March 24, 2008

Truth, Belief, and Science

Posted by Eric at 10:30 pm | Category: Links, Philosophy, Science

Dr. Free-Ride over at Adventures in Ethics and Science has written a post on a topic near and dear to my own heart: the boundaries between science and belief. Her position is that there is no conflict in the mind of a scientist who believes in a supernatural deity. It’s a topic that I’ve been mulling over for the past 10 years or so.

In some ways, it is very easy to fall into the strict regime in which “if one believes that the scientific method leads to truth, that is the methodology to be applied in all beliefs of truth, public and private.” I think many scientific atheists hold to this sort of belief very often.

There is, of course, that very crucial “if” in that sentence that many philosophers will jump on. Does the scientific method lead to truth, exactly? Maybe not. No truly introspective scientist says that they know the entire truth; all we have are models that work well enough. Where does the photon “go” when it’s absorbed by an electron? Well, we don’t know; we don’t even really know what “absorbed” means, but as long as we do good bookkeeping on the energies and momenta of all the particles, it seems to work out pretty accurately, at least as far as we can tell.

In addition, can I know that my senses are the same as others’? What makes a schizophrenic’s reasonings about the truth of the world less “true” than mine? Does the fact that I have to wear glasses make a difference on my conclusions about how something looks? So truth is, perhaps, more a social construction, especially truth based on scientific observation. “Truth” needs to be verifiable to someone else, and based on some standard that others can replicate, not based on my unique, idiosyncratic senses.

But I feel something fundamentally wrong with simply saying that the scientific doesn’t lead to some sort of generation of knowledge. After all, the keyboard under my fingers feels real enough. I can hit it with other things, to ascertain its existence. Is the possibility that I’m some sort of “brain in a vat” really going to bar me from asserting the truth that my computer exists? Though Dr. Freeride might characterize my belief in the reality of the world I sense as a scientifically unjustifiable metaphysical commitment, can it really not be? Am I not justified in asserting that the world is real until I have reason to believe it’s not? (Being unplugged from the Matrix might do it.)

Is there, then, truth other than this sort of knowledge?

Follow the comments thread by the RSS 2.0 feed. You can leave a response, or trackback from your own site.

8 Responses to “Truth, Belief, and Science”

  1. Dan Says:
    March 25th, 2008 at 5:00 am

    I think Lawrence Krauss said it exceptionally well in an interview for SEED recently:

    That, and the oft-misunderstood fact that science doesn’t prove things to be true. Science only proves things to be false. That’s all it does. But that alone is something that doesn’t happen in almost any other area of human activity. The fact that you can say, “That’s garbage, don’t talk about it any more.” The earth isn’t flat. We don’t need to have critical thinking classes to debate or discuss it.

    And I would extend that thought to reference Bertrand Russell by saying that for things that cannot be falsified, such as the existence of a deity, we can confidently say that some things appear much more likely or unlikely than others.

  2. Eric Says:
    March 25th, 2008 at 6:24 am

    The whole “falsifiability” thing is a bit oversold in science, I think. Technically, yes, science can prove things false, but really, when an experiment gives a result different from expectations, what tends to be falsified first? “I probably messed up a step,” is the first explanation that comes to any grad student’s mind. Any hypothesis comes with a million auxiliary hypotheses, such as, “my reagents haven’t suddenly gone bad since the last time I tested them” and “my control reactions will be equivalent to my actual reactions except for one difference.” When we test things in science, we test (and falsify) in bundles of hypotheses. That makes falsifiability a less than useful philosophy for science, I think.

    I have to agree with your extension. We generally tend think that experiments produce some sort of knowledge. Sure, scientists might claim it’s only statistical knowledge (”likely” and “unlikely”), but that’s still a claim to a form of truth, I think. The problem is that framing the claim statistically doesn’t sidestep the fundamental metaphysical problems that Janet brought up. “Likely” is still a social construction, rather than a personal reality. Perhaps for someone, they think it’s likely that visions of angels appear when they walk near that tree, but it’s still not verifiable by others.

  3. Dan Says:
    March 25th, 2008 at 8:25 am

    Technically, yes, science can prove things false, but really, when an experiment gives a result different from expectations, what tends to be falsified first?

