Sorry, Robert.  I got here the end of October 2006, so I missed the first half year of your posts.  However, I read all the threads as soon as I registered.  Since then I’ve entered over 400 posts, however, they haven’t been uniformely distributed among the threads.  I responded only to those that seemed to me to be worth discussing, and weren’t merely playing semantic games.  I certainly may be mistaken in my evaluation, but I just wasn’t interested in slogging though the endless verbiage of this thread for, what I saw as, a very tiny bit of interesting challenge.  However, since this thread has generated five pages of responses, it’s obvious that others do not share my opinions.

Occam

Hello Mr. Gulack,

You said the following:

Pragmatic Naturalist cites observation and experiment as an alternative to Occam’s razor.  What we need to realize here is that, for any particular experimental result, or any particular observation, there are an infinite number of theories that “explain” that result or observation (in the sense of being consistent with it).  If I note that heavy and light objects accelerate at the same rate, it could be because invisible demons like playing with heavy objects, for example.
So the basic question is, what is it that science does to narrow down the list from infinity to the few theories that are genuinely worth considering?  And the answer is, you go with the simplest theory that covers all currently available evidence.  That becomes your working hypothesis until more data becomes available.  That is the principle known as Occam’s razor.  It is not an alternative to observation and experiment.  It explains what you DO with observations and experiments.  It is the sole and only method for analyzing factual claims, and, for that reason, it is not only relevant to the issue of human freedom, but the sole means we have of speaking intelligibly about the issue.

I suppose you’re right that observation and experiment are not really “alternatives to Occam’s razor.”  I guess I was confused because I thought we were talking about “factual analyses” before.  But, I admit, it seems clear that Occam’s razor has to do with theory selection, as you indicate. 

Nonetheless, at the risk of being “cavalier”, I just don’t agree that Occam’s razor is “the sole and only method for analyzing factual claims” or for deciding between competing theories.  It is one way; but it is not the only one, and it is certainly not the decisive one.  To defend this claim I’ll provide an example and an alternative methodology.

First, the example:  Was Einstein’s theory of gravity simpler than Newton’s?  I don’t think so.  (Perhaps Occam’s razor would even—mistakenly, in this case—favor Newton.)  Einstein’s theory of general relativity was just more successful for dealing with large astral bodies than Newton’s theory of universal gravitation.  In fact, I think that Einstein was motivated to produce his theory of general relativity because Newton’s theory proved inadequate in making certain predictions. 

Newton’s theory is fine—i.e. it makes basically the same predictions as Einstein’s—for events on earth (where the gravitational force is relatively weak), but its not so good when it comes to much more massive bodies, like our sun.  That leaves us with a pragmatic choice to make: we can still use Newton’s (simpler) theory for mid-sized objects if we want to, or, we can break out general relativity. 

That brings us to the “alternative methodology” part of my response.  The alternative was just mentioned above: it is the “pragmatic method”.  (I’m not just a naturalist, I’m also, by name, a pragmatist.) The direction of inquiry is often guided by particular problems that we want to solve.  That is the key.  (I think that Einstein versus Newton is a case in point.  That is why I used it as an example.)  We want to understand nature and control it when we can—it is in our interest to do so.  Science and inquiry—and the acquisition of knowledge generally—are driven by this problem-solving goal. 

So, how does the pragmatic method help us to choose between, as you put it, “an infinite number of theories that “explain” a result or observation”?

I used this quotation from William James in another thread about truth, but it seems relevant here as well.  James says:

“I am a natural realist. The world per se may be likened to a cast of beans on a table. By themselves they spell nothing. An onlooker may group them as he likes. He may simply count them all and map them. He may select groups and name these capriciously, or name them to suit certain extrinsic purposes of his. Whatever he does, so long as he takes account of them, his account is neither false nor irrelevant. If neither, why not call it true? It fits the beans-minus-him, and expresses the total fact, of beans-plus-him. Truth in this total sense is partially ambiguous, then. If he simply counts or maps, he obeys a subjective interest as much as if he traces figures. Let that stand for pure “intellectual” treatment of the beans, while grouping them variously stands for non-intellectual interests. All that Schiller and I contend for is that there is no “truth” without some interest, and that non-intellectual interests play a part as well as intellectual ones. Whereupon we are accused of denying the beans, or denying being in anyway constrained by them! It’s too silly!” (William James, “Letter to Dickenson Miller, Aug. 5, 1907,” The Letters of William James, pp. 195-6)

