometimes point out that the earth is actually an oblate spheroid–because, you know, only perfectly spherical objects can be called spheres.
Actually, I did finally figure out why the issue matters, somewhat (other than to space science, GPS, etc.). It was because, other than a few bumps and what-not, earth was assumed to be a perfect sphere by ancient Greeks, reasoning according to shapes, forms, geomety, perfection, some combination of those and perhaps others. French scientists, as I recall, showed that it is oblate, kind of a big deal at the time.
By the approximate terms normally used, of course, earth is a sphere.
Glen Davidson
How do you choose The Model?
Blas,
Some of this:
Maybe some “goodness of fit” R sqaured type stuff
Maybe some http://en.wikipedia.org/wiki/Akaike_information_criterion for parsimony.
All basic science, my friend.
Well I said all the models (at least the coherent models) intersec in that, fit to the data, they differ in the extrapolations because they make different assumptions to make them.
‘Well I said all the models (at least the coherent models) intersec in that, fit to the data,’
This isn’t really true. They don’t all “Fit the data”, they all fit the data to different degrees of accuracy. This is why we employ the statistical tests above.
AIC even trades off between accuracy and simplicity.
Yes, all models are imperfect have trade offs, differents acceptable levels of error.
So we can qualitatively assess them and chose the best. Also, as new data and discoveries roll in, they should help us filter the good from the bad even more. This is why predictive novelty (like finding Tiktaalik) is so good for evolution. ID predicts:
I’m not sure what happened, but I had intended to post that last comment under the “Historical vs Observational Science” thread. I must have had two window open.
Moved.
I’ll eventually move this post (that I am replying to), and my reply to the moderation page.
I stand corrected — Erastothenes’s claim to fame was in calculating the circumference of the earth, not showing that it was spherical to begin with. Thank you!
Seems like as good a place as any:
“Mindpowers” Murray: ” Unless the jurors are swabbing cheeks and conducting DNA tests themselves, the DNA evidence is in principle nothing more than the testimony of an expert witness. The jurors have no means of ascertaining the DNA “facts” for themselves; they entirely rely upon the testimony of what they assume to be a highly credible witness.”
Denyse: “There is absolutely no evidence that space aliens exist. ”
KF: “… 500+ witnesses at the core of the Christian contention…”
Yes, it’s not like anyone else could check out the DNA evidence, after all.
Oh, but that would still be a matter of “testimony.” Sorta, but somehow we believe identical or repeated results more than single testimony about something that isn’t already shown by solid evidence to actually exist.
By the way, I suspect that if Murray were called to the stand as an “expert witness” that he’d be relatively easy to impeach based on what he’s written on the internet. The leap to a “designer” of life by ignoring normal markers of design that life lacks (ability to break from inheritance patterns, notably) and using misleading technical terms to suggest a similarity between life processes and technology that doesn’t exist, should disqualify him easily.
Glen Davidson
Glen:
Yeah, “I don’t care if my beliefs are true” is not what you want your witness to say under oath.
Richardthughes,
Denyse: “There is absolutely no evidence that space aliens exist. ”
But aliens designed and built Stonehenge, pyramids, the Nazca Lines, and other cool stuff! They even removed the top of a mountain so that they would have a flat place to land their spaceships, and aliens still live among us! joey says so and joey is NEVER wrong! I’m shocked, shocked I tell you, that fearless and all knowing joey isn’t screaming at o’bleary that she’s wrong wrong wrong, and that she’s a cowardly assface!
KF: “… 500+ witnesses at the core of the Christian contention…”
The “Christian contention” is a crock. There were no witnesses to something that never happened (the alleged ‘resurrection of jesus’) and christianity is rotten to the core (just like gordo and his ‘ilk’ and ‘enablers’). What a worthless excuse for a man gordo is.
There are actually a lot of really interesting questions and problems at work in the concept of testimony and the epistemology and ethics of testimony (for example, here). I’ve been reading lately about the concept of “testimonial injustice” in Epistemic Injustice as part of my new research project.
As usual, Murray and the others at UD come right up to the point where an interesting philosophical question can be asked — and stop, because they’re so convinced that their prejudices are correct that they are unable to imagine an alternative.
This comes up in the permutations of the EAAN, as oft-quoted from C. S. Lewis and Haldane as well as from Plantinga — they are so convinced that the EAAN must be right that it never even occurs to them to see what the naturalistic responses to the EAAN might be, let alone consider whether or not those responses are adequate. (I think that those responses are adequate, but my broader point here is that the UD people think that the debate is settled in favor of Plantinga. Apparently there are some controversies not worth discussing?)
Likewise, it is so blindingly obvious (“self-evident”) to them that objective morality is inconsistent with naturalism that it never even occurs to them to inquire into whether or not that is indeed the case. So they never feel any need to read Dewey or contemporary philosophers who think that objective morality and naturalism are compatible (e.g. Johnson or Casebeer). Of course there are other philosophers who do think that objective morality and naturalism are not compatible, but that just shows that there is a debate here worth having — and it’s a debate that’s very far from being settled!