Is theory choice constrained only by epistemic factors?

Discussions on the philosophical foundations, assumptions, and implications of science, including the natural sciences.

Re: Is theory choice constrained only by epistemic factors?

Postby Lomax on December 9th, 2016, 12:24 am 

BadgerJelly » December 9th, 2016, 4:43 am wrote:Either way would appreciate it if you could explain in laymans terms.

NoShips is asking what we think about the underdetermination of evidence. To take a simple (and oversimplified) example, suppose you are testing whether a new toaster works. You put the bread in, wait five minutes, and remove the bread, still untoasted. Is the toaster nonfunctional? Perhaps there was a localised power-cut? Perhaps it's special, untoastable bread. Perhaps it did toast, and you are dreaming, or on drugs. Perhaps the toaster works all but a statistically insignificant amount of the time, and by sheer freak coincidence your breakfast fell into the realm of statistical insignificance.

The point is that we can always come up with more than one theory to account for our observations. Serpent is right when he says we can empirically test which of the above theories is most likely, but even those tests will be open to the same problems. And so on. This is known as the Duhem-Quine Thesis.

So the idea behind NoShips's (and Quine's) thinking is that we ultimately can't confirm a theory based solely on empirical observation. Evidence is a necessary* but not sufficient factor. Quine suggests that we should and do rely on other factors, such as the simplicity, elegance and conservatism** of the theories.

I think NoShips is asking for a counter-argument. Donald Davidson, a philosopher of epistemology, metaphysics, mind and language, and a former student of Quine's, argues that in order to know how to choose between theories (regardless of what factors we use), our reasoning about such things (whatever it may be) has to be innate and a priori, or else we wouldn't even be able to match the factors to the theories which best fit them***. In other words our proof theory must be something we just intuitively know. While I don't know how to prove Davidson wrong, his argument suffers several shortcomings. It can't tell us how we know, why we all disagree about logic, who is right, or how to find out. And arguably it is circular. And arguably so is the Duhem-Quine thesis.

Sorry if that wasn't very lay; these things are mindbending. Men as formidably intelligent as Chomsky, Kripke and Strawson (Sr.) seem to miss the point when debating with Quine. I did my best.

* This is an oversimplification too, but a necessary one, for now.
** Their resemblance to what we already believed.
*** Lewis Carroll made a similar argument more than a century ago, in the form of a children's short story.
User avatar
Lomax
Forum Administrator
 
Posts: 3664
Joined: 01 Jul 2010
Location: Nuneaton, UK
NoShipsScruffy Nerf Herder liked this post


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 12:52 am 

That is kind of what I thought. I think Husserl covers this.
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 2:26 am 

Oh dear, that scary badger is here, and he does bring the most dreadful friends to parties. If you promise to stop mentioning Germans whose names begin with H, I promise to bring world peace to mankind within six months.

Lomax's summary is closer (I think! -- who knows what that Husserl is on about) to the issue I want to examine.

Let's start with some clarification then. To begin with, at least, I'd like to restrict the definition of "evidence" to that which is observable. How exactly the observable/unobservable distinction is to be drawn shall be left simmering on the back burner in the meantime, though may come back to haunt us later.

Furthermore, I will assume, as I daresay almost everyone does, that evidence is an epistemic virtue; there is a relationship between evidence and knowledge. The accumulation of evidence adds to a theory's epistemic warrant. The number of letters in a theory's name, on the other hand, or the number of Creationists who believe or dismiss a theory, presumably has no bearing whatsoever on a theory's epistemic warrant. These are not epistemic factors.

That scientists appeal to evidence in choosing and appraising their theories is entirely uncontroversial. The observation of a particular behaviour in coconut crabs, say, might be taken as evidence to either confirm or disconfirm the theory of natural selection.

The naive (in my opinion) claim I want to scrutinize in this thread, then, is:

C1: "Scientists are constrained in their theory choices only by evidence and logic."

In other words, only epistemic factors (evidence and logic) are deployed in theory choice.

Consider the egregious Lysenko affair in Russian biology wherein Stalin more or less dictated how science was to be done. "I'll tell you which theory is right, comrade, or else you die." Here we have an obvious example of socio-political (i.e. non-epistemic) factors influencing the workings of science and scientific theory choice. Many of us would be inclined to say science was not being done at all. I offer this merely for illustration; not as a counterexample to C1.

Back to reality, how would one go about disproving C1? Well, one way would be to adduce a case where two logically incompatible theories, T1 and T2, are empirically equivalent (i.e., they rule in and rule out precisely the same set of observations; their evidential fit is identical), yet one theory was selected over the other. In such a case, clearly, epistemic factors (evidence and logic) were insufficient to break the deadlock, so how was the deadlock broken? It would appear non-epistemic factors must have played a role.

