Talk:Occam's razor/Archive 2
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 |
Tautologies
Wikipedia's definition: Tautology refers to a use of redundant language in speech or writing, or, put simply, "saying the same thing twice".
But this is definitely *not* the same thing as additional elements in a model which doesn't conform to Occam's Razor. The problem with these additional elements is that they *are* meaningful, as the example in science shows. Adding those indetectable creatures results in new predictions, very important new predictions. Thus the additional elemenst aren't simply restatements of existing data or model elements. They are adding new info, which is exactly the problem. We should remove all references to Tautologies in the article, a process which I've begun. —The preceding unsigned comment was added by 69.209.182.55 (talk • contribs) 17:45, March 9, 2006.
This would be a BIG mistake. In philosphy, many phrases have a different meaning. From the Philosophical Dictionary --
tautology
Logical truth. A statement which is necessarily true because, by virtue of its logical form, it cannot be used to make a false assertion.
Example: "If neither John nor Betty is here, then John is not here."
It would be MUCH wiser to alter the Wikipedia definition because IT is wrong and incomplete, than to go back through other people's articles and edit out things that are correct.
I suspect this could be used as a stern warning to people regarding editing other people's work on Wikipedia!!!!
Make damn sure you know what you are talking about. And never consider Wiki the ultimate source. (-This was an unsigned comment)
Maybe I'll take the time to add at least one instance of the term back in at some point... The concepts of redundant tautology and "unnecessary entity" are all but identical - propositions that remain true as long as the rest of the theory remains true, doing away with which leaves the theory intact. --AceMyth 05:05, 16 February 2007 (UTC)
Apologetics and Occam's Razor
Regarding my 02:35, 17 March 2006 edit-- I do not have a substabtial objection to an external link to an aplogetics essay per say. However, such an essay must add something over and above the wikipedia entry. This essay, I contend, does not constitute a substantive addition, so I removed it.
I am willing to engage contrary views. --Johnny Logic 02:53, 17 March 2006 (UTC)
- The article was mine. I did not include it in the article because it was a POV, but there was a substantive section on Occam's Razor's application to religion in the article proper, but not much at all within that section on how a religious person might answer. Paragraphs which sound like they are going to provide that response end dismissing that response. Eg., "However, such arguments are problematic on at least two counts" This then leads to several more paragraphs showing how the religious rebuttal is unjustified. When that whole section is looked at, it is almost completely one-sided. It seemed less work, rather than heavily edit that section, to simply provide an external link where the 'religious' side can have its say.
- One thing in my essay that is not covered in the article at all is the double standard by which a singular God is invoked as a violation of Occam's Razor but the multi-verse, often entertained by the same people invoking it against theism, escapes cut free. This in itself I think would constitute a substantive addition, as it certainly deserves attention in the article proper. Anyway, allowing a response in an external link to a series of objections that are otherwise unanswered seems to be enough to justify its presence, imo. (naturally my essay isn't an exact rebuttal, because it wasn't written with the wiki article in mind) Sntjohnny 03:19, 17 March 2006 (UTC)
While the essay certainly represents a POV, it is inappropriate for the following reasons:
- Factual (e.g. mischaracterizations of abiogenesis and the many-worlds interpretation of QM), conceptual (e.g. stating that the a priori must be true or all reality collapses) and grammatical errors (e.g. Wilhelm of Ockham, get’s, than/then, etc.).
- Neglect of scholarship (e.g. Sophisticated characterizations and justifications of OR abound [see OR sections on Statistics, Science and Justification], however, the author chooses to address his own straw man version).
- Rhetorical, rather than analytical, tone (e.g. “One would think it would be obvious that the argument has shifted, but sadly, you will probably have to throw up your pinky to deflect the razor anyway. They [skeptics] keep it in a little sheaf on their belt, and I suspect they sleep with it under their pillow”).
- Use of fallacies, including straw man, non-sequitur and false dichotomy (e.g. Arguing that some skeptics are inconsistent in their application of OR has nothing to do with whether or not the postulation of God represents a violation of OR) .--Johnny Logic 22:17, 18 March 2006 (UTC)
I'm not really the sort of guy who really gets into the whole Wiki-Wars because they always boil down to the 'victor' being merely the one who is willing to sit there day in and day out to make sure they get the last one and often nothing to do with substance. So, I'm not going to re-post the link, but I am going to respond to the charges of inappropriateness.
- 1. Factual. It is merely your assertion that these are mischaracterizations. It is a brief essay speaking to Occam's razor, where these are invoked as illustrations. Their mention cannot possibly be elaborate enough to fully encompass all the facts about the concepts, but you would be a fool to deny the potential applicability of Occam's Razor to the Many World hypothesis. At anyrate, if I replied to this assertion of yours with "no it isn't a mischaracterization" I will have contributed an assertion with about as much substance as your assertion. Normally, one wants to substantiate one's assertions, don't they? I can't really believe you included grammatical errors in your rebuttal. That is pedantic to the extreme. Really digging, man. Really digging.
- 2. It isn't a strawman version. That of course is YOUR POV. Again, one expects folks to substantiate their claims, but we see that without substantiating that assertion you invoke it in all four of your points. That's bad scholarship right there. But it is true, I don't invoke 'scholarship.' The point of the essay was to provide a perspective reflecting the use of the razor in religious areas, highlighting the hypocrisy of scientism which in very snobby fashion dismisses Intelligent Design or Theism itself because it is not parsiminous while doing far worse in regards to the Multiverse and to a lesser degree abiogenesis.
- 3. It contains plenty of analysis, along with rhetoric. Rhetoric alone isn't enough to make something irrelevant.
- 4. Use of fallacies? Again, that begs the question of whether or not it really is a strawman, which you'd have to demonstrate. While it is true that the fact that skeptics do not use it consistently does not necessarily- necessarily in its technical sense- mean it is an irrelevant objection to theism, the fact that the principle is only used by skeptics when applied to theism and very rarely in regards to other examples... again, the multiverse, which practically begs for an application of the Razor... surely is relevant to this wiki entry, as nearly every non-theist I've ever met who invoked Occam's Razor had no problem tolerating the Multiverse. In the context of the use of the razor in areas of theism, which this wiki entry addresses, my essay is perfectly appropriate.
However, you sir have decided to suppress a certain point of view instead of letting people make up their own minds, which is against the spirit of wikipedia. Sntjohnny 18:33, 19 August 2007 (UTC)
When to choose the tougher choice?
I understand the whole concept of Occam's Razor, and for the vast majority of all times I would agree with it's use. My question is more from an anthropological type of scenario:
If you have a choice to make regarding how you are going to deal with a situation involving someone whom you know to be wrong beyond the shadow of a doubt. Your options are bountiful, but you resolve it down to two. The easiest one, and the hardest one. The easiest one being that you choose to let it go, not say a word, blow it off and sit and stew and allow yourself to continue getting mad. And the hardest choice, to confront the person, tell them how they were wrong and earnestly be willing to help them change their faults.
In this situation, the easiest choice is not necessarily the best option. You may end up doing yourself more harm in the long run and allowing the other person to continue doing harm to themselves and everyone around you. The harder choice, in my opinion would be the best option, by allowing the person to be able to know something about them that bothers you and or other people.
What do you think? Can and should this kind of logic devoloped by philosophy be used broadly across the board? Thewhippo 07:38, March 23, 2006 (UTC)
- Occam's Razor does not deal with concepts like "easy" or "hard. It deals with simplicity concerning theories. It does not provide *moral* answers except so far as one system ethics can be said to be simpler than another and explain the data equally well. —The preceding unsigned comment was added by Palthrow (talk • contribs) 17:07, March 23, 2006.
Ockham's Tazor
JA: You hate to see a really good typo go. Sounded like a policy of truly polite police: Go ahead, carry a taser, but use it only when you really, really, really need it. Jon Awbrey 17:38, 23 March 2006 (UTC)
Kierkegaard
I removed this text from the 'in religion' section: "On the other hand, Kierkegaard argued that there were no testable predictions of the existence of God and further argued that the concept of faith made any testable observations self-defeating. It is difficult to explain humankind's unique understanding of good and evil and its ability to love and hate -- relative to the rest of the animal kingdom -- from a purely evolutionary standpoint. In this sense, it is not reasonable to simply combine all of the material observations of our universe and apply Occam's Razor to justify the non-existence of God. Indeed, William of Ockham himself did not make this leap, being himself very well educated in religion."
My reasoning was primarily that it makes no sense in the context of Occam's Razor. The whole point of Occam's Razor is that you don't need to falsify a theory to discount it. For every set of data there are an infinite set of unfalsifiable theories. If you have no way to choose between those theories you have no basis on which to believe anything in particular. So restating the fact that God is nonfalsifiable (which is already strongly implied by the previous discussion, if God were falsifiable then Occam's razor would be irrelevant in the discussion) is completely unecessary.
