Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!cs.utexas.edu!usc!elroy.jpl.nasa.gov!ames!saturn!xanthian From: xanthian@saturn.ADS.COM (Metafont Consultant Account) Newsgroups: news.groups Subject: Re: sci.philosophy.objectivism Message-ID: <10618@saturn.ADS.COM> Date: 25 Jan 90 08:58:50 GMT References: <1990Jan13.140242.14111@twwells.com> <9001140024.AA18363@apee.ogi.edu> <8P1157Gxds8@ficc.uu.net> <8250@portia.Stanford.EDU> <1990Jan17.083725.1215@twwells.com> <9001182300.AA28308@apee.ogi.edu> <10507@saturn.ADS.COM> <900@athen.sinix.UUCP> Organization: Advanced Decision Systems, Mt. View, CA (415) 960-7300 Lines: 181 [The only _possible_ excuse for continuing this discussion in news.groups is that it still, marginally, pertains to the question of whether *philosopy.objectivism should have "*" replaced by "sci".] In article <900@athen.sinix.UUCP> es@athen.UUCP (Dr. Sanio) writes: =In article <10507@saturn.ADS.COM> xanthian@saturn.ADS.COM (Metafont Consultant Account) writes: =>In article <9001182300.AA28308@apee.ogi.edu> mehuld@APEE.OGI.EDU (Mehul Dave) writes: =>[...] =>>I think philosophy is most certainly a science and therefore it should => =>The premier modern discriminator between science and non-science is =>that a science must produce hypotheses/theorems which are =>"falsifiable"; in the context of the current discussion, that means =>that if I produce _one_ instance in which your system/theory =>demonstrably fails, you will stop arguing with me that it is right, go =>away, and fix it. => =>This is not a characteristic of any system of philosophy of which I am =>aware. Instead, into the teeth of massive demonstrations of error, =>the true believers continue to espouse their "inerrant" faith. Regard =>any discussion on the net for corroberation, the current one in =>news.groups, for instance. => =Let me point out first that I share the opinion that the group xxx.philos.o'ism =doesnt fit into the sci hierarchy. At least we have the beginnings of a basis for further agreements! ;-) =To the rest, I disagree. The criterion of "falsifiability" reflects =the view of just one branch of the epistemological discussion, that =of positivism (Popper et alii). Other epistemological schools share =it only up to a certain amount, or even reject it as less useful or =even meaningless. Sorry, but no. I guess I should have said "The premier modern discriminator _among scientists_...", but that leads immediately to a circularity problem, since it is exactly the question of who constitute "scientists", that we are trying to answer here. I am profoundly uninterested in what the philosophers of knowledge have to say about science. I am profoundly interested in what the scientists whose science leads to successful and practical results have to say. Falsifiability is, for example, the basis on which "creation science" is rejected _as a science_ by the vast majority of science practitioners. By design, "creation science" avoids positing testable arguments, and instead its proponents spend their intellectual resources explaining away contradictory evidence. =In fact, the systematics of "science" and "arts" in =the academic field existed long before positivism. IMHO, =falsification is an important access to academic reasoning, as well =in fields which are clearly not "science" (sociology, history etc), =but it isn't all even in "science". Try to apply it to medicine, =psychology, education, economy and you are doomed to failure. Sorry, no. The statement of the falsifiable hypothesis must be made a little more carefully in fields where the evidence is statistical rather than capable of a direct test, but each of these sciences uses falsifiability directly in research every day. For example, in medicine: For a sample of statistically significant size, we can state with a 95% confidence level that the application of anti-melanoma protocol A will lead to a five year survival rate 30% higher than protocol B, a placebo protocol, all medically significant other variables being controlled for by random sampling of the test cohorts. is testable, falsifiable hypothesis. Confirmation testing by independent researchers of unexpected results are normal medical research practice. If more that 1 of 20 such tests fails, the hypothesis is rejected. Behavioralists have used falsifiable hypothesis since Pavalov was a little boy with a pet cocker spaniel. For example: In a Skinner box, with 95% confidence, a pigeon trained to peck a triangle in preference to a circle to elicit food will fail to extinguish that behavior (despite its unprofitable calorie budget) as long as at least one food pellet is elicited for every 300 pecks, to starvation. is a classic, testable, falsifiable