    “Tends to be,” does not win out in science. Take spontaneous generation, for instance.

    And I don’t get it when people bring up “social constructions” for things. You don’t need a social construction to say that it’s unlikely that Santa Claus, monsters under your bed, fairies, or thunder gods in the sky really exists. So “unlikely” is not a social construction. Believing in the likelihood of unlikely things, however, yes, I’ll agree that that is a social construction.

  4. Eric Says:
    March 27th, 2008 at 12:20 am

    What I’m arguing is more of a Kuhnian framework for science, in that most science is not about falsification. That only happens at paradigm shifts; most science is mundane, about fleshing out the frameworks we have now. If my experiment doesn’t work, I don’t then presume to have falsified the theory of atoms and quantum mechanics.

    I think the way scientists work in practice is that theories that have been well-tested are given a privileged status in later testing; not all hypotheses are created equal. So it’s not just that theories are falsified; some theories are made more “likely”, as you said.

    Spontaneous generation failed because it didn’t live up to the standards that the scientific community set for “theory” (and yes, I do mean community). Maybe in one hallucinating, deranged scientist’s perception of reality, spontaneous generation does exist, but in the community, it is not replicable, and so we say that spontaneous generation is not true, even though to that person (whom we perceive as mad) the spontaneous generation theory is perfectly replicable. So yes, I do think that science happens in a community. We don’t just value hearsay, after all; we look for replication as the final word, even if replication doesn’t always get published.

    I think your definition of social construction and mine are different. What I mean by saying knowledge is a social construction is that we have “community standards.” And I think you really do need community standards for knowledge, because that’s how we evaluate theories in science. Who came up with P values of 0.05 and 0.01 as the most common ones for evaluating statistical “significance”? The scientific community eventually came to a collective decision as to what a “good” P value is. Similarly, the idea that a truth needs to be testable and replicable, and to what extent something needs to be replicable, is also something the scientific community came to a consensus for. And ultimately, hypotheses are tested in a community, because no scientist would bet their lives on the evidence of one person replicating an experiment over and over.

  5. Dan Says:
    March 27th, 2008 at 5:46 am

    most science is mundane, about fleshing out the frameworks we have now. If my experiment doesn’t work, I don’t then presume to have falsified the theory of atoms and quantum mechanics.

    Right, and that amounts to the ‘bookkeeping’ aspect of science. Even within that normative mode of science however - especially experimental sciences - you design your experiments as tests of hypotheses, right? Moreover, even when you are simply doing normative science and adding details to existing paradigms, manuscripts and especially proposals are laden with the language of revolutionary science or falsifiability.

    I think the way scientists work in practice is that theories that have been well-tested are given a privileged status in later testing; not all hypotheses are created equal. So it’s not just that theories are falsified; some theories are made more “likely”, as you said.

    Privileged? To use a clichéd example, I suppose in a sense I do feel that Newtonian physics is more “likely” than alternatives in most cases, just as I think that the central dogma of molecular biology is more likely than alternatives in most cases. But I don’t think that those qualify as unwarranted views simply because they are paradigms - any critically thinking scientist would be quite open to the fact that exceptions exist or could exist, or that the paradigm is flat out wrong. Do you think that you are unreasonably biased in favor of the truth of the central dogma of molecular biology?

    Spontaneous generation failed because it didn’t live up to the standards that the scientific community set for “theory” (and yes, I do mean community).

    Exactly. As you said, “but really, when an experiment gives a result different from expectations, what tends to be falsified first?,” which implies that the experiment tends to be falsified first, not the expectations. And I gave the example of spontaneous generation, which is a prime example - for decades, some prominent scientists (take John Needham for instance) claimed that failures to prove spontaneous generation were the fault of the experiment, not the idea. And for over a century after Needham, prominent experimental biologists still espoused some form of spontaneous generation. But in the end, as you say, the theory was recognized to be at fault - falsifiable facts win out in the end, even where social constructions have long persisted.

    What I mean by saying knowledge is a social construction is that we have “community standards.” And I think you really do need community standards for knowledge, because that’s how we evaluate theories in science. Who came up with P values of 0.05 and 0.01 as the most common ones for evaluating statistical “significance”?