The rest is clearly not NPOV. Most anthropologists would disagree that it is hard to explain humankind's unique understanding of good and evil from an evolutionary standpoint. Also this general idea has already been discussed in an earlier paragraph about the comparison psychological theories and God based theories. —The preceding unsigned comment was added by OverZealousFan (talk • contribs) 03:32, April 8, 2006.
Occam or Ockham? — Datapoints
Which spelling do other reasonably authoritative sources prefer?
Here's a list of what a quick look through my bookshelves and some online searching came up with. Other readers may like to extend it... Jheald 00:47, 13 April 2006 (UTC)
Occam
- (1836-7) William Rowan Hamilton, Lectures on metaphysics and logic, sec. xxxix; published 1859, vol II p. 395: "We are therefore entitled to apply Occam's razor to this theory of causality". (The earliest citation quoted in the OED (1989); the phrase is presumably introduced earlier in the lectures)
- (1898) Brewer's Dictionary of Phrase and Fable. "William of Occam", "Occam's Razor".
- (1911) Britannica. "William of Occam". ("Occam's Razor").
- (1946) Bertrand Russell. History of Western Philosophy. ("William of Occam", "Occam's razor").
- (1965) Funk and Wagnalls New Standard Dictionary of the English Language. ("Occam's razor", "William of Occam").
- (1970) Brewer's Dictionary of Phrase and Fable. ("Occam's Razor", "William of Occam").
- (1981) Concise Oxford Dictionary of Quotations. "William Occam", "Occam's Razor".
- (1988) Gull, S.F. "Bayesian inductive inference and maximum entropy". In Maximum Entropy and Bayesian Methods in Science and Engineering, Volume 1: Foundations, ed. G.J. Erickson and C.R. Smith. ("Occam's razor", "Occam factor".)
- (1988) Stephen Hawking. A Brief History of Time. (Occam's razor).
- (1989) Oxford English Dictionary (2e). (Occam's (also Ockham's) razor).
- (1991) MacKay, David. Bayesian interpolation. Neural Computation 4 (3): 415-447. ("Occam's razor", "Occam factor").
- (1992) P.C.W. Davies. The Mind of God. ("Occam's razor").
- (1997) Deutsch, David. The Fabric of Reality. (Occam's razor).
- (1998) New Penguin Dictionary of Science. ("Occam's razor (Ockham's razor)).
- (2003) Jaynes, E.T. "Probability Theory: The Logic of Science". ("William of Occam", "Occam's razor").
- (2003) MacKay, David. Information Theory, Inference, and Learning Algorithms. ("Occam's razor", "Occam factor").
- (2003) Oxford Dictionary of Statistical Terms. ("Occam's razor", "William of Occam"). OUP/International Statistical Institute.
Ockham
- (1912) Catholic Encyclopedia. "William of Ockham". ("Ockham's Razor".)
- (1961) Jeffreys, Harold. Theory of Probability. ("William of Ockham's rule", "Ockham's Razor".)
- (1961) Webster's 3rd New International Dictionary. ("Ockham's razor", or "Occam's razor")
- (1962) Kneale & Kneale. Development of Logic. ("William Ockham", "Ockham's razor".)
- (1962) Runes. Dictionary of Philosophy. ("William of Ockham", "Ockham's razor", "Ockhamism".)
- (1983) P.C.W. Davies. God and the New Physics. ("Ockham's razor").
- (1991) Garrett, Anthony. "Ockham's Razor". Physics World 4: 39-42.
- (1991) Jefferys, William H. and James O. Berger. "Sharpening Ockham's Razor On a Bayesian Strop".
- (1994) Scruton, Roger. Modern Philosophy. ("Ockham's razor", "William of Ockham" (sometimes written Occam)).
- (1994) Oxford Dictionary of Philosophy. Ockham's razor.
- (1995) Oxford Companion to Philosophy. "William Ockham", "Ockham's razor".
- (1999) Collins Concise Dictionary. "Ockham's razor". (Also spelled "Occam's razor".)
- (2002) Oxford Dictionary of Statistics. "Ockham's razor".
- (2003) Molé, Phil. "Ockham's Razor Cuts Both Ways". Skeptic 10 (1): 40-7.
- (2004) Beck, James L. and Ka-veng Yuen. "Model Selection Using Response Measurements: Bayesian Probabilistic Approach." Journal of Engineering Mechanics 130 (2): 192-203. ("William of Ockham", "Ockham's razor", "Ockham factor".)
- (2004) Kelly, Kevin T. "Justification as Truth-Finding Efficiency: How Ockham's Razor Works". Minds & Machines 14 (4): 485-505.
- (Current) Britannica. "Ockham's razor". (Also spelled "Occam's razor"; "William of Ockham", also spelled "William of Occam".)
- (Current) MacTutor: William of Ockham, Ockham's razor [1]
- (Current) Modern placename: Ockham [2]
- (Current) Stanford Encyclopedia of Philosophy. "William of Ockham". ("Ockham's Razor".)
Historians of ideas appear currently to overwhelmingly use the spelling Ockham when referring to the man himself, if the list of references at the MacTutor site is correct. [3]
Split decisions
- (2004) Standish, Russell K. "Why Occam's Razor". Foundations of Physics Letters 17 (3): 255-66. ("William of Ockham".)
My personal preference would be for Ockham's razor, and William of Ockham; despite the considerable impact that David MacKay's paper (building on the earlier paper of Steve Gull, and argumentation by E. T. Jaynes before that) has had, particularly in the machine learning community.
"Occam" was clearly widespread in the late 19th century; but it looks as though "Ockham" became the dominant spelling, at least in the philosophy literature, in the 20th century. Jheald 00:47, 13 April 2006 (UTC)
JA: I'm guessing that this case is probably analogous to other Scholastics writing in Latin, who used Latinizations of their original names. I'll use "Occam" on the condition that we go back to calling Descartes "Cartesius", and so on. Most scholarly literature that I've seen uses "Ockham". Let's make it formal and put in a name change request. —The preceding unsigned comment was added by Jon Awbrey (talk • contribs) 03:12, April 13, 2006.
There seems to be minor edit-warring about the spelling in the article on this. Since the page move was declined in April, and the page remains as Occam's Razor, I reverted the spelling in the article to match. References to "William of Ockham" seem to be appropriate. NOW ... just because teh OED prefers Occam's as the primary spelling, doesn't mean it's right. I concur with the statement above that the 20th century use of Ockham's Razor is valid. Perhaps a new page move request, with the datapoints listed very concisely above (thank you, Jheald), is in order; having it one way for the theorem, and another for the man, is silly. But until then, let's NOT go back and forth over and over on the spelling in the article. It makes the article look less than authorititative that it can't even get a spelling done consistently (that is, Occam's Razor in one paragraph, and Ockham's Razor in the next)(and NO, one should NOT change spellings of referenced works). David Spalding (☎ ✉ ✍) 16:36, 16 January 2007 (UTC)
- Hi, I reverted some of your changes because I thought they were referencing "William of Ockham" explicitly. That's what I meant by "um, used Ockham when it's clear it's William of Ockham." --Kjoonlee 17:24, 16 January 2007 (UTC)
- Good catch! David Spalding (☎ ✉ ✍) 18:29, 16 January 2007 (UTC)
Ockham's Razor
JA: The great consensus in all of the more informed literatures is that Ockham's razor is an aesthetic, economic, heuristic principle, at most ancillary to scientific method and not fundamental to it, too vague in practice to settle any real dispute, frequently violated, and in the actual history of science almost never, if ever, decisive in the choice between competing theories. Jon Awbrey 05:20, 8 April 2006 (UTC)
- Sorry JA but you are flat wrong about this. All you need to do is READ the section to realize that. Give it a serious read, and then explain how science can proceed in any instance without Occam's Razor. Look at the underdetermination issue. That is not a trivial issue which can be swept under the carpet. I think what you are noticing is that we use Occam's Razor IMPLICITLY much of the time we use it. We don't even explicity consider the infinite number of alternative theories because we depend on Occam's Razor to simplify the analysis. However, without a principle like Occam's Razor we have no way to choose between the infinite number of competeing and radically different theories, and science ceases to function entirely. Its important to note then, that what is really happening is that Occam's Razor is being used so frequently and so extensively that you don't even realize that its being used. Nevertheless, its important to realize on a philosophical level.