James here, I think, lays out the keys to the pragmatic method.  The universe, or living organisms, or freedom, or whatever we are trying to explain, is like a bunch of beans on a table.  The beans are there.  They are real, not figments of our imagination; and, we can describe them (i.e., reality) in various ways.  We choose certain ways rather than others because it is beneficial to our goal of solving (certain) problems—and therefore, for living successfully.  The “beans”—i.e. the cold, hard world that we have to live with—faces us with problems.  It’s not easy being alive.  It’s a lot of hard work.  We are, as individuals, temporarily resisting the trend (in the rest of the universe) toward increasing entropy.  And, as a species, our practice of acquiring knowledge and finding out about—and learning how to control and predict—the workings of nature provide us with tools to better cope with this precarious existence. 

My point here is that, when dealing with the cold, hard world, the simpler (theory) is not always the better (though it has its obvious benefits at times, I admit).  The more successful is the better.  And this “success” is measured in consequences.  (That is why I brought up experimentalism before: it’s a way of assessing consequences and testing theories.)  The pragmatic method—at its core—has its eye on these consequences.  That is why I think it is apt to describe this method in terms of its solving problems for us. 

Thanks for your time,
PN

All that you have to do to realize that everything that William James says about pragmatism is absolute nonsense is ask yourself this simple question:

“How do I know what particular CONSEQUENCES have occurred?”

If James is right, you can only judge what consequences have occurred by evaluating the consequences of believing that those consequences have occurred.  And one can only evaluate THOSE consequences by . . .

In other words, James is leading you into an infinite regress, in which there is never ANY method for determining the truth value of any statement whatsoever.

Once you realize that the only method for evaluating what consequences have occurred is to use Occam’s razor, you will easily see that you should also use the Razor to evaluate the truth of the original statement in the first place.  There is just no escaping organizing evidence in the simplest possible manner.

Bob Gulack

Okay.

You win.

FWIW I’m with Pragmatic Naturalist on this one. Occam’s Razor is a useful tool, not an infallible holy text. Sometimes the simpler answer isn’t the right one. As for relying on real-world outcomes (aka “consequences”) as a guide to the truth or utility of a theory, it is fundamental to science. Absolute truth is a philosophical and religious concept. Science deals with practical truth. There is no more an infiinite regress in looking at outcomes than Xeno’s Paradox is a real barrier to an arrow hitting a target. Theoretically one can imagine that evaluating consequences requires evaluating consequences and blah blah blah, but in the real world a certain degree of certainty is good enough.

mckenzievmd,

Thanks for the support.  Very well said!

PN

mckenzievmd - 17 June 2007 09:39 PM

. . . Occam’s Razor is a useful tool, not an infallible holy text. Sometimes the simpler answer isn’t the right one. . .

Maybe not, but it’s usually a hell of a lot easier to read.  :grin:

Ok, I agree that Brennan’s answer was succinct, cogent, rational, and quite clearly stated, but I couldn’t help making my comment.

Occam

I never said the Razor was an infallible text.
By its very terms, the Razor confesses NOT to be.
All it says is, adopt as your current working hypothesis the simplest theory that covers all currently available evidence.
That is how scientists work (see Einstein’s discussion of Galileo).
The point here is to realize that the Razor is the right tool to use, and then use it on the free will factual claim.
BOB GULACK

Another way of putting the matter is to ask,

“Under what conditions do we become aware that ‘the simplest answer is not the right one’”?

Clearly, it’s when new evidence has emerged that is not consistent with the previous “simplest answer.”

In other words, the fact that Occam’s razor leads to the continuous development of scientific ideas is not a disproof of the Razor, but a proof of it.
BOB GULACK

I think your argument is a bit circular. The simplest answer is correct unless new evidence shows it is not, then the next simplest answer is correct. No one is trying to disprove Occam’s Razor, only to demonstrate that it is not the only criteria by which to determine which of several hypotheses is correct and it is not always correct. If there is one case in which Occam’s Razor selects the wrong hypothesis, then it must be true that it is not an infallible way to make such a selection. And if there is one example of using another method to select the hypothesis that turns out to be correct, than it cannot be the only method. I think there is no difficulty finding such examples. I routinely try to assume all of a pateint’s complaints are related to a single cause because this is parsimonious, and it is true more often than not. But it is hardly unusual for a patient to have more than one pathology with complaints related to each, and if I blindly follow Occam’s Razor instead of keeping an open mind for the possiblity of a complex, multifactorial etiology for the complaints, I’m going to incorrectly diagnose a fair number of patients. And if I have a complex hypothesis (multiple causes) and a simple hypothesis (one cause) and I use statistical probability instead of simplicity to select between them, there are plenty of cases in which this yields a better choice.