    Point taken. Indeed, social constructions do exist, and should exist. I still think that they only facilitate discussion of facts and how they are described (meta-facts, if you will), but do not directly influence the reality itself which we are trying to observe as scientists.

    Sorry, reading over this and my previous two comments before hitting “submit comment” I think I’m sounding pretty harsh. I don’t mean to, really - I just think that Kuhnsian philosophy of science is given more credit than it’s due - maybe Kuhn’s model should be called a “sociology of science” instead. ;-)

  6. Eric Says:
    March 27th, 2008 at 10:32 am

    which implies that the experiment tends to be falsified first, not the expectations.

    As you said, in the end, the theory was recognized to be faulty, but only after long amounts of testing and repetition, which is what I’m getting at. Self-correction in science takes a long time, I think, especially for theories that seem to be very valid.

    But I don’t think that those qualify as unwarranted views simply because they are paradigms - any critically thinking scientist would be quite open to the fact that exceptions exist or could exist, or that the paradigm is flat out wrong.

    True, every scientist works under the possibility, but you don’t see many scientists jumping up to that conclusion first, right? Shifting such monumentally well-tested theories often starts with someone saying, “Why can’t I get this to work? What am I doing wrong??”

    I really do think that science works within theoretical frameworks, with experiments preferentially falsifying certain hypotheses over others. Every “fact” is the result of a lot of interpretation and a lot of assumptions about how things have worked. I think that’s part of the reason why science demands controls so often, because otherwise we’d never be able to reasonably narrow down what is being falsified, but even so, we have very strong assumptions about how the universe works: thermodynamics, homogeneity of solutions, and so on. There are no truly stable “facts”, only interpretations of raw readouts.

    As an example, consider astronomy before Copernicus; everyone was working under the (what they thought was) well-observed phenomenon that everything revolved around the earth. So they invented things like epicycles and all sorts of extra modifications to their astronomical theories to make it fit the data, and fit the data it did, much better than Copernicus’. It was only with Kepler that the major theory started to be questioned by a large number of people. They were operating under a paradigm.

    Now, that’s a descriptive view on how I think scientists operate, but I also think there’s really no other way. We need models to tell us what experiments to do next, and in order to interpret out experiments. I think biological models tend to be more fluid and more challenge-able because the field isn’t that mature yet.

    I don’t always agree with Kuhn either; to some extent, he’s the other extreme away from Karl Popper, but I do think he has quite a few valid criticisms of Popper. Real, everyday science probably follows neither philosophy and happens somewhere in between the two, or a little further off from the Kuhn-Popper spectrum.

  7. Dan Says:
    March 27th, 2008 at 3:22 pm

    Good points, all of them. I think I’ve come a long way in evaluating the strong and weak points of Kuhn and Popper, and still have a long way to go.

    Point of fact, a couple weeks ago I got an interesting comment on an old post of mine, from a postgrad in Philosophy of Science. (No comment permalink available, but it’s at the bottom of this discussion). The interesting bit:

    Philosophy of Science has come a long way since the days of Popper and Kuhn, and I suggest that you read ‘What is this Thing Called Science?’ by Alan Chalmers (I’m not sure what edition the book is up to, but read the latest one - he keeps the book up-to-date). This is a very good book, clear and well written, and provides an excellent overview of philosophy of science (including all the major players: Popper, Kuhn, Lakatos, Laudan…) and the problem of demarcation between science and non-science.

    In particular, Dan, you may like to read the chapters on Bayesianism and New Experimentalism. Most philosophers of science shared your concerns about falsificationism versus verificationism (or deduction versus induction) - and Bayesiansm and Experimentalism provide some sort of response to these concerns.

    Be sure to read the earlier chapters first - that way you will understand how philosophers have arrived at these later views.

    That comment just put Chalmers’ book at the top of my Amazon Wishlist.

  8. Eric Says:
    March 27th, 2008 at 6:24 pm

    I’m in the middle of reading a book by Deborah Mayo, which was published a few years ago, called “Error and the Growth of Experimental Knowledge.” Problem is, it’s more of “primary literature” type book instead of a survey, so it isn’t something I can easily pick up in between experiments. I’ll check out the Chalmer book; thanks for the suggestion!

Leave a Reply