cbo@utai.UUCP (07/03/87)
Eric Raymond (eric@snark) writes: | in order to believe that science is anything but futile you must at some | level be an empiricist and/or operationalist. Yes, there are scientists | who profess to have other philosophies in connection with their religious | beliefs or whatever -- but *in the lab* they must function as operationalists | or give up. Who knows how the terms "empiricist" and "operationalist" are being used above. But under normal readings of these terms, it seems to me that empiricism is just as religious as realism, yet far less satisfying and far less plausible. It insists upon a principled metaphysical and semantic distinction between the nontheoretical (which we can directly observe) and the theoretical (the unobservable, which we can only infer), in which the former of these is taken as somehow primary, and the latter as defined. What is this but an unsavory religion that places humans and their evolutionarily produced idiosyncracies at the center of what there is? Raymond quotes Carnes (I think): | > Your view of knowledge is, in | > broad terms, an empiricist one, which is orthodoxy in the hard | > sciences because it explains them well. But the philosophical | > question you and others are begging is whether an empiricist | > epistemology is adequate for the kind of understanding we can have of | > the specifically human world studied by the social sciences. Hey, dont give the empiricist too much! What ever happened to the unity of the sciences? I dont see that empiricism for the "hard sciences" is any better off than it is for the "social sciences". | Knowledge is instrumental. Humans gather knowledge in order to *predict* | things and *do* things. Humans form beliefs (including philosophical | beliefs) in order to organize knowledge so they can *do* things and | *predict* things. What knowledge *is* and what people gather knowledge in order to *do* are two different things. People gather knowledge for many different purposes. The question dividing various kinds of anti-realism and realism is: can a scientifically sensible story be told in which the instrumental utility of knowledge is accounted for without appealing to notions of correspondence truth? Common sense folk psychology suggests the answer is no: generally, the reason that knowledge is "useful" in predicting and doing "things" is that it is true. Attempts to escape this result by appealing to its "usefulness" in satisfying desires run aground on the embarrassing problem that satisfaction conditions for desires are the same kind of thing as truth conditions for beliefs. A desire, after all, is for a certain proposition to be rendered true. | Empiricism, verificationism, and operationalism *work*. | Repeatably. Consistently. Powerfully. After 300 years of this, the arrogance | displayed by the likes of Charles Taylor and implicitly defended by Mr. | Carnes would be laughable if such people didn't retain a disproportionate | amount of influence on the culture. As science, empiricism/verificationism/operationalism seems to fail for the reasons outlined above. As methodology, it fails because no one could seriously do any work if she thought the theoretical terms in her theory were defined operationally or in terms of the ways of verifying sentences containing them. The reason is simple: as soon as a theoretical term was introduced ("electron here!" means "the dial moved in such and such a way"), there would be no motivation for discovering other ways to detect instances of the items referred to by that term. But scientists are forced to do this all the time: it is a fundemental aspect of their nonempiricist methodology. I dont defend Charles Taylor; I have to admit that he seems hopelessly mired in the hermeneutic circle. But what makes you think he is any more arrogant than, for example, yourself? | > What the ontology of | > mainstream social science lacks is the notion of meaning as not | > simply for an individual subject; of a subject who can be a `we' as | > well as an `I'. | This is nonsense. There is nothing in general empiricism or the empiricist | program for the social sciences that excludes discussion of intersubjectivity. Far from being nonsense, this seems to be quite plausible to me. The simplest and most consistent version of empiricism leads logically to solipsism. When theoretical entities are fictions, or defined in terms of sense data, as different versions of empiricism would have it, there is little ground on which to build an intersubjective meeting place. | Linguists, for example, deal with intersubjectivity all the time. What is the | 'meaning' of a word but an intersubjective reality? Sure, linguists in the Chomskian tradition deal with intersubjectivity, but where did you get the idea that they are empiricists? Chomsky considers himself to be a Cartesian rationalist. On the other hand, philosophers of language who are empiricist end up being behaviorists and repudiating the notion of meaning as incoherent because they have nothing real to ground it in. This is the main thrust, for example, of Quine's philosophy of language based upon "stimulus meaning". | > The exclusion of this possibility, of the communal, | > comes once again from the baleful influence of the epistemological | > tradition for which all knowledge has to be reconstructed from the | > impressions imprinted on the individual subject. | | Mr. Taylor is not only historically wrong but actually incoherent here. | The ontological status of individual impressions versus intersubjective | reality has nothing to do with the empiricist versus anti-empiricist | distinction. Careful there! I am not a scholar of the history of philosophy, but to me this seems roughly correct (certainly it is not incoherent, thats probably just arrogance on your part). The original British empiricists were, I think, either forced to a skepticism that their heart was not into (Hume), or to an idealism where intersubjectivity was only maintained by the invocation of a supreme Diety (Berkely). Neither of these views seem to "work" very well to me. Calvin Bruce Ostrum (cbo@ai.toronto.edu) Department of Computer Science University of Toronto --------------------------------------------------------------------- "It may be true that the unexamined life is not worth living, but the overexamined life is nothing to write home about, either". Daniel Dennett. ---------------------------------------------------------------------