- Also note that if you were right, most of the section should be eliminated because it dirrectly contradicts your understanding, so your edits would be inadequate. But you are clearly wrong so it doesn't matter. Also, just incase it seems like I'm being a little extra abrasive, I probably am responding to your quip about OverZealous fantasy in your comments. Since you are editing an article that has existed in this form for a while, and since you clearly don't even have a grasp of the ideas already here, it seems like you should be a bit more humble. OverZealousFan
JA: I have summarized as ably as I could in a few short lines the gist of the last hundred years or so of literate discussion on Ockham's razor. If you dispute my summary, then you are entitled to ask for citations, and I will supply them. If you feel that I have missed signficant material from significant sources then please add whatever material you see fit. If I do not recognize the significance of the material you add, or think that the reader may not regard it as common knowledge, then I will ask for citations. But I will remove unsourced nonsense with minimal notice. Jon Awbrey 03:26, 12 April 2006 (UTC)
JA: OverZealousFan (OZF) is missing the point that there is yet to be given an explicit operational formulation of the ostensible criterion in question. Until there is, it is simply nonsense to say that the TBA priniciple is implied in every choice among theories. There is no principled principle to speak of yet. Jon Awbrey 04:07, 12 April 2006 (UTC)
- How do you believe that science deals with the problem of underdetermination then? Even to get to the point where you are choosing between a finite number of theories you first must rule out the infinite derivatives of each theory, which all fit the evidence equally well and all give radically different predictions. Its not nonsense to say that Occam's Razor is used implicitly its fact. In formal logic it is not hard to give a precise definition of Occam's Razor. There are also numerous other formal definitions in statistics and elsewhere. However, formal or not, if you are really going to hold on to your ridiculous and unsupported claim that Occam's Razor is not used implicitly in every instance of scientific inquiry, then please tell the rest of us how you believe that the problem of underdetermination is solved in science? Answer this one question reasonably and I'll shut up forever, and you'll become a world famous philosopher, etc.
- Just to make sure you don't ignore the question again, I'll repeat. If Occam's Razor is not used implicitly every time a scientist proposes a theory, then how are scientists dealing with the problem of underdetermination.
- One more time to make it perfectly obvious to everybody who reads this. There are infinite variations of every base theory which all fit the evidence as well as the theory, and all of which make conflicting predictions. There is NO scientific theory which does not have this problem. If scientists are NOT always using Occam's Razor implicitly than on what basis do they select among those infinite theories?
- I think a good faith effort of discussing this requires that you at least TRY to answer the question. Give it a try. How are they doing this magic?
- Of course you won't be able to come up with an answer because you are wrong. But at least your lack of an answer should be evidence to everybody who reads this discussion.
JA: Sorry if I've missed some of your remarks but I have been traveling the last three weeks and using begged and borrowed computers at conference sites and airport kiosks. But I think that I have already responded in some detail to the question about underdetermination on the Philosophy of Science talk page. Please let me know if I have missed other remarks. Jon Awbrey 03:44, 13 April 2006 (UTC)
JA: For ease of reference, my discussion of underdetermination is found under Point 3 at Talk:Philosophy_of_science#Material_for_examination. Jon Awbrey 03:50, 13 April 2006 (UTC)
- If I understand your point on that talk page, Jon, it's that infinitely many heuristics are capable of solving the underdetermination problem; Occam's Razor doesn't have to be the chosen heuristic, so it's not fundamental to the scientific method. Now that I understand you (I think), I agree, and I've changed the first sentence in the In science section to this: "Occam's Razor has become a basic tool for those who follow the scientific method." (It was "basic perspective" before, and had been December 31, 2003, but "tool" works better with the razor metaphor.) — Elembis 06:56, 13 April 2006 (UTC)
I agree, the under determination argument for the scientific utility of Ockham's razor is nonsense. The only way to legitimately eliminate a contending scientific explanation is to show that it differs in its predictions from the other explanations and then look to nature as to how well those predictions agree with reality. In cases I am aware of where there was no difference found in the predicted results of competing explanations one or the other was not discarded based on some claim of simplicity or "parsimony". It was shown that the competing explanations were different representations of the same underlying idea. It sure would be nice if advocates of Ockham's razor would apply Ockham's razor to Ockham's razor. It is unnecessary to science. Gkochanowsky 19:11, 7 February 2007 (UTC)
- The following discussion is an archived debate of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.
The result of the debate was no consensus. —Nightstallion (?) Seen this already? 17:16, 18 April 2006 (UTC)
Requested move
- Occam's Razor → Ockham's razor … Rationale: More standard usage and capitalization, consistency with related articles … Please discuss/vote at Talk:Occam's Razor — Jon Awbrey 03:18, 13 April 2006 (UTC)
Survey
- Add *Support or *Oppose followed by an optional one-sentence explanation, then sign your opinion with ~~~~
- Support move to "Ockham's razor". The spelling "Ockham" accords with standard current usage in most scholarly literature. Jon Awbrey 03:34, 13 April 2006 (UTC)
- Support. Unless "Occam's razor" is significantly more common than "Ockham's razor", which it isn't, the razor should be spelled like William's name (which is almost always spelled "Ockham"); also see Wikipedia:Naming conventions (capitalization). — Elembis 05:25, 13 April 2006 (UTC)
- Current count from Google: "Occam's razor" -"Ockham's razor": 723,000 "Ockham's razor" -"Occam's razor" 676,000 both: 17,500. In December 2002 the count was 27,300 for the first search, 19,100 for the second. Jheald 07:32, 13 April 2006 (UTC)
- Support move to "Ockham's razor". You can transliterate it however you like, but (I feel like a fundamentalist here) well, you already know how the arguments go about this. Lacking a compelling argument to revise to some "modern" form, or a compulsion to feel like one is on the faux leading edge, the original and most commonly used transliteration is preferable...Kenosis 05:40, 13 April 2006 (UTC)
- Oppose How come when I type into Google "ockham's razor" (with or without quotes) it asks me "Did you mean: occam's razor"? But no question is asked if I type in "occam's razor". BTW, over here Google is reporting: about 783,000 for occam's razor; about 747,000 for "occam's razor"; 702,000 for ockham's razor; about 694,000 for "ockham's razor". Ewlyahoocom 11:50, 13 April 2006 (UTC)
- Support. GHits are too close to be used as a guide in this case. How did Wm. of Ockham himself spell his name/hometown? Today the village is officially spelled Ockham. AKADriver 13:59, 13 April 2006 (UTC)
- All of his writing was in Latin and English spelling of the time was not standard, so I don't think he can answer the question for us. Jonathunder 15:37, 13 April 2006 (UTC)
- Support. It should conform with the philosopher's name in the main article--Aldux 16:02, 13 April 2006 (UTC)
- Oppose. I agree that Ockham is more accurate than Occam. Even modern ecclesiastical Latin follows this practice. However, Ockham himself spelled his name in medieval Latin, as Occam, and most Romance languages (see Spanish, Italian, and Portuguese Wikipediae, for instance) kept the Occam spelling in this expression. Spanish and French Wikipediae do a weird thing: they crossreference "Guillermo de Occam" to "Guillermo de Ockham", and "Navaja de Ockham" to "Navaja de Occam". These anomalies show that the expression "Occam's razor", thus spelled, became part of many langages' jargon, even though the modern practice favors the Ockham spelling in the philosopher's name. Louie 18:04, 13 April 2006 (UTC)
- Oppose after reading the previous page move discussion. The traditional spelling is "Occam" and usage has not changed overwhelmingly. Jonathunder 18:32, 13 April 2006 (UTC)
- Oppose The principle is more commonly known as "Occam's Razor". Besides the main Google web search, look at the difference in hits on Google News (15 Occam vs. 1 Okham), Google Groups (98,400 Occam vs. 13,300 Ockham), and Google Scholar (4,150 Occam vs. 1,150 Ockham) older ≠ wiser 14:30, 14 April 2006 (UTC)
- Oppose The fundamental point is that spelling was not standardised when William was around anyway, so they're both correct. Since we have to pick one, google has more google hits of occam.WolfKeeper 16:41, 14 April 2006 (UTC)
- Support. I don't care about how it's spelt in other languages. (Plus we're currently telling people that Occam's Razor is named after William of Ockham, which is just silly.) Proteus (Talk) 18:46, 14 April 2006 (UTC)
- I think you didn't get my point. The point is: most modern languages (even ecclesiastical Latin, if that's modern ;-) tend to follow the modern spelling in William of Ockham's name, yet still preserve the expression "Occamis novacula", which may be archaic, but deeply entrenched as to be preserved.--- Louie 17:43, 16 April 2006 (UTC)
JA: And you are missing the point that Late Latinization of an English name forces that choice, for the lacc of a k in the preferred orthography of Late Latin. Jon Awbrey 17:56, 16 April 2006 (UTC) Cf.
- Actually, where did the 'ckh' come from, anyway? I remember a similar case: John Peckham's name latinized as "Pecham". This spelling seems more consistent with the actual English pronuntiation and Latin spelling at the time than the weird-looking (and perhaps incorrect) 'Occam'.--- Should I read here 'Ocham'? I bet there are instances of this spelling.--- Perhaps the 'ckh' spelling is a correction on the confusing "ch" inherited from Latin spelling: read "Pecham" in modern English pronunciation, against "Peckham"; it seems that the intermediate 'k' corrects the tendency to read the Esperantian ĉ (like in 'church') here. Could anybody confirm this conjecture? Louie 19:56, 17 April 2006 (UTC)
- Oppose Occam seems to be the traditional phrase --Henrygb 02:20, 15 April 2006 (UTC)
- Weak oppose. Both are extremely common, but I've found that "Occam's" is slightly more common for the razor itself (even though I personally prefer "Ockham's"). The main rationale for moving seems to be that it makes it more consistent with the name of the person, and that we'll look "silly" for saying that something named "Occam" was named after something named "Ockham", which are both rather weak arguments. Wikipedia's naming conventions were created precisely for this purpose: to avoid warring over which name is "better" (an abstract and ultimately empty concept), and instead simply determine which is more common. -Silence 20:59, 16 April 2006 (UTC)
Discussion
- Note: This move was discussed extensively, but without consensus, in 2002; see Talk:Occam's Razor/Archive 1#Article naming issue.