Again, Occam’s Razor is a powerful and useful tool, and we all agree on this. I think the only disagreement is with some of your more strident statements about it being the only way that we ever find out what is true.

There are two possibilities here.
Either we can agree with Einstein that the basic principle of science is parsimony, or we can argue that there is more than one way to do factual analysis.
If there is more than one way:
(1) What are the competing principles?
(2) How do you resolve conflicts among the principles?
BOB GULACK

Either we can agree with Einstein that the basic principle of science is parsimony

I believe you may be referring to this quote:

“The supreme goal of all theory is to make the irreducible basic elements as simple and as few as possible without having to surrender the adequate representation of a single datum of experience”

In this case, I think you are overlooking the importance of the qualification. The basic principle of science is the adequacy of the representation as judged by practical outcomes, and while parsimony aften furhters this goal, it is not itself sufficient for science. The best theory is that which best explains or represents the phenomenon udner consideration, regardless of its complexity. Again, simplicity usually is best, but not always.

I wouldn’t say that I am arguing there are competing principles of factual analysis, but that simplicity is one charactersitic of a theory which suggests it may be superior to others. Plausibility is another characteristic, similarity to other proposed theories or mechanisms which have turned out to be correct is another, etc.

I am sorry to have to continue what might seem an overly technical argument, but here goes:

(1) “Similarity to previously successful theories” is another example of simplicity.

(2) The term “plausibility” means nothing, unless it means “similarity to the kind of thing we already accept”—in which case, see (1), above.

There still is nothing really going on here besides Occam’s razor.  One cannot even refer to “outcomes” because there’s no way to judge what the “outcome” has, in fact,  been without using the Razor.  Let’s take the French experiments following the discovery of X-rays, which purported to discover another new kind of radiation, but turned out to be all self-deception.  The so-called “outcomes” there all eventually failed the test of the Razor and were discarded.

The reason I’m pushing this so hard is, of course, because, until we settle on a criteron for factual truth, we can’t have any discussion as to what is true or false in any particular case.

Bob

Well, I don’t see how “similarity to previously successful theories” is the same thing as “simplicity” at all. And plausibility means, essentially, consistent with what is already accepted as known, as opposed to a theory which requires a rejection of established knowledge. Again, I don’t see this as the same thing as similar to other theories or as simple. Now, I suppose we could start playing the dictionary game, arguing over what words like “simple” mean, but usually a discussion that has reached that point is de facto doomed. By simple, in terms of a scientific theory, I mean the fewest number of constitutive elements/unknwons/variables, and perhaps a relativley straightforward arrangement of those elements. “Simple” says nothing about the relationship of the theory to other theories or established knowledge.

As for outcome, I do think you’re making this more complicated than it needs to be, Bob. Theories allow for predictions and experiments measure whether the results are consistent with these predictions. I don’t see how it is impossible to know anything about this process without the principle that the simplest explanation is generally the best.

1. I’m not really clear on what you mean by “similarity to previously successful theories.”  That was your suggestion for a factual criteron, not mine.  But if you mean, for example, that our ideas on how apes evolved can inform our ideas on how whales evolved, then that would, indeed, be an example of using the same scientific principle to explain more than one thing—i.e., an example of simplicity.  (The more data can be explained by the same theory, the stronger that theory is under Occam’s razor.)

2. You have now defined “plausibility” (which, again, was your suggestion, not mine) to mean “not rejecting previously established knowledge.”  But don’t you see that, once we have explained earlier data with a simple theory, thus “establishing” knowledge, it is simpler to come up with a theory that doesn’t overturn our earlier work?  Once again, the old data and the new data would both be explained, and this would be simpler under the Razor.

3. As for your allegation that the results of experiments are absolutely clear and speak for themselves, I have already given the example of the French scientists who made fools of themselves trying to surpass Rontgen’s work on X-rays.  People very often fool themselves about research, and very often have to interpret research that is on the edge of experimental error.  All of this assessment is done using Occam’s razor—comparing how likely it would be for the experimenter to have made an error vs. how likely the new theory is.

Bob