Flickr

  • www.flickr.com
    Abiola_Lapite's photos More of Abiola_Lapite's photos

« Are You An Idiot? | Main | Yet More Korean Racism »

March 09, 2006

Comments

Chuckles

Both interpretations actually derive from human psychology. It makes it just so much easier to claim that \"X Group is Intellectually Inferior because it lacks Y gene\" instead of multiplying entities in the form of Geography, Culture, Politics - etc; with the neccesary plural entities and their deconstructions that these would also imply.
The reality simply seems to be that with regards to human behavior in social settings; entities *are* multiplied *beyond* neccessity: Since what we deem \"neccesary\" to explain certain social phenomena is contingent upon our knowledge at the time.
You would think that provisionalism would imbue more \"race researchers\" with caution, as opposed to opinions about the so obvious lower intelligences of certain groups.
Though many loathe to admit it; Occam thinking guides a lot of Molecular Biology: One gene and that\'s it; knock this gene out and wham, bam, kaboom.
For materialists, genetic determinism is an incredibly strong temptation and when reinforced by cognitive modules like Occams; it becomes nigh irresistible.
Its the old \"simple\" versus \"simplistic\" thing.

Barbar

[Non apology "apology" deleted. As a reward for the ridiculous barb attached to your "apology", I won't give you even the satisfaction of seeing your last comment on here remain intact. You don't come into my yard and insult me on my own property. - A.L.]

Abiola

What a wonderful way to show contrition - tack on an insult unto a supposed "apology". You can go to hell: any new comments by you will simply be deleted.

dsquared

surely no mathematical model postulates either more or fewer entities than any other mathematical model and they are thus all on exactly the same ground with respect to Occam's Razor? I think that's what this paper is basically saying; Occam's razor is about entities, not complexity. It would be quite odd for a medieval churchman to have said that things had to be assumed to be as simple as possible, which I suppose is why he didn't.

Notoriously, even in its most defensible form, Occam's Razor is a quite debatable principle of philosophical logic. It is very hard to explain what the relationship is between a class and its members without postulating a few entities.

Andrew

"Occam thinking guides a lot of Molecular Biology: One gene and that\'s it; knock this gene out and wham, bam, kaboom."

Occam thinking guides almost all sciences, not just molecular biology. Scientists are always talking about their theories as being the most "parsimonious" interpretation of the data and so on. As for single-gene knockouts, it's actually quite common for them not to "work" as expected, i.e. you knock out a gene and see no phenotype, due to redundancy, compensentory effects, cross-talk, etc. or just because your model was wrong. People publish them because after you've spent two years of your life making a knockout mouse you usually want at least something to show for it. And then it's only the 'simple' knockouts with a clear phenotype that stick in people's minds. (And, indeed, it's 'simple'/'elegant' explanations in general that stick in people's minds - and get published in Science and Nature, and bring their discoverers fame and tenure... until people do more experiments and find, invariably, that it's a whole lot more complicated than we thought.)

dsquared

I think you have it backwards, Andrew; people describe their models as "parsimonious" because they've heard that "Parsimonious" is a good thing to be (precisely because of this cargo-cult of Occam's Razor). And because there is no objective definition of "parsimony", you can almost always get away with it.

Most models are typically made up of complicated relationships between a small number of basic entities (and therefore "parsimonious" about the entities they postulate) or simpler relationships between a larger number of entities (and are therefore parsimonious in the number of relationships).

Abiola

"Most models are typically made up of complicated relationships between a small number of basic entities (and therefore "parsimonious" about the entities they postulate) or simpler relationships between a larger number of entities (and are therefore parsimonious in the number of relationships)."

Indeed, and that's before even getting into how we define and count "entities." Then there's also the question of whether one ought to prefer more complex relationships between "simpler" entities to simpler relationships between more "complex" ones (Ptolemaean cycles and epicyles vs. Keplerian ellipses), or how one ought to best trade off simplicity (however defined) against goodness of fit. If all these issues were so easy to settle, AI research wouldn't be a field in perpetual infancy.

gene berman

I think everybody here has a handle on the razor--which is a good thing, I guess.

As Chuckles pointed out, it's a concept deriving from (or at least in accord with) human psychology. Apparently, we're programmed with (along with the ability to appreciate logical relationships) prejudice toward parsimonious explanation in preference to those less so. And, to the extent that significantly less parsimony-preferring individuals have existed, they've been culled by environmental pressure, which is the only sort of "proof" we may ever have that the simpler explanations are more in accord with reality.

There are two related matters of interest to discussions in economics. The first is the notion of the "higher productivity of roundabout methods" (Bohm-Bawerk more than a century ago). The concept is not a law but a very generally prevailing condition: all those methods of production now in existence reflect the best efforts of the most qualified aspirants to bring forth the greatest production with the least costs in the shortest time. (And, thus, any attempt at improving either the quality of the product or reducing associated costs will necessarily involve more complex methods requiring a longer time period--including concomitant interest--before satisfaction may be had.) There is nothing that makes the rule true in every case--indeed, it is not true in every case; on the other hand, it is so true in almost every case that it may be safely said that all those cases in which it is not true must involve "breakthrough" or unanticipated invention in science or technology (or of previously unrecognized sources of satisfaction-generating commodities).