JA: We are talking about informed use, which is not always the same as popular use. When I type "Peirce" into Google, it says "Did you mean: Pierce?" It used to do that even when I typed the exact phrase "Charles Peirce", but it's gotten smarter in the last year or so. There's a lesson in that. Jon Awbrey 12:04, 13 April 2006 (UTC)
I think Google keeps the Did you mean algorithm a trade secret, so good luck reverse engineering it.
- I know how it works in broad terms. Google watches what the users search on. If they do a search, and then immediately do another search with a different spelling of the same word, then it gets offered next time as 'did you mean'. So it's the users helping each other out. :-)WolfKeeper 16:46, 14 April 2006 (UTC)
But I'm not sure it even matters how Mr. Ockham spelled his own name. Did he ever refer this principle as a razor? Shouldn't we rename it "Law of parsimony" or "Principle of economy" or something even more historically accurate? Ewlyahoocom 15:34, 13 April 2006 (UTC)
I think you should keep the Occam spelling, at least for the sake of consistency with other Wikipediae. The pervasiveness of the Occam spelling throughout Western European languages in this construction suggests that this spelling is somehow standard, even though modern practice favors spelling Ockham in Brother William's name. Louie 18:10, 13 April 2006 (UTC)
JA: Many people are missing the fact that the spelling "Occam" is the Medieval Latinization of the English name "Ockham", and that this spelling is forced by the fact that there is no "k" in the preferred orthography of Late Latin. Since this is the English WP, I think that we should use the English instead of the Latin spelling. Jon Awbrey 18:18, 16 April 2006 (UTC)
JA: Cf.
- This being the English Wikipedia, we should use whatever spelling is most common -- which appears to be Occam for the principle. older ≠ wiser 19:13, 16 April 2006 (UTC)
- The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.
GR & QM
I think the following needs a bit of work: "But in the 20th century, general relativity and quantum mechanics complicated scientific views of nature, dealing a blow to the razor's metaphysical footing."
I'm not sure that GR / QM disobeyed the principle of parsimony, and I think the "dealing a blow" statement is value-laden and mildly POV against. What's the definition of complexity inherint in this statement, and why would GR be considered "more complex" than Newtonian mechanics? If the argument is that "because GR is taught in grad-school and Newton is taught in undergrad" then I think that would be insufficient. Thanks, --209.128.81.201 16:40, 16 April 2006 (UTC)
- I wrote that sentence, and I intended it to mean that modern science isn't as elegant and simple as people expected it to be, not that GR/QM violate Occam's Razor in any way. Occam's Razor was once supported primarily with the idea that simpler things were more likely to be true — that's what "metaphysical footing" means — but sometimes the truth is complicated. For reference, here are the three sentences that my sentence replaced: "Prior to the 20th century it was believed that these associations are the very justification for the razor — that nature was in some sense simple and that our theories about nature should reflect that simplicity. With such a metaphysical justification came the implication that Occam's Razor is a metaphysical principle. From the beginning of the 20th century, these views fell out of favor as scientists presented an increasingly complex world view." — Elembis 00:37, 18 April 2006 (UTC)
- Maybe it's original research for me but I doubt the metaphysical standing of Occam's razor isn't in question but only the definition of what counts as "simple" vs. what counts as "complex." Newtonian mechanics is pretty complex I guess if you're living in the 18th c. and don't know the calculus. But I understand I think the concern, or the sentiment behind your edit and the previous, if we take the "nature &it is simple. Will people living in the 23rd c. argue that QM is too complex or more complex than say (insert counter-theories to QM here?) Don't know. Thanks, --63.138.93.195 02:29, 21 April 2006 (UTC)
I have to say I'm glad that this section is present in the discussion, because the respective statement within the article itself is, in my view, an extremely provocative and controversial statement. In fact, I would suggest that it be removed, seeing as how, in reality, Quantum and Relativity theories are (in my opinion, anyway) shining examples of good scientific theories that obey/pass Occam's Razor. To say that these theories bring Occam's Razor into question demonstrates (as far as I can tell) a lack of understanding of the razor or the theories or both. But maybe it's just myself who doesn't understand. I'm hesitant to remove the statement myself, but I would, personally, advocate its removal. Does anyone else agree? - 61.9.204.168, 22nd April, 2006
- I understand the concerns you two (three?) have, and I have removed the sentence in question. The point (as I understood it) was never that GR/QM violated the razor or were poor scientific theories. But a misleading explanation is worse than none at all (and I don't think my edit was any more misleading than the one it replaced). The fact that we didn't have a source for the explanation made it even worse. Thanks for the comments. — Elembis 05:52, 22 April 2006 (UTC)
- Thanks Elembis, and thanks 61.9. BTW, 209.128 and 63.138 (and me right now) are 1 person. 61.9 is someone else. Maybe I should create an account. At any rate in my mind the original statement(s) has some validity if presented as a counter-argument to the (Victorian?) idea that "Nature's rules BY FIAT are simple, and cognitively accessible." I can't think of a metaphysical reason why that should be so, and a strict interpretation of the razor might imply that. But thanks.--63.138.93.195 01:35, 23 April 2006 (UTC)
- I think the sentence should stand. That is because you can't make a claim about the importance of Ockham's for selecting explanation X unless you also show all the other explanations {A,B,C...} that were not preferred and how they failed Ockham's razor in some way. Fact is none of the current crop of explanations were preferred for anything remotely resembling Ockham's in any form you care to state it that William of Ockham never wrote. Gkochanowsky 02:56, 25 June 2007 (UTC)
Regarding QM I was surprised to see no reference to Interpretation_of_quantum_mechanics, as that seems an excellent current example of scientific theories with competing ontologies but equivalent predictions.
Also the unspoken assumption of Occam -- that it is necessary or wise to choose just one theory out of a set of alternatives -- can legitimately be questioned. I know David Bohm, with F. David Peat, expounded on this in Science, Order, and Creativity, starting from a commentary on Thomas Kuhn's well-known "Structure of Scientific Revolutions". Bohm proposes that keeping two or more theories concurrently "in play" is often to be encouraged. -Chuck 209.162.144.2 22:57, 7 October 2007 (UTC)
Kneale & Be Humble
JA: The Kneale & Kneale paragraph is an invaluably succinct summary of three different variations on the so-called razor, along with their actual provenances, from a standard reference on the history of logic, and serves to correct a large number of historical distortions in a very compact space. Yes, the general scholarly estimate of Ockham's power and influence as a logician is less than the cult of Occam would like to admit, but it is sourced from an authoritative source, and at least the Kneales were not making it up out of their own heads week by week. Until you can say the same, leave it be. Jon Awbrey 23:12, 17 April 2006 (UTC)
- We have a Variations section for variations. If "numquam ponenda est..." and "frustra fit per plura..." express the same idea as "entia non sunt...", why should they be mentioned in the History section? Do you think a new phrasing constitutes a meaningful historical event or a new way of thinking about the razor? Do you think two untranslated Latin variations of "entia non sunt..." help readers understand the history of the razor? Or do you simply think "numquam ponenda est..." does have a different meaning? Any explanation would be nice. And while we're on the topic, would you kindly list the "large number of historical distortions" the Kneales' quote sets right? I still don't understand why anything more than the second sentence of it belongs in the History section. — Elembis 00:20, 18 April 2006 (UTC)
JA: The quote is valuable for the amount of detailed information that it summarizes in a very compact form, relating the more apochryphal popular statement to the better documented extant statements, and it is not taken for granted that all of them are saying the same thing. Nor is it automatic that the common interpretantion as a heuristic of science is justified to read into what is basically a maxim of nominalism. Given the number of retrospective readings of anachronistic ideas into Ockham's mottos, and I have seen a large number of truly outrageous fictoids of late, it is crucial to have accurate and complete data on this score. Smudging the problems of interpretation into oblivion is a distortion of the facts. Jon Awbrey 13:45, 18 April 2006 (UTC)
- With all due respect, I think you're still just waving your hands without actually explaining how the Kneales' quote (KQ), aside from the second sentence, offers important information from a historical standpoint. (Again, unless "numquam ponenda est..." means something that "entia non sunt..." doesn't, History doesn't need to quote it, and it belongs in Variations instead.) You mention interpretations relating to science and nominalism, and interpretations (especially Ockham's own) are surely historically relevant, but KQ doesn't mention interpretations except to say that "entia non sunt..." "[n]o doubt . . . represents correctly the general tendency of [Ockham's] philosophy", and I included that line in my edit. You also mention "retrospective readings of anachronistic ideas into Ockham's mottos", but you don't explain what these poor readings are or how KQ addresses them at all. It's easy to say something is important, but a clear explanation is needed, and broad statements about "outrageous fictoids" are no substitute. Please state, without resorting to vague and abstract language, the specific historical insights that your preferred presentation of KQ makes. — Elembis 19:15, 18 April 2006 (UTC)
- Requested a clear and concrete explanation of the quote's historical value
JA: I believe that I gave a clear and concrete explanation of the value of that specific quote. I said "The Kneale & Kneale paragraph is an invaluably succinct summary of three different variations on the so-called razor, along with their actual provenances, from a standard reference on the history of logic, and serves to correct a large number of historical distortions in a very compact space". I personally find having three variants in paradigmatic comparison very useful, or I would not have spent the hours that I did trying to remember where I read it.
