Re: Polanyi on science (was Re: [asa] C.S. Lewis on ID)

From: Don Winterstein <>
Date: Mon Dec 01 2008 - 00:55:09 EST

OK, I certainly didn't understand your sentence the way you explained it. Unless a scientist's result defies prior experience, general practice in the scientific community is to accept his result conditionally, which usually means we accept it unless someone demonstrates that it's flawed. The more it deviates from expectation, the greater the need we feel to put it to the test.

As for "irretrievable confusion," chances are good I'll get over it after enough elapsed time. : )


  ----- Original Message -----
  From: Murray Hogg<>
  To: ASA<>
  Sent: Sunday, November 30, 2008 4:11 AM
  Subject: Re: Polanyi on science (was Re: [asa] C.S. Lewis on ID)

  Hi Don,

  This was VERY poorly worded on my part and I fear I have created an irretrievable confusion - but let me try to clarify my point nonetheless.

  By using the phrase "the concept of verification in the sciences" I have inadvertently steered the discussion away from the real point at issue.

  I have made it appear that I am questioning the idea that one can repeat an experiment and determine whether the results of the original researcher(s) were, or were not, valid. This was NOT what I had in mind, and I acknowledge that it would demonstrate an utter ignorance of science to suggest it. So I hope you'll accept my apology for a major miscommunication and not allow it to colour my attempt to clarify what I DID mean.

  First, the topic under discussion was primarily epistemology of testimony and NOT methodology of science (although the first impacts the way we understand the second).

  What I was really rejecting was NOT the possibility of verification in science, but necessity of verification of testimony. But to try to clarify, let me explain one of the two major categories in epistemology of testimony - the idea of a reductionist epistemology of testimony.

  Reductionist epistemologies of testimony hold that one can only believe another's testimony IF one can verify the truth of their claim(s) for oneself. So - to give a specific example - a strongly reductionist epistemology of testimony would hold that I am NOT justified in believing your claim to have been involved in investigations in seismology simply because you say so. Rather I have to find OTHER grounds for believing your claim. Perhaps I need to find the lab you worked at and check that you really did work there. This is a REDUCTIONIST epistemology of testimony because it is the claim that testimony is only to be accepted when it is "reduced" to a more basic category (usually direct empirical confirmation).

  This sort of approach was made popular by David Hume (possibly by a misreading of Hume, actually, but that's another story!) and it has some claim to being "the" Enlightenment view. It undergirds many of the arguments against miracles, tradition, scripture, etc which have pervaded skeptical attacks on religion - Hume's famous argument against miracles, for instance, is the claim that it is more likely that testimony regarding a miracle is flawed than that a law of nature be broken (the definition of miracle in Hume's philosophy). It isn't the claim that miracles are impossible as often misunderstood, but that testimony regarding miracles cannot be accepted because testimony is an inherently unreliable epistemological category.

  But it is not only in philosophy of religion that such a view won popularity, it ALSO became popular in the sciences - particularly so when seized upon by empiricists to bolster their own claims about scientific method.

  Basically, the claim arose that science doesn't deal with such shoddy categories as testimony. NO, Sir! NOOOOOOOOOO, I say! We men of science base OUR knowledge on FACTS! Evidence, Sir! Evidence! Give me the evidence and to blazes with all your fishwive's talk of gods and miracles!

  Er, oops, sorry, brief turn of melodrama there - but you get the point. A distinction was made between science grounded in evidence and religion grounded in testimony (tradition being a form of testimony). And because it was held that evidence is a reliable source of knowledge and testimony is not, therefore, science provides a reliable body of knowledge and religion does not.

  And now to the rub of it;

  What epistemologists of testimony have pointed out is that every scientist relies, in fact, on a pretty large body of knowledge which s/he HASN'T actually gone out and checked for her/him self. Yet at the same time we find that science does give us extraordinarily reliable results. Indeed, it's ironic in a way that your response to my badly worded remarks merely demonstrates the point: the fact that we can so often verify by experiment other scientist's claims shows that scientist's testimony is more often than not reliable. All of which raises a conundrum for those who want to hold to a reductionist view of testimony because one of the fundamental claims being made is that unverified testimony CANNOT be a ground of reliable knowledge.

  The point isn't that the scientist can't verify SOME of the things s/he has been told and, as you point out, it's common practice to take somebody's say so and verify whether or not it's so.

  But understand that THIS is not what reductionist epistemologies of testimony are about. Such epistemologies of testimony are making the radically stronger claim that one is NEVER justified in believing ANY testimony WITHOUT verification.

  In such a theory you would not be justified in believing ANY claim made by your research partners - no matter how qualified, no matter how experienced - unless you yourself verified their claims. You'd have to, by the way, verify their claims as to qualifications and experience also! If you set a post-grad student to work taking readings on a piece of equipment, you would not be justified in accepting their record of results unless you, yourself, stood over them and verified each and every reading. Indeed, you couldn't trust the equipment unless you verified it's theory of operation, its design, process of manufacture, and calibration. And the list just goes on, and on, and on.

  What this boils down to is one has two options: either (A) accept that reductionist epistemologies of testimony are at least overstated or (more likely) just plain wrong; or (B) make the claim that science (being shot-through with testimony) does NOT provide reliable knowledge.

  (B) is absurd, thus (A) is to be accepted.

  I hope that all makes sense and I haven't made a further hash of it? But if I've not lost you, I can now go back to my original statement (" turns out that the concept of verification in the sciences is a grand fiction") and clarify what I was attempting to get at. I'm not opposing the claim that experimental verification is an intrinsic part of scientific practice and that it actually works. Rather I'm opposing the claim that testimony is NEVER reliable UNLESS experimentally verified. I should have written something more like;

  " turns out that the claim that scientists NEVER accept testimony UNLESS verified is a grand fiction."

  I hope that clarifies the issue somewhat? Although, as I say, I worry that I might have caused an irretrievable confusion.


  Don Winterstein wrote:
> it turns out that the concept of verification in the sciences is a grand
> fiction.
> I don't know how it is for everybody else, but in my experience this
> doesn't ring true. I was involved with a new kind of investigation in
> seismology. Once we published our data, every lab of consequence in our
> field conducted experiments to check out our results. Another case that
> everyone knows about is the Pons & Fleischmann cold fusion result: Once
> they publicized their findings, many groups checked on them. Scientists
> are extremely competitive, and if someone comes up with a new result,
> lots of groups will check it out. In fact, my PhD thesis work involved
> checking out an Aussie's claim that he'd detected free quarks in cosmic
> ray air showers. Got everyone excited there for a bit.
> Maybe what you say is true in a limited sense, but I can only
> guess what that might be. Scientists are forever checking up on one
> another, often secretly hoping to find that someone has screwed up.
> (Pointing out someone's error is an easier way of making a name for
> oneself than making an original contribution!) Another motive is to get
> competent fast in a new kind of investigation in hopes of finding
> something worth patenting.
> Don

  To unsubscribe, send a message to<> with
  "unsubscribe asa" (no quotes) as the body of the message.

To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Mon Dec 1 00:55:46 2008

This archive was generated by hypermail 2.1.8 : Mon Dec 01 2008 - 00:55:46 EST