Later (and in what context I've long forgotten), Mises raised the matter of the future of human progress (in "standard of living") absent the driving force of technologic innovation responsible for such dramatic improvement in recent centuries. His observation was that, even were invention and dramatic innovation to CEASE ALTOGETHER, there are so many "recipes" for improvement in existence (and constantly coming into existence but remaining unused because unprofitable--consuming more resources or time than those in use) that mankind could never "run out" of ideas for improvement. Thus, the advent of fiber optics (and wireless) technology, while greatly reducing demand for copper for certain uses, may greatly enhance its use for some other purposes heretofore constrained by the price height due to its now-dwindling former uses. It is even possible that, at some lowered price, demand might exist for even more of the good than existing formerly. (And might bring to development inventions or other innovation formerly impossible because of the cost of the copper required.) At present, we produce roadways (and many other goods) less durable than might be desired under only slightly different conditions--chiefly reflected in the interest rate). Substantial improvements in all such production can be achieved simply by using a little more of whatever substances are required--paying off in greater durability, less repair, etc. (leaving the savings minus the cost of extra material for a greater amount of the other things--not excepting "leisure"--constituting
our standard of living.) Even the quantities of various resources available change dramatically with their relative requirement
(and price. If oil were to double in price, "reserves" would expand greatly-- to include deposits now unprofitable to exploit).

From what we are able to understand of evolutionary process, an organism whose perception of its environment and its abilities (in man, the cognitive processes) to react to these perceptions will be favored to reproduce the more closely perception and cognition conform to reality. Even the concept of favoring the "simple" explanation is far from simple. And, I might add, so is the concept "explanation." For instance, what we understand as explanation most generally involves prescription: how to treat, to use, to react to whatever may be the phenomenon we seek to explain. Indeed, there is no evidence that we could ever even think about anything unconnected to its significance for some aspect of our behavior.

Throughout recorded history, most people have had an aversion toward snakes. The serpent stars in the biblical role of the "fall from grace," and our observance of the nearly universal occurrence of this aversion, combined with our realization of the age of the human species, leads inescapably to the conclusion that the aversion has much older roots and that the explanation may have some different cause than the deity's curse upon Suborder Serpentes. Some snakes are poisonous, their bites potentially and often fatal. A primitive behavior of avoidance, fear, and hostility has obvious survival payoff; inconsequential for such survival is any recognition that most species are harmless (and,when killed, both venomous and harmless
are equally useful as food or source of other utility). A more correct perception of reality has, consequently, little survival payoff and thus lesser likelihood of even engaging mens' attention, under constant pressure for direction to matters of seeming importance. There can be no doubt that the correct explanation of the threat posed by snakes is far less simple but that the historic was (evidenced by survival of its adherents) not only parsimonious but (measured by result) "good."

A fellow put a sign "NO ELEPHANTS ALLOWED ON THIS PROPERTY" in his front yard. When his neighbor remarked that there weren't any elephants within hundreds of miles, he replied, "Works damn good, huh?"

The important thing to understand is that people generally agree (though not necessarily across time) on the matters of most importance to them and disagree on those of declining importance. Long before any formulation of thought regarding the higher productivity of specialized effort, people were specializing their efforts and cooperating (the latter inseparable from the former). "I'll move it in and out--you move it 'round and 'round and it'll pay off for both of us, baby" might not be the absolute original but it's pretty far back, all things considered.

To earlier men, "God made the world and everything in it" was about as parsimonious as you can get. And whether or not everyone believed such a thing--even in those days--is beside the question. The plain fact is that every historic society of which we know from the past and on which all presently existing societies are based have been "believing" societies. The few attempts of which we are aware to construct societies in which disbelief is foundational have been dramatic (and nightmarish) failures at the very same time as the more traditional (believing) societies have come to include
larger and larger numbers of unbelievers (of both the militant and uninsistent varieties) as well as larger and larger numbers of those for whom their belief is, at best, partial, hedged, and of dwindling significance for behavior. "A designer made the world and everything in it" is simply a ploy designed to restore the prestige of ancient belief by appeal to its original--parsimony; indeed, it has no positive program of explanation whatever--simply an attack on the innumerable (and presently-considered) and incomplete unparsimonious explanations of the very people who profess to revere parsimony. (And my view of the creationist intellectual leadership--as out-and-out "conmen"--ain't bad itself in the parsimony category.) Parsimony is, itself, chimerical; though always attractive and, quite possibly somewhat logically supportable, it is the same quality which
makes "Bush lied," "the Jews are behind it," and other conspiracy accounts so salable and so durable. Wherever there is no completely staisfactory explanation for any phenomenon of any importance, we can be certain that competing explanations will be offered by those "in the business" (e.g. scientific specialists of several types, theologans, conmen, as well as assorted attention-seekers and nutcases--all only imperfectly distinguishable from each other). Take your pick. We never pick our explanation on the basis of its parsimony but, after we've picked one--it will certainly appear to have that recommending feature.