JA: The chief historical distortions are these: (1) To cite a version of the principle that is not found in the extant works of Ockham as "Ockham's razor". (2) To replace concrete statements with "general tendency" paraphrases. Even if we all currently agree that the paraphrase captures the gist of someone's philosophy, we might turn out to be wrong in some important detail, and it's best to keep the exact statements ready to hand as a check. (3) More controversial, perhaps, but still deserving a voice: Many practicing scientists and philosphers of science do not see nominalism as a cornerstone of modern science, but something that can block its growth if taken to extremes. To confound a heuristic guideline about simplicity in theory, which is a modern reading of the maxim, with a maxim of nominal thinking, that was its original sense, is a serious philosophical confusion. Jon Awbrey 20:00, 18 April 2006 (UTC)
- Thank you, Jon. Now that I understand your reasons for preferring the full quote, I can improve the History section (which I think should describe how the razor has been used and promoted throughout history, especially by Ockham, Scotus and Aristotle) and better recognize similar problems in the future. — Elembis 22:06, 18 April 2006 (UTC)
Near perfect example of a bad example
JA: Nothing's perfect, but this example is so bad that it affords us a near perfect example of what's wrong with some popular but half-baked misreadings of Ockham's razor vis-a-vis scientific method. But it's late, so will discuss tomorrow. Jon Awbrey 03:54, 29 April 2006 (UTC)
For example, it's easy to think of alternative theories to Newton's famous statement that every action has an equal and opposite reaction. One such theory could be that for every action there is an opposite reaction of half intensity, but undetectable creatures magnify the reaction so it appears to be equal in magnitude; these creatures will all die in the year 2055, at which point the observable nature of the universe will shift. (Variations of this theory might mention 2056 instead, or 2057, and so on to infinity.)
Until 2055, the creature hypothesis will appear identical to Newton's law, leaving a scientist with no way of telling which, if either, is right. This is the problem of underdetermination: For every set of data, there are infinitely many theories which are consistent with it. Without a method to choose between these competing and theoretically possible theories, science ceases to function entirely and becomes useless for all practical purposes, and perhaps even useless in principle. Occam's razor is such a tool, expressing a preference for theories with fewer unsupported elements. Thus, it deems the third law of motion in its original form preferable to the version with the added undetectable creatures.
- Well, that wasn't very constructive. "Bad", "wrong" and "half-baked" Why? By what standards? I'm obviously missing something, because I can't imagine how a heuristic drawing on probabilistic justifications could be of use in any other way aside from precisely the aforementioned kind of situation. Or, for that matter, how that kind of situation could be resolved by anything other than a heuristic drawing on probabilistic justifications. Yes, I am aware that you take objection to the citing of anything but the pure maxim of nominalism originally invoked by Occam as "Occam's Razor", but just tossing all the modern variations aside solely due to them being modern would be grossly prescriptive and unbefitting of an encyclopedia. --AceMyth 05:03, 29 April 2006 (UTC)
- If I can hijack a statement by Imre Lakatos: "These scientific games are without genuine epistemological content unless we superimpose on them some sort of metaphysical principle which will say that the game, as specified by the methodology, gives us the best chance of approaching the truth." (The Methodology of Reseach Programmes [Cambridge: the University Press, 1980], p. 122). The razor appears to be exactly one such metaphysical principle. Whether it is correct or not, the assumption that the simplest explanation of the facts is the most likely to be accurate provides a criterion for selecting between competing sufficient hypotheses when other criteria are unhelpful. » MonkeeSage « 09:04, 29 April 2006 (UTC)
JA: Okay, I had a good night's rest. I have this feeling that I'm going to need it. I thought we had discussed this issue several times already, but maybe that was somewhere else. I will have to integrate this by parts, as they say in maths. Jon Awbrey 13:26, 29 April 2006 (UTC)
JA: 1. First off, the various principles that fall under the rubric of "Ockham's razor" (OR), as commonly interpreted, are very strictly speaking "aesthetic principles", at most being used as "regulative principles". They are not even formulated in a way that could justify calling them anything else, they have no a priori warrant except human convenience, which is to say, the motive of "wishful thinking" that reality be simple enough for us to grasp. On critical examination, the assumption in question is more pleasing than plausible, however necessary in pragmatic terms as a hope that permits inquiry to proceed undaunted. Jon Awbrey 14:15, 29 April 2006 (UTC)
JA: 2. Not that it necessarily matters what I think, but I do not personally dismiss the value, nor even discount the necessity, of aesthetic, economic, heuristic, normative, pragmatic, or regulative principles in scientific inquiry. But they have to be recognized for what they are, and not confounded with axioms, definitions, empirical facts, or scientific laws. Jon Awbrey 17:10, 29 April 2006 (UTC)
JA: 3. All of the literature on heuristic principles, of which the motley assortment of rules-of-thumb known as the OR fill but one tray in the toolbox, observe that they tend to come in opposing pairs — "neither borrower nor lender be", etc. — indeed, this is one of the marks of a maxim that can do no more than call attention to a particular axis of frequently pertinent concerns. Regulative principles like these are simply never formulated determinately enough in operational terms to be tested by experiment, by inductive, quantitative, statistical measures. The only kind of test they ever get is summed up in observations like this: "The species of creatures who habitually acted on instinct X eventually died out of terminally wrong hunches".
JA: 4. Sorry for the delay in attending to the specific example, but it seemed like a good idea to review a number of general considerations affecting the regulatory status of economic principles in science. Let's now get down to the brass taxonomy of the example in hand and see what we can learn from what I claim are its outstanding defects. First off, one might well ask: Why was it necessary to concoct such an ad hoc-ey example in the first place? Are they really no compelling examples from the history of science where some razor variant was actually used in a critical and decisive manner to advance our scientific grasp of some phenomenal domain? And the answer is: No, there aren't all that many non-fictional examples. To tell the truth, I can't even think of one. No doubt some people will be shocked at this. Still, there it is. Jon Awbrey 16:40, 30 April 2006 (UTC)
JA: Defect 1 = the ad hoc, cooked up, fictional character of the example. It was most likely made up on the model of Nelson Goodman's grue/bleen example in his "new riddle of induction. We can discuss what that was about in a while. I have noted the fact that it is actually tough to come up with genuine examples in the history of science where anything like a razor principle was crucial in deciding what actually transpired in the choice of theories. This brings up another problem that affects this example, along with the lion's share of the retrospective examples that one commonly sees mentioned.
JA: Defect 2 = the post hoc, reconstructive, retrospective, revisionary-historical character of most textbook examples that one sees. One of the big lessons that they try to drill into your head in courses in both historical method and statistical method has to do with the illusions of post hoc analysis. A similar issue comes up with the criterion of predicative power in science, that everybody knows is important. It's very easy to predict things that have already happened. It reminds me of that unctious uncle in Pearl S. Buck's The Good Earth, at least in the movie version that was on TV all the time when I was a kid. He was always saying "I predicted it!" about some important turn of events, but after the fact, when the fact was that he had usually predicted the exact opposite. It's the same with post hoc analysis and explanation. It's all too easy to tailor-make some measure of simplicity that will fit whatever style of culottes are dictated by any old revolution that has already happened. In the US they call that "Monday morning quarterbacking". For more general consumption let's call it the Elastic Razor. As you might guess, it's double duty, but doesn't really cut the mustard. Jon Awbrey 18:24, 30 April 2006 (UTC)
GK: This also brings up the point that in science the preference in explanations goes to the explanation that does the best at predicting that which has not been discovered. Not that which satisfies Ockham's razor (whatever that is). New discoveries predicted by explanations before the fact are usually what gets you the noble prize in theoretical physics. Too bad the prize is not given to people who excel at predicting that which has already been discovered. We would all be Nobel Laureates. In the upcoming bake-offs between the various cosmological explanations the deciding factor will have nothing at all to do with Ockham's razor. If it did then it would not be science.