What might these non-epistemic factors be then? Here are a few options: simplicity, coherence with our general scientific body of knowledge as a whole, explanatory goodness, elegance (whatever that is), social factors, psychological factors, political factors...

The options towards the beginning of my list are probably not particularly controversial even to scientists themselves; those towards the end, I expect, will be fiercely resisted. For example, let's say T1 and T2 are empirically equivalent, but scientists adopt T1 and disregard T2. We ask them why:

"We chose T1 because it's simpler than T2", or "We chose T2 because even though both theories are empirically equivalent -- they fit the evidence equally well -- we feel that T2 provides a better explanation of the evidence."

In these cases, it was simplicity and explanatory goodness, respectively, which broke the deadlock epistemic factors could not break.

Now, at this point some might want to argue that simplicity and explanatory goodness are epistemic factors. Occam's razor was mentioned earlier. For some this is just a pragmatic methodological principle: "all else being equal, it makes pragmatic sense to choose the simpler theory"; for others it transcends pragmatics: "all else being equal, the simpler theory is more likely to be true".

Likewise for explanatory goodness. Some (e.g., Bas van Fraassen) argue that explanation falls into the domain of pragmatics; others argue, to the contrary, that a theory providing a good explanation (whatever that means) for a body of evidence adds to the theory's epistemic warrant. (so-called Inference to the Best Explanation).

Pace Husserl, I haven't come across anyone in my own reading who argues that social, political or historical factors bear any relationship to epistemic warrant.

As Lomax notes, philosophers blether a lot about the underdetermination of theories by evidence, the doctrine that for any body of evidence an indefinite number of theories can be constructed to accommodate that evidence. In such case, all these theories are empirically equivalent. Real world science examples of underdetermination are much harder to come by, although there do appear to be one or two. The Ptolemaic vs Copernican model of the cosmos circa 1600 may be one. It seems that James Cushing has drawn our attention to another.

On Cushing's account, we have an example of two theories (Copenhagen vs Bohm) which are empirically equivalent, yet one reigns supreme while the other languishes in obscurity. If Cushing is right then we must look to non-epistemic factors to explain this choice of one theory over the other. Cushing's conclusion is that it was a simple matter of historical contingency -- Bohr et al got there first.

Therefore (*drum roll*), there is more to theory choice in science than evidence and logic. Get off my plane, C1.


Not sure how all this squares with Husserl, badger, but I'd love to hear your analysis. Probably won't understand it though :-(
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 2:38 am 

I'll do my best not to mention H starting to annoy myself! Haaha

Anyway, I have a whole thread for that and more to come ;)

I would like to ask if we are not talking about "evidence" or "logic" then we are talking about "non-evidence" and "non-logic"? Or is that a logical presumption? See the issue here?

We cannot say something is not evidence. "Something" means it is evident. This means science has to set out what evidence is valid for its purpose by way of some form of judgement. "Logic" is the realm of language, but in science mathematical logic holds sway. Verbally/politically "logic" is bound by the meaning. Science does not deal directly with "meaning" only empirical data. The scientist deals with the politic being a human being.
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 3:12 am 

BadgerJelly » December 9th, 2016, 3:38 pm wrote:
I would like to ask if we are not talking about "evidence" or "logic" then we are talking about "non-evidence" and "non-logic"? Or is that a logical presumption? See the issue here?



Erm, but we are talking about evidence and logic. I'm suggesting there's more to theory choice, in at least some cases, than just evidence and logic.

The naive claim (C1) is that only evidence and logic are relevant to theory choice.
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 3:20 am 

What is logic and evidence? That is the obvious question that needs to be addressed is it not?

Or rather what is evidence without judgement and logic without application. Mathematical logic is auite different from rational thought in our non-absolute everyday language. To judge what evidence I need to look for is done through the rationality of everyday language (including technical jargon) not through the preicision of mathematical absolutes (numbers).

Understand?
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 4:00 am 

BadgerJelly » December 9th, 2016, 4:20 pm wrote:What is logic and evidence? That is the obvious question that needs to be addressed is it not?

Or rather what is evidence without judgement and logic without application. Mathematical logic is auite different from rational thought in our non-absolute everyday language. To judge what evidence I need to look for is done through the rationality of everyday language (including technical jargon) not through the preicision of mathematical absolutes (numbers).

Understand?



I think I'd rather just stick with our intuitive understanding of the terms, lest we get sidetracked all the way to Burkina Faso, but they are good questions. If scientists call it evidence, then that's good enough for me at the moment.