Chuckles

[...God made the world and everything in it...]

Which of course is the fun part. Some time ago, I was going through an article on Occams and Gravity and the author said something like "Saying God pushed the apple down is multiplying entities beyond the neccesity of mass and distance required to explain gravity" - the foolishness of this explanation lay of course in the fact that "God pushed the apple down" has only one causative entity and the latter example has two (assuming of course a simple definition of entity, i.e. verbs/actions wouldnt be entities). The whole notion of empirical justification comes into play here: But not even the latter is an exhaustive empirical justification of the phenomenon in question - and at atomic/subatomic scales it gets even weirder.
It really is funny when people start thinking that "parsimony" automatically imbues their arguments with some mystical accuracy. And I am not really suprised that this principle got so solidified in a monotheistic culture. The old catch about cutting out elements in a theory that do not increase its predictive power is a very interesting application of the Razor, but it is also false, in so far as science is descriptive. There is no rational reason to omit molecules in a biological network, which appear to have been vestigially coopted in a signaling cascade, seeing as the cascade would function quite properly should they be knocked out. Now, parsimony might dictate that a theoretical prediction of these networks omit these seemingly multiplied entities, but it would nowhere approach an objective description of reality.

I agree with Dsquared in his reply to Andrew when he says that the reason scientists describe their theories and models as being parsimonious is because of the Occam cult - as though parsimony were some badge of scientific righteousness. Yet, even this description by scientists is merely an attempt to call good old fashioned human simplism something else.

Andrew

"Now, parsimony might dictate that a theoretical prediction of these networks omit these seemingly multiplied entities, but it would nowhere approach an objective description of reality."

Good points in reply to my post. But on this point - in my experience, biologists don't throw out data in the name of parsimony. Usually it's more like they have a favored interpretation of their data, the reviewers say "but what about this other possible explanation" and they cop out by saying "our explanation is more parsimonious." As dsquared said, usually they (we?) get away with it, but no one ever thinks that just because the favored explanation is parsimonious, that it is definitely correct. Hence the other favorite cop-out, "Further experiments will be needed to address this point."

Steve Sailer

Maybe we like Occam's Razor because it's proven pretty useful over the last 700 years?

The alternatives to Occam's Razor turn out to be mostly Occam's Butterknife and Occam's Vegematic.

dsquared

I'm glad to see that great minds think alike on the genuinely important issues, although rather disconcerted to find out that we disagree so violently on the relatively trivial questions.

gene berman

"D"--ol' buddy, I was gonna compliment you likewise but, on reconsideration, thought better of it and took a pass.

I don't think I (or anyone else commenting hereabouts) has any other than a high opinion of your intelligence or of your canny disputational abilities. What is annoying is to see such obvious talents devoted to the perverse worldview you so consistently espouse and defend with arguments of moral equivalence, etc. In the society you seem to advocate, there are only Kool-aid drinkers and Kool-aid preparers; you seem not likely to be one of the former and why anyone would choose to be the latter has no explanation of any complimentary sort--the very least offensive being the opportunity for playing childish games or adolescent exhibitions of one or another types of prowess. The characteristic of this case in which you seem in general agreement with others (with whom you normally dispute) is that it seems not to involve the subject which normally arouses your most fervent and clever defense: criticism of collectivist modes of societal organization and governance.

gene berman

Steve:

I'm surprised at your comment, especially as most of us here couldn't find much difference between Occam's razor, Occam's butterknife, Occam's Vegematic, or even Occam's clamshell.

There's an old saying about how "parsimony is in the eye of the beholder" and, if you ever had a parsnip in the eye, you'd know what I'm talkin' about.

But I'm willing to be made wiser, maybe even somewhat humble. Just name me one problem or phenomenon whose correct solution or explanation has been furnished by resort to Okham's observation. As you've said, the lustre of this pearl has been lighting our way for 700 years. But has it opened even a single oyster?

It would be as well to make a case that the apparent indwelling tendency to upgrade the regard in which an explanation is held the more closely it approaches the description "simple" is the common shortcoming of the naive, the credulous, and the plain stupid. It is with Occam's Razor that conmen part the marks from their money. Both the naive and trusting soul and his opposite, the paranoid mistruster of all--make their errors precisely by reliance on the good Earl's advice.

The only thing I can tell you for sure about Occam's Razor is that, posited in the reverse, in cartoon styling, it sure helped make Rube Goldberg a household name.

NingunOtro

Well, I'm not aware of his skills on criticism of collectivist modes of societal organization and governance, but if it takes all this to argue about a principle as easy as Occams Razor, I think he should seriously do some profitability calculus on the rest of his reasoning processes. Sure he should stop thinking if his loses are bigger than his earnings.

Occams Razor merely predicts that under hostile circumstances (like global evolution, or dialectic argumentation) simple constructs are more probably to maintain integrity and survive than more complex ones. Want an example? Ask yourself why the Kalashnikov is still the number one weapon among the collectivities he so well knows.

The comments to this entry are closed.

Notes for Readers