- So you're, in essence, saying that what makes the "alien" example a bad example isn't its conclusion but rather that it presents it as if it were the result of deductive, rather than inductive, reasoning?
- Also I'd like to note that Occam's Razor does not lie utterly outside of the scope of axiomatic justifications, as demonstrated in the following paragraph from the article:
"The razor's strict form, which prohibits irrelevant assumptions in a given theory, is justified by the fact that all assumptions introduce possibilities for error. If an assumption does not improve the accuracy of a theory, its only effect is to make the theory more error-prone, and since error is undesirable in any theory, unnecessary assumptions should be avoided."
- GK:I am saying that the "alien" example was a bad example because there are far more different predictions that can be made with an explanation that includes aliens than from one with only lightening. That the two explanations may account for the burning tree but would not make the same prediction for other phenomena, such as the overall weather, or shiny objects in the sky, strange whirring noises, or other markings on the ground, surgically gutted cattle nearby and so on. The two explanations are only equivalent for the single case of the burning tree. But they do not make the same predictions in general. The explanation that does the better job of predicting other concurrent phenomena that can be observed is the most preferred scientifically. It is not as if science is out to explain only one thing in the universe.
--AceMyth 17:23, 29 April 2006 (UTC)
- The razor is no less scientific because it is inductively justified (i.e., probabilistic rather than certain); on the contrary, according to Russell: "the general principles of science, such as the belief in the reign of law, and the belief that every event must have a cause, are as completely dependent upon the inductive principle as are the beliefs of daily life." (The Problems of Philosophy, §6). If we discount the razor because it relies on induction, then we must discount the scientific method as it equally relies on induction; and so on. Soon we would have no definitions, axioms, facts or laws (for which of them is derived completely from a purely deductive knowledge of the world?) — soon we would have no science. » MonkeeSage « 22:18, 29 April 2006 (UTC)
JA: Now wait, which is it? (1) The razor relies on induction? (2) Induction relies on the razor? Maybe it's some kind of Relier Paradox? Jon Awbrey 23:00, 29 April 2006 (UTC)
- 1. Induction must rely on the Razor to reach any useful conclusion. 2. Scientific inquiry must rely on induction to reach any useful conclusion. Ergo 3. Scientific inquiry must rely on the Razor to reach any useful conclusion. Induction and the Razor do not rely on one another per se; rather, the razor is the only valid mode of induction, which in turn is an inseparable fundamental of scientific method (see Popper's naive falsificationism for the absurdities which would result from denying this). --AceMyth 23:40, 29 April 2006 (UTC)
JA: So now we have the Straight Razor: "The razor is the only valid mode of induction". Is that really what you meant to say? Jon Awbrey 23:50, 29 April 2006 (UTC)
I meant "mode" as a synonym for, say, "frame of mind" or "modus operandi". How can you draw any conclusions by means of induction without using the Razor (implicitly or explicitly)?
That aside, keeping the example out might be a good idea regardless, as it makes the "in science" section flow much better. Perhaps an example of this sort would be better off at underdetermination than here, as it serves to elucidate that concept better than that of Occam's Razor. --AceMyth 00:20, 30 April 2006 (UTC)
JA: Here's an example of applying OR:
- Phenomenon. You are saying something about scientific method that I find surprisingly simplistic.
- Hypothesis 0. You do not have much acquaintance with the actual practice of scientific method.
- Hypothesis 1. You do have a considerable acquaintance with the actual practice of scientific method, and in addition know something else about it that renders it all surprisingly simple for you.
- Evaluation. H0 contains 1 clause and has a message length of 80 bytes. H1 contains 2 clauses and has a message length of 180 bytes. Therefore I reject H1 and accept H0. Jon Awbrey 03:20, 30 April 2006 (UTC)
I would say that is a reductionistic example, as we don't know what supporting evidences there are for H0 or H1 (i.e., we don't know if both are actually sufficient explanations of the phenomenon); but granted that they are, I do not see how H0 is simpler (in the OR sense of appealing to fewer explanatory entities) — both hypotheses involve a number of other propositions which are not indicated, and without knowing the supporting evidences for each hypothesis, a judgement of simplicity would have to be suspended (e.g., mabye H1 requires 50 other postulates about motivation and personal psychology, while H0 requires only 20). BTW, I don't know about Ace, but I'm not arguing that the razor is necessarily/always accurate, just that it is considered to be a valid selection criterion in many cases. » MonkeeSage « 03:47, 30 April 2006 (UTC)
JA: OIC. I knew there'd be a catch, or 22. Jon Awbrey 03:54, 30 April 2006 (UTC)
JA: OR'. Hypotheses that appear simple may have unknown numbers of epicycles on their deferents. Jon Awbrey 04:00, 30 April 2006 (UTC)
- 1. I may have less experience than you in the practical application on the philosophy of science, logic, debate and all that, but I do recall learning early on something about argumentum ad hominem being a logical fallacy, even when veiled by ill-conceived wit. 2. Stripped of that, what's left of your response is the equating of my claim, that the so-called "modern interpretation of Occam's razor" is in fact an inseperable aspect of inductive reasoning, to the claim that any two theories can be validly compared based on which could be formulated using shorter sentences. Those two propositions, needless to say, are hardly equivalent. If you're that much more knowledgeable on this subject than me please be kind and enlighten me by addressing that claim directly and demonstrating exactly what is wrong with it rather than pointing and laughing, which benefits neither any of us nor the article. --AceMyth 06:05, 30 April 2006 (UTC)
JA: I will review the record again to see if I have been careless in any particular, but I do not presently think that I have referred to any matters of person in my criticisms of arguments, assertions, examples, or interpretations. But I do call them like I see them in these regards and I will continue to do so. The stakes demand that nonsense be labeled as such.
JA: I do not keep track of who adds what to the article, but the plain fact is that there is a lot of gross and misleading oversimplification of scientific method in the article, most of which is trained out of most researchers in their first statistical methods course, if not sooner, and the rest of which has been shot down repeatedly in various scientific and philosophical literatures, for as long as I can remember anyway. The fact is that a couple of people are currently defending this nonsense, and so I am presently addressing them, purely for the cure of souls, as it were. Jon Awbrey 15:51, 30 April 2006 (UTC)
Fine, then. I'll be the first to enthusiastically support such endeavours. My current understanding of the problem of induction and how the razor relates to it agrees with many points in this article to a significant degree, which is why I've been prompting you during all of this discussion to share exactly what is wrong with it, and why, rather than dismissing it by hand-waving. I'm not going to learn anything new about the scientific method from vague accusations such as "nonsense" and "gross oversimplification" and neither are the potential readers of this article. --AceMyth 17:44, 30 April 2006 (UTC)
- Okay, I've gone ahead and read some of your comments in other pages. Your main concern seems to be that there exists no rigorous definition of exactly what "simplicity" is, and thus no rigorous version of Occam's razor to begin with. Also that the problem of underdetermination alone does not automatically necessitate the existance of one single principle which solves it, let alone it being this undefined criterion of simplicity. Did I get it all or is there more to it? I want to be able to address the issue fully and without the interference of any phantom arguments that are constantly alluded to yet are never presented.
JA: Yes, that's a big part of it. Razors of all kinds remain vague: (1) too vague to be treated as empirical hypotheses, (2) and thus too vague to be tested up or down by any conceivable experiment, (3) too vague to be given a uniquely preferable mathematical definition, (4) and thus too vague to prove mathematical theorems about. If you look at historical examples, different thinkers' choices of razors are as varied as their theories, indeed, razors are every bit as theory-laden as theories. The progress of science on any question has never, so far as I can tell, been settled by razor, guillotine maybe, but no razor. This is because all of the combatants are using some principle of simplicity — it's only human nature to do so — but they all have different razors in their kits. Some say that it's simpler to assume that things are just as they appear: the ground appears rock solid, therefore the earth does not move except during earthquakes; the sun appears to move through the sky, therefore it does. Jon Awbrey 21:32, 30 April 2006 (UTC)
- Meanwhile, of course, you're welcome to be bold and make any edits to the article you see fit. If they're properly cited and make sense I, personally, will not revert them (and yes, by "making sense" I do not mean "strictly adhering to my own point of view"). Because really, my only mental commitment is to what reason dictates and I was thus far under the impression that the necessity of Occam's razor in inductive arguments is something that reason dictates. I will not throw tantrums if you show otherwise. --AceMyth 19:16, 30 April 2006 (UTC)
Jon: How do you select between the Newtonian theory of imperceivable "tethers" to explain gravitic effects and the Einsteinian theory of curved space? It's not by direct verification or falsification, obviously. Don't you select the Einsteinian theory because that theory fits better with other theories, and requires less explicative entities (viz., simpler in relation to the body of knowledge, simpler in itself)? If so, how is that fundimentally different from the razor? » MonkeeSage « 23:17, 30 April 2006 (UTC)
JA: During the years when I was taking physics in high school and college, secondary and post-secondary school physics educationalists were deeply engaged in a series of curriculum reforms, part of the thrust of which was directed toward cleaning out a lot of the "dust bunnies" of mythological physics history that had collected over the years, and they made us read numerous accounts of primary source physics-in-progress without always telling us what the answer-at-the-back-of-the-book was going to be, so maybe I got sensitized to these issues at an early age. At any rate, I hope that you can, with some reflection, see the absurdity of suggesting that we trajected physics from Newton to SR, to GR, to QM, and beyond to QED by razor alone. Jon Awbrey 18:12, 1 May 2006 (UTC)
Oh, I see. I've long been aware of /that/ problem, which is why I think the only way the concept of the razor - because I still think there is a true razor at the basis of all of this that is valid - could only hold water as a coherentist, subjectivist guide. So the very razor you're using, itself, must either be a direct consequence of logic or become a theory which could actually be false, and thus not worthy of becoming a methodological principle for the same reason nobody bases mathematical proofs on unproven hypotheses.