Scientists, themselves, will disagree over what counts as evidence, of course. Observation statements (i.e. empirical evidence) are fairly uncontroversial, I think, whereas my examples of simplicity and explanatory goodness, I expect would be disputed. Some scientists might say, for example, simplicity is a purely pragmatic virtue, thus has no bearing on epistemic warrant (i.e., the simplicity of a theory does not count as evidence for its truth). Others might argue that simplicity does bear on epistemic warrant (i.e., simplicity does count as evidence for a theory's truth -- Einstein speaks this way at times, I believe).

For those who consider simplicity and explanatory goodness to have some bearing on epistemic warrant, simplicity and explanatory goodness might be termed "non-empirical evidence".

But what is scientific evidence really? Simply reasons for holding a particular belief? Some kind of logical relationship between an observation statement and a theory? Or something else? Peter Achinstein is the best man to read on this. He identifies several (four) different ways which scientists employ the term -- probably without being aware of the differences.
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 4:21 am 

If scientists disagree over what is "evidence" then some would say that some scientists would call X evidence but not Y, and other Y evidence but not X. What concerns the scientist in making such a judgement? Is it a "reasonable" judgement?
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 5:24 am 

Well, always wanted to visit Ouagadougou...

It's in situations like the one you advert to above, badger, that our conflicting intuitions vis-à-vis the meaning of the term "evidence" are brought to the surface.

J.C. Maxwell supposedly said (quoted in Laudan somewhere) that the luminiferous aether is more highly confirmed than any other theoretical entity, or put another way, there's lots and lots of evidence to support our belief in said aether. (whether he actually said this or not doesn't matter)

We're now told the aether doesn't exist, and one intuition screams that there can be no evidence for an entity that does not exist. Another intuition, meanwhile, leaves most of us feeling decidedly uncomfortable at the suggestion Maxwell et al were behaving irrationally or unreasonably. It seems we'd like to say both that there was evidence for the aether and there is no evidence for the aether.

The same could be said for our contemporary scientific beliefs: black holes, say. Presumably a physicist would tell us there is strong evidence to believe in black holes. On the other hand, science is a fallible business; it's by no means inconceivable that black holes will go the way of the dodo and the aether. And then what do we say? Was there evidence for black holes or not? After all, how can there be evidence for a non-existent entity?

I'll reproduce below a blog I wrote some time ago:



It seems our members are forever at each other's throats over what constitutes "evidence". The religious poster may claim the evidence for God is all around; the UFO fan will adduce abundant eye-witness testimony as evidence for extraterrestrial visitation; meanwhile the science fan is liable to dismiss both -- and none too graciously either -- while trumpeting the superior virtues of "scientific evidence".

You're not likely to hear him or her explain precisely what that means though.

Well, why all this discord, chaps? Could it be that there's more than one concept of evidence in play? Could it be that this is the case even within the domain of science, which is primarily our concern here? After all, scientists themselves are by no means univocal regarding precisely what counts as evidence. Consider, for example, those entities such as the luminiferous ether, endorsed by the entire scientific community throughout the 19th century, and even hailed by J. C. Maxwell as "the most highly confirmed entity in all science", much as we're often told today of the "overwhelming evidence" for evolutionary theory.

(Although once again, how the rest of us are to know what constitutes "overwhelming evidence" is left unexplained. )

If the ether turns out to be illusory, as is now apparently the case, what are we to say of all that overwhelming evidence? Did these sombre, costive, bearded Victorians have evidence or not? Should we say:-

a). There is good evidence that the ether exists
b). Between 1800 and 1900 (to simplify) there was good evidence that the ether exists. After 1900 there was not.
c). There never was good evidence that the ether exists

Philosopher of science, Peter Achinstein, endeavors to add some clarity to the evidentiary imbroglio. In his highly recommended, if rather dry, "The Book of Evidence", Achinstein reviews and rejects as inadequate previous philosophical explications of evidence (raven paradox and all that) and identifies the following four concepts of evidence, all of which, he claims, are used by scientists, more than one of which may be applicable simultaneously.


1. Subjective evidence - If e is believed to constitute evidence for hypothesis h, then it does.

On this construal of evidence, neither h nor even e need be true. All that matters is that e is BELIEVED by some individual or group to be evidence for h. If you think it's evidence, dahlin', then it is. Insofar as scientists believed there was evidence supporting the existence of the ether, between the years 1800 and 1900 there was.

And if e is YOUR evidence, then you believe e constitutes veridical evidence (see below)


2. ES evidence - Given all that is known in a particular context or "epistemic situation", e provides ample justification for a belief in h. Such an interpretation of evidence is objective in the sense that it represents a relationship between evidence and hypothesis based on a given epistemic context, irrespective of whether anyone actually believes e, h, or that e constitutes evidence for h.

Thus, relativized to certain epistemic situations such as that which may have obtained in the 19th century, there is good evidence that the ether exists.