hypothesis. (Whether it is _true_ or not, I have no idea!) Similarly, in education: For a sample of statistically significant size, we can state with a 95% confidence level that the application of Head Start programs will lead to a 1.2 year level improvement in reading skills tested at completion of grade three, against a cohort of children not given Head Start training, control being exerted for other causitive factors by random selection. Economics has some problems controlling for "other things being equal", but, for example, "decreasing the tax burden on a governmental units' population will lead to increased well being and prosperity" has been massively disproved by the California experience, and _honest_ economists are beginning to realize that a taxation rate that allows for maintenance of the economic infrastructure is indeed essential to well being and prosperity. This does not justify waste of tax resources, but it does seem to defend proper use of them. =>-- =Concerning your "true believers" argument: did you follow the Beckmann =discussion in sci.physics or cold fusion arguments? No, but I followed it in the popular and scientific press. =I'm not expert enough to decide who was right or wrong in all =details, but in result I could say : "Hey, at least one of the =factions (or both) ignore massive demonstrations of error". Sorry, no. It is certainly likely that one or several experiments were poorly thought out or poorly designed. The potential rewards, in terms of benefit to humankind, personal prestige, and personal economic gains, were so great as to promote "fast science", a fairly chancy proposition at best. Nevertheless, the cold fusion example (and by the way, _confirming_ experiments are still being reported, among other locales, from within the Soviet Union, so there is still hope remaining) _exemplifies_ falsifiability; the result reported by the Utah experimenters was so unexpected that numerous laboratories, worldwide, immediately attempted to confirm or falsify it. That the experiment was 1) extremely dependent on poorly documented features of the experimental apparatus, 2) very difficult to perform, 3) quite dangerous, and 4) involved several poorly integrated branches of science (fast and slow neutron detection, atomic mass spectroscopy for tritium byproducts, and thermal budget balance, among others), led to a (continuing to this day) series of inconsistent results, which is why the attempt to confirm or falsify the result _decisively_ and _repeatably_ is still ongoing; the science, and the example of falsifiability as a method of science, is exactly what is wanted for science to work, however painful (in terms of expense and blemished reputations) the process might be. =Read the history of science, and you will learn that ignorance and =dogmatism don't end at the borders of "science". Not to mention greed, charlatinism (sp?), and a host of other human ills. It is to combat human frailities that the peer review system of science exists. That it is fueled more by self interest than by altruism is usually true; that it works anyway is a fine compliment to the strength of the basic paradigm. =And, by the way, the criterion you cite as the decisive discriminator =(falsifiability) was developed in a "non-scientific" branch, =sociology. The _single_ sociology class I've ever taken left me the impression of a science just waiting for a suitable set of test subjects. I think that sociology is doomed to be a science on the model of astronomy; no experiments allowed, but any hypothesis presented can be falsified by appealing to the observational record for contradictory evidence, and tentatively confirmed by confirming evidence, the lack of contradictory evidence, and an appeal to Occam's razor to assure that the hypothesis is useful/necessary to explain the observations. Since astronomy has progressed by leaps and bounds using such relatively unsuitable tools, perhaps sociology can do a "go thou and do likewise" act. =Any comments? I must be a stranger to you; this question referenced to the man from xanth is eminently superfluous, as hundreds of USENetters will gladly, (and occasionally loudly and angrily) testify! ;-) =>Again, my opinions, not the account furnishers'. =So are mine => =>xanthian@well.sf.ca.us xanthian@ads.com (Kent Paul Dolan) = =regards, es es@athen.uucp sanio@netmbx.uucp (Erhard Sanio) -- Again, my opinions, not the account furnishers'. xanthian@well.sf.ca.us xanthian@ads.com (Kent Paul Dolan) Kent, the (bionic) man from xanth, now available as a build-a-xanthian kit at better toy stores near you. Warning - some parts proven fragile. -> METAFONT, TeX, graphics programming done on spec -- (415) 964-4486 <-