Under these criteria, the meaning of "simple" must be independent of any whim outside the strict realms of logic: Out of all consistent systems which could account for any set of evidence, you choose the one (or any of the set) with the least postulative elements ("postulative elements" manifesting as, say, taking as true something the perceived probability of which within the system is only around 60%, or in the worse case, taking something "for granted" just because, without any way to collect data that would relate to it). This doesn't set a foot outside being a direct consequence of probability theory. Applied to perceived probability rather than objective probability thus making it quite fallible, but seeing as the perceived probabilities are pretty much what we're stuck with...
Taking the example you used, if you want to suddenly say that "things are as they appear", now that sets a foot outside the scope of probability theory, it assumes something about the universe that we have no guarantee about being true. Though sometimes it co-incides with the form I think is the better one, sometimes... It does NOT. It needs, as you said, to be empirically tested, and together with all the data we've collected by now about how the universe actually does work, this thing just does not form a coherent system. Not anywhere near "the universe may be simple, but its definition of simple is very different to ours", that is.
This is also why the razor was never crucial in the history of breakthroughs in science. This is true in the same way that transitivity was never crucial in the history of breakthroughs in mathematics. --AceMyth 14:21, 1 May 2006 (UTC)
JA: "Things are as they appear to be" is a principle of simplicity. You argue that it's too simplistic to get us very far in science. I agree. But the point is that all known razors are too simplistic to get us very far in science. That is why the move from one step to the next is never achieved by razor. Circles are simpler than ellipses, therefore the motion of the planets must be founded on circles. What could be simpler? Jon Awbrey 18:34, 1 May 2006 (UTC)
- Was that directed at my comment to Jon above, or JA's previous comment? » MonkeeSage « 16:57, 1 May 2006 (UTC)
- JA's. --AceMyth 17:30, 1 May 2006 (UTC)
Jon: I think you must be talking about your own razor, not the nominalist razor. The real razor dictates: "don't multiply entities [i.e., reasons, causes, &c] beyond what is necessary [to explain the phenomena]." Your supposed examples refer to simplicity of presentation/form and simplicity in abstraction from other relevant data and explanatory power. Eliptical planetary orbit is simpler in terms of its relationship with other known data and its explanatory power (whereas circular orbit would require more postulates, e.g., epicycles — facts not in evidence, except in order to support the circular orbit hypothesis — to explain the observed phenomena which eliptical orbit explains as a postulate in itself); so it is a perfect example of the razor. Why select eliptical over circular when you cannot directly observe, verify, or falsify either hypothesis? Because it requires fewer entities to explain the phenomena, and fewer entities in connection with the body of knowledge! » MonkeeSage « 09:35, 3 May 2006 (UTC)
JA: YAFEORC. Yet another fine example of retroactive casuistry . The fact is that there is no prior reason for saying that a conic of two parameters is simpler than two conics of one parameter, so razors in general are useless in deciding the issue before it's decided by reasons quite independent of simplicity. Jon Awbrey 16:30, 4 May 2006 (UTC)
- Jon: Since we are not going to use the example in the article anyhow, I guess we can just agree to disagree on the issue. I still don't see what structural/mathematical/presentational simplicity have to do with the razor (which, as I understand it, deals with ontological/epistemological simplicity as a property of explanations, understood in the context of the phenomenon and the body of knowledge). You may be interested in reading John L. Taylor's essay, "Ockham's Razor in Russell's Philosophy" (PDF), and in knowning that New Scientist called the razor "salutary, indeed indispensable." ("The blunting of Occam's razor", Colin Tudge, no. 1971, 1994). I leave you with the words of Wittgenstein: "Occam's maxim is, of course, not an arbitrary rule. . ." (TLP, 5.47321). » MonkeeSage « 01:17, 5 May 2006 (UTC)
In Medicine: Diagnostic Parsimony
I think that saying that diagnostic parsimony is useless is, in itself, misleading. True, you /could/ in theory find some very rare disease that would explain all the symptoms. But that would be ignoring the obvious non-parsimonious assumption that out of all people, /that/ person happens to have the disease. Suppose you have a patient with two symptoms, one of which can be caused by a disease that roughly 3% of the population has on average and another that can be caused by a disease that roughly 5% of the population has on average. Then you have a third possible disease that would account for both but only 0.1% of the population has on average. Even though the latter explanation postulates only one cause, by simple probability calculations the former is 1.5 times as likely.
That is, of course, not to say that the process of hypothesis formulating and testing in medicine is not absolutely crucial. --AceMyth 11:26, 10 May 2006 (UTC)
[MJP] Mike Palmer is working on a method of calculating the degree of complexity of a theoretical model. For example, particle-only, wave-only and quantum models of e-m radiation are analysed and a ‘validity quotient’ is derived, which is a measure of the extent to which the model is consistent with repeatable experimental observation. Occam can be applied immediately, unambiguously and independently of ‘personal opinion’, to select the best model. Without an approach such as this, we remain stuck where we always have been: with a method that arrives at the ‘best’ view via a semi-democratic opinion poll amongst a self-appointing and self-regulating scientific elite. The outcome is supposed to be the ‘best working theory’, but must always be a woolly consensus, by definition. To date there is not one single historical instance where Occam has been used in a rational manner. —The preceding unsigned comment was added by 193.113.37.9 (talk • contribs) 11:16, May 16, 2006 (UTC).
- It appears to be a problem in the users, not the razor itself. We may not always be able to ascertain which solution requires the lest amount of entities and best accounts for the observations, but in the cases where we can, the razor is a helpful selection criterion. For example, for a US patient presenting with a skin rash, who claims to have been exposed to chemicals consistent with such a rash; the most likely diagnosis, based on the razor, would be dermal irritation due to exposure to an irritant. This would be preferred over a theory about a pig in Borneo carrying a rare, rash-inducing bacteria, having been brought to the US in the 1930's, deficating in a field, the bacteria being dormant until the field was recently re-planted, and then spreading to the man through ingestion of vegetables grown in that field. The later scenario may actually prove to be correct, but the former has a much higher statistical probability (and hence the razor is inductively justified). » MonkeeSage « 17:17, 16 May 2006 (UTC)
[MJP] Summary of above point:
O1 = patient presenting with a skin rash
T1 = exposed to chemical X consistent with rash
T2 = caught rare pig disease Y consisten with rash
The most likely diagnosis, based on the razor, would be T1. T2 may actually prove to be correct, but T1 has a much higher statistical probability (and hence the razor is inductively justified).
Clarification:
Occam relates to multiplication of entities. Entities are:
- theoretical concepts
- theoretical laws
- assumptions
Chemical X is known to be present in many USA locations, and for T1 needs to be assumed present in others where not yet known as fact, i.e. some assumptions needed in support of T1. Disease Y is known to be present in few USA locations, and for T2 needs to be assumed present in many other locations not yet known as fact, i.e. many more assumptions needed in support of T2. Here we have knowledge that enables a measure of probability to be derived for T1 and T2, and so the greater probability leads via the assumptions-entities route to the application of Occam and a determination to be made (even if false).
There is a problem with Occam, in that until knowledge is complete, it can produce incorrect selections.