3. Veridical evidence - The real McCoy.

Veridical evidence provides, in Achinstein's terms, "a good reason to believe", to be contrasted with a good justification for belief in the case of ES evidence. Even though both are objective and neither depends on what anyone actually believes, veridical evidence differs importantly from ES evidence in that while 19th century scientists may have been perfectly justified in believing in the ether given all that was known at the time, their ES evidence nonetheless does NOT in fact constitute a good reason to believe. There can be no veridical evidence for the ether hypothesis inasmuch as h is false.

Veridical evidence is analogous to the signs or symptoms of a disease. The presence of a particular rash may provide good reason to believe that the patient has a certain disease, regardless of what anyone knows or doesn't know about such things. Veridical evidence need not be conclusive though. To use Achenstein's own example, evidence that Jones bought 999 tickets in a 1000-ticket (fair) lottery provides a very good, although not conclusive, reason to believe the hypothesis that Jones was the winner. If Jones was not the winner, though, there was no veridical evidence to be had; veridical evidence requires that h be true. Which brings us to...


4. Potential evidence

More modest than veridical evidence. All veridical evidence is potential evidence, but the reverse does not necessarily hold. Like veridical evidence, potential evidence requires that e be true, but unlike veridical evidence, it does not presuppose the truth of h. That rash can constitute potential evidence for the patient having that particular disease, even if it turns out that the patient does not, in fact, have the disease.

While veridical evidence demands that e be the kind of thing that IS a good reason to believe h, potential evidence demands only the e be the kind of thing that CAN BE a good reason to believe h. (If I'm understanding Achinstein correctly)

Hope that's clear. If not, buy the book, dammit, and support Prof Achinstein's upcoming weekend break in Asbury Park, a rare splurge thanks to double-figured book sales.
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan
BadgerJelly liked this post


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 6:30 am 

At the end of the day, stuff apparently happens. We guess how it apparently happens and occasionally it seems to be correct. From this apodictic frame we ordinate our worldly sense of being.

Or in simplier and more specific terms. I like strawberries!! People who don't like strawberries should either be burnt at the stake or put to work in harvesting and developing strawberry farms.
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 6:38 am 

More of a mango and lychee man myself. Oh, and durian.
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 6:47 am 

NoShips » December 9th, 2016, 6:38 pm wrote:More of a mango and lychee man myself. Oh, and durian.


Mayne the differences of science and philosophy are like human mango eating habits. One group eats them ripe and juicy and the other also eats them unripe with salt and chilli whilst the others look on in disgust and partial amusement. Scientists need to rediscover the mango and look for other uses.

Being a strawberry man that basically puts me in the madhouse I guess ... :(
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 6:50 am 

What's all this got to do with Husserl the Giant Killer?
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Re: Is theory choice constrained only by epistemic factors?

Postby BadgerJelly on December 9th, 2016, 6:59 am 

NoShips » December 9th, 2016, 6:50 pm wrote:What's all this got to do with Husserl the Giant Killer?


You mean Husserl and the Giant Peach?
User avatar
BadgerJelly
Resident Member
 
Posts: 5364
Joined: 14 Mar 2012


Re: Is theory choice constrained only by epistemic factors?

Postby NoShips on December 9th, 2016, 9:36 am 

Lomax » December 9th, 2016, 1:24 pm wrote:
Sorry if that wasn't very lay; these things are mindbending. Men as formidably intelligent as Chomsky, Kripke and Strawson (Sr.) seem to miss the point when debating with Quine. I did my best.



Well, since I'm bored and have nothing better to do... (Godfather II tonight). Is this what you mean, Lomax?

When Quine proposed his indeterminacy of translation thesis, Chomsky responded something like this: "No, no no, that is not indeterminacy. That is a phenomenon very familiar to philosophers of science; namely, the underdetermination of theories by evidence."

Which is to say (on Chomsky's view), there is a fact of the matter as to what you, me, badger, or the radical field translator means; it's just we can't be sure what it is. (It may be underdetermined, but it's there somewhere.)

Quine says nay nay nay, it's much worse than that: "There is no fact of the matter". *panic* (anyone thinking of quantum physics now?)

Have I got this right? If so, the underdetermination thesis I might (or might not) defend, is -- contra Quine -- that if a whole emporium of empirically equivalent, i.e. observationally indistinguishable, theories are laid out before us, exhausting all possible theories that could accommodate the body of evidence at hand, there is nonetheless, a true one among them. We just don't know which one it is.

That would make me a scientific realist....and maybe a Chomskian ... for today anyway.

Wot about you, guvnor? To repeat, is this what you mean? *hiccup*
User avatar
NoShips
Banned User
 
Posts: 1852
Joined: 07 Oct 2016
Location: Taiwan


Previous

Return to Philosophy of Science

Who is online

Users browsing this forum: No registered users and 6 guests