eric@snark.UUCP (Eric S. Raymond) (07/10/87)
Mr. Ostrum responds with a generally lucid, closely-reasoned argument which is an excellent rejoinder to what he thought I was advocating. The fact that he had my position not quite right is my fault; I used the term 'empiricist' too loosely (as I explained in my reply to Gene Ward Smith). In that posting, I gave informal definitions to each of the terms I am using. To recap: rationalist = "I believe what my reason tells me" empiricist = "I believe what my senses tell me" pragmatist = "I do what works" verificationist = "'Truth' is what independent observers can agree on." operationalist = "'Truth' is what one can predict experimental results with." In my reply to Mr. Carnes I waved the banner of 'empiricism' because of his characterization (historically correct) of the scientific stance as 'broadly empiricist'. I think I have since made it clear that I am actually arguing for a sophisticated form of operationalism as the 'natural' stance of science. In article <3984@utai.UUCP>, cbo@utai.UUCP writes: > empiricism is just as religious as realism, yet far less satisfying and > far less plausible. It insists upon a principled metaphysical and semantic > distinction between the nontheoretical (which we can directly observe) > and the theoretical (the unobservable, which we can only infer), in which > the former of these is taken as somehow primary, and the latter as defined. I can agree with this as a criticism of Hume et al. A properly formulated operationalism gets around having to make 'theoretical' vs. 'nontheoretical distinctions at all -- 'theories' and 'percepts' are both simply elements of predictive models held in the minds of observers. > The question dividing various kinds of anti-realism and realism is: can > a scientifically sensible story be told in which the instrumental utility of > knowledge is accounted for without appealing to notions of correspondence > truth? Common sense folk psychology suggests the answer is no: generally, > the reason that knowledge is "useful" in predicting and doing "things" is > that it is true. One of the things that makes these discussions difficult is that human brains aren't wired for philosophical correctness. The fact that 'folk psychology' argues for correspondence truth is very poor evidence; 'folk psychology' argues for a great many utter fallacies. That last sentence is an *assumption*! > Attempts to escape this result by appealing to its > "usefulness" in satisfying desires run aground on the embarrassing problem > that satisfaction conditions for desires are the same kind of thing as > truth conditions for beliefs. A desire, after all, is for a certain > proposition to be rendered true. Ah, but 'satisfying desires' isn't the criterion an operationalist uses -- 'predictive strength' is. That is capable of being evaluated without this circularity, which I think is mostly a linguistic artifact anyhow. > As methodology, it fails because no one could > seriously do any work if she thought the theoretical terms in her theory > were defined operationally or in terms of the ways of verifying sentences > containing them. No? Quantum physicists *have* to work that way because their objects of study behave so counterintuitively. In fact, what's generally misled you (I think) is a confusion between 1) the intuitive means by which theories are formulated 2) the methods by which they must be analyzed and confirmed I used to be a metamathematician. One of the jokes in that small, odd subculture is to the effect that all mathematicians are Platonists before publication and Formalists afterwards; that is, we think and talk about mathematics informally as though we were studying some kind of ideal world of existent things, but we ground our *proofs* of mathematics in a strict formalist insistance on the rules of 'a game played with marks on paper' -- and this is precisely as it should be. There is a similar though less widely recognized duality in scientific thought. A scientist in a field where the objects of study behave 'intuitively' may (most of the time) think as a realist and find his process of theory-building assisted thereby; indeed, as Mr. Ostrum points out, it is decidedly difficult for humans (wired as they are) to function otherwise. However, when the chips are down, the theory has to be evaluated operationally -- *by testing its predictive power*. The analogy is exact, and instructive. To re-emphasize: the 'philosophical stance of science', is implied by its criteria for *validating* theories, *not* by the 'folk psychology' of how scientists dream up their theories. Mr. Ostrum then devotes several paragraphs to a dissection of naive empiricism, with which I largely agree. I am skeptical of his claim that only Chomskians deal with intersubjectivity (it seems to me any linguist would have to or give up the field) however. Would any linguist care to comment? -- Eric S. Raymond UUCP: {{seismo,ihnp4,rutgers}!cbmvax,sdcrdcf!burdvax,vu-vlsi}!snark!eric Post: 22 South Warren Avenue, Malvern, PA 19355 Phone: (215)-296-5718