Key Point:
Occam is a means of choosing between rival theoretical models. It is widely understood in the scientific community as having been part of the rationale for major decisions in choosing a ‘good’ theory over a weaker one. This is rarely the case, and is a common misunderstanding that leads to a widescale falsely-based belief in the 'validity' of the ‘current working model’ - in many fields. The 'validity' is never measured/proven/objectively justified. —Preceding unsigned comment added by 193.113.37.9 (talk)
- If the lack of pre-theoretical assumptions is the only way to acheive validity, then nothing can ever been known as valid, as none have observed every instance of time or space or the relationships of objects to each other or to time and space. Thus an experiement can't actually prove anything (on this account), no matter how many times it's repeated, since it takes a pre-critical assumption about uniformity to take the results from description of specific instances to general principles and limiting concepts (i.e., the application of deductive processes require inductive inferences). The truth of pre-theoterical assumptions in general,or in specifical instances, can be challenged; but their presence doesn't de facto invalidate a theory (as all theories require their share of such assumptions). » MonkeeSage « 00:16, 23 May 2006 (UTC)
Re: newly inserted passage: Can't this be said more clearly and in a way that does not sound like musing of non-physicians about what physicians do? Perhaps a quote with a citation? Or just a summary of something from the literature where physicians are discussing the tension in plain terms?? .... Kenosis 14:38, 24 May 2006 (UTC)
- "It is, however, questionable whether a single disease which is so complex that it can account for multiple symptoms, is simpler than a combination of several simple diseases just because it is nominally referred to as one item and they are referred to as many; if features of items are included as entities of explanations, then the single disease, not to mention the circumstances under which it may be contracted, could prove to be far more complex than an array of simpler and more common diseases, and thus Occam's razor would select for the simpler solution: multiple common diseases."... 14:38, 24 May 2006 (UTC)
I have removed this passage requesting an analysis and reconsideration of it before reinsertion in the article. (1) "Multiple common diseases" is poorly phrased. (2) The assumption that Occam tends toward a diagnosis of multiple diseases with a single common cause is reasonable except for one very important issue. Physicians are responsible for treating the multiple diseases and must identify them in order to treat them. A physician cannot reasonably say, for instance, "Well she has a whole bunch of multiple common diseases all attibutable to a sodium deficiency and long-term dehydration." Treating the sodium deficiency does not heal , e.g., the infections, dilapidation of tissues in the body, the ringworm, the additional virus she caught while her immune system was weak, and the bad spinal condition that developed and caused the bone structure to change over time, etc., to give just one scenario ....Kenosis 17:12, 24 May 2006 (UTC) What the paragraph above appears to be asserting, it should be said, has also been known to physicians as "Osler's rule" (or, "one disease to a customer"), this extreme late-19th Century version of Occam's razor attributed to William Osler. Perhaps needless to say it's a very controversial position today... Kenosis 18:51, 24 May 2006 (UTC)
- Hey guys, I'm the poor sucker who put the Iccam dictum thing in in the first place. Thanks for the "hickam" thing... man I've been looking to find a reference to "iccam" for so long and it never occured to me that I might have misspelt it!!! :-)))
- Talking about statiscitics in clinical medicine is useful up to a point. There is insufficient data about the true rates of diseases and more improtantly symptoms (becasasue after all a doctor is presented with symptoms and not a disease) for doctors to aply statistics to figure out what is wrong with a patient. I do not think that it is misleading to say that Occam's razor is "uselss" in clincial practice. On the very rare occasion that Occam's razor is invoked it is purely by chance. Allow me to explain:
- A doctor seeing a patient will find out about what symptoms they have. The doctor will then think of a number of causes of those symptoms, starting with the most obvious and serious. E.g. coughing up blood, chest pain, weight loss, he'd think oh oh... this person may have lung cancer, or bronciectasis, or a pneumonia, or a few other things. The tests are done to exclude/include those diseases, so the patient will have a chest X-ray, a bronchoscopy, their sputum cultures, some blood tests etc. Some of you will notice that the three symptoms can actually be caused by three independant things e.g. coughed up blood totally at random (it can hapen) the chest pain is because of heart disease, and the weight loss is because they are going through a stresful time in their lives and they arn't eating properly. The point is that when taking the patients medical history, the doctor will ask questions that follow the pattern on the illness e.g. coughing up blood... in someone with lung cancer the blood will be bright red, thee is liable to be a lot of it, it's likely to have gone on for some time, they're inevitably also smoking etc. I.e a symptom of "coughing up blood" is a whole lot more complicated than that... how long have they coughed up blood for, what does the blood look like, how much came up, how often do they cough up blood. All these pieces of information inform your diagnosis, not just "coughs up blood." In other words all that statistical stuff that you guys are refereing to is wasted, 'cos there are no hard figures on this kind of stuff for all possible populations.
- Someone mentioned that Occam's razor may explain some cases where probabilities just add up. Thats nice. But how is that useful clinically? You still need to exclude the common causes, you still have to run the same tests. It just doesn't help you in the practical world at all. Thus Occam's Razor is simply useless in diagnosing a patient. So Ocam's razor came up with the right diagnosis in one case out of hundreds... would you consider that to be in any way useful? This is not an "important limit", this is an absolute contraindication to the use of Occams razor. Sure you'll consider a rare diagnosis if the symptoms fit, but then again you'll also consider a common diagnosis if the sympotms fit, Occam's razor just doesn't help you decide which is more likely.
- gerg 13:56 26th May 2006 aka Gergprotect
The edits I just added are very very ugly indeed... I guess I wrote them to get across the point of what I mean, rather than as a finished polished article... that the: "However, this principle has no pratical influence in diagnostic medicine." and "Statisctical analysis of a patients symptoms does not help because of the paucity of data, which is caused by the difficulty in aquiring such huge amounts of information on so many possible permutations, and also because even if a person is statistically more likely to have any given disease, you still need to run the same tests that you would have had to if you had not analysed their symptoms statisticaly, making it an excercise in pointlessness." bits... Gergprotect 13:27, 26 May 2006 (UTC)
[MJP] Key Point (Round 2 – Scientific Validity):
Occam insists that rival theories be compared. If the Razor is not applied objectively, then there is no scientific way of selecting the better one. In the wider scientific community there is a misunderstanding of what ‘validity’ means - that the consensus view defines ‘valid’ (this is corrected by http://en.wikipedia.org/wiki/Scientific_consensus). Pre-theoretical assumptions are mandatory for an objective comparison. As such a methodology is currently incomplete, scientists should not assert that any theory is ‘valid’, but rather refer their work to a given ‘suspect working model’. This is not to say that such a theory IS invalid, but merely that it cannot be assumed to be valid (the common mistake). Until a pre-theoretical mechanism is available, the ‘better’ one is selected by ballot – what effectively happens now. And if /when the mechanism appears, it is relativistic, then it could only be assumed to deduce the least bad theory, not a theory that will remain without known flaws. A relativistic mechanism might produced in due course, but it is hard to see how an absolute mechanism could emerge, and obviate ultimate ignorance. The objective scientist and philosopher recognises this position – uncomfortable because it is one of humility.
Occam's Razor in medicine
That section of the article has this to say:
When discussing Occam's razor in contemporary medicine, doctors and philosophers of medicine speak of diagnostic parsimony. Diagnostic parsimony advocates that when diagnosing a given injury, ailment, illness, or disease a doctor should strive to look for the fewest possible causes that will account for all the symptoms.
This principle has very important limitations in medical practice. Among the reasons is that it is more likely for a patient to have several common diseases, rather than having a single rarer disease which explains all their myriad of symptoms. A variant of Occam's Razor typically taught to all medical students learning how to make diagnoses, is the expression, "When you hear hoofbeats, think horses not zebras." This is to impress upon the future doctors that the more common ailment is the most likely. The actual process that occurs when diagnosing a patient is a continuous flow of hypothesis and testing of that hypothesis, then modifying the hypothesis and so on. At no stage can a diagnosis properly be made or excluded because it doesn't immediately appear to fit the principles of Occam's razor. The principle of Occam's razor does not demand that the diagnostician necessarily opt for the simplest explanation, but instead guides the medical practitioner to seek explanations, without unnecessary additional assumptions, which are capable of accounting for all relevant evidence.
"Hickam's dictum" is a modern counterargument to the use of Occam's razor in the medical profession. Put succinctly it states: "Patients can have as many diseases as they like!".
Diagnostic parsimony: "A doctor should strive to look for the fewest possible causes that will account for all the symptoms. . . . When you hear hoofbeats, think horses not zebras."
I wonder if this is at all controversial. There is, it seems to me, a critical difference between developing a theoretical construct that will account for some set of phenomena and diagnosing a patient with some set of symptoms. Tha difference lies in the desired outcomes. The goal of the diagnosis is proper treatment; thus, when multiple diseases account for all the symptoms that present on initial visit, failing to consider the less common ones is not IMO an optimal strategy.
I mention this because I know (from reading) of just such lazy diagnoses. For example, Robert H. Goddard suffered from tuberculosis all his life, but died of cancer. Yet when the symptoms of cancer began to manifest, the Baltimore doctor he saw commented that he had never heard of cancer superimposed on tuberculosis. (See Lehmann, This High Man.) Another example: a patient with Huntington's chorea had to see 31 doctors before one identified the disease. And this was a known disease at the time; in fact, Woody Guthrie's well-publicized death from HC had occurred some years before. (See Walter Bodmer, The Book of Man.)
So when I read that medical schools are teaching future doctors, "When you hear hoofbeats, think horses not zebras," I have to hope that the article has phrased that poorly -- for zebras' hooves, too, make that clopping sound. --Chris 20:50, 8 July 2006 (UTC)
And in this subcategory, should we reference the "House, (the TV show) episode titled Occam's Razor? --LeafPlacement 19:34, 24 July 2006 (UTC)