Popper

It's a truism that the favorite philosopher of every scientist is Karl Popper. (In my own experience, this truism is mostly true.) Popper, or so the story goes, stood up for empirical fact when the post-modernists were descending into Deleuze and Derrida and difference. His popularity among experimentalists is also a side-effect of simplicity, as his fundamental idea is easy enough for just about anyone to understand: Popper famously pointed out that science never proves things true, it merely proves things false (this is the falsifiability doctrine). In other words, scientists proceed in stuttering steps, advancing by saying what theories are wrong. The truth is just what (temporarily) survives.

At first glance, it's easy to not notice how radical an idea this is. But look closer. For Popper, all discovery is really just criticism. Although we think of scientific truth as being somehow more stable than literary or cultural truth - literary fashions come and go, but gravity remains - the opposite is actually true. Scientific truth is true precisely because it is open to change, willing to reverse itself and admit its errors. Les Demoiselles d'Avignon will always be a magnificent painting, but a scientific idea that is no longer true...well, what good is that?

But here's my question (and it's a question that I really don't know the answer to): If Popper's model of science is reasonably accurate - and plenty of scientists think it is - then I'm a little worried about what this means for the distant future of science. After all, if our knowledge of everything emerges from criticism, from the falsification of yesterday's truth, then the scientific process can only keep going by proving itself wrong. If our truths were ever perfect - if science could ever come up with a theory of everything that wasn't flawed or couldn't be improved upon - then those truths wouldn't be scientific, for they wouldn't be falsifiable. According to Popper, a science without limits or imperfections or fixable flaws isn't science: it's metaphysics. In this sense, the end goal of science - the construction of a perfect mirror to reality - isn't just unrealistic: it's also unscientific.

This isn't directly related, but I love this quote from Richard Feynman: "Philosophy of science is about as useful to scientists as ornithology is to birds." I happen to think that isn't true - philosophers can actually be quite useful - but it's still pretty pithy.

More like this

My original post on the (possible) limits of science generated lots of thought-provoking feedback. On the one hand, some people argued that I was conflating the persistence of statistical uncertainty with genuine mystery: Of course, there are built in uncertianties in science especially with the…
One of the issues involved in the evolution/creationism battle is the question of demarcation - what separates science from non-science? One of the most popular and, in my view, compelling arguments against Intelligent Design Creationism (IDC) is that it is not a genuine scientific theory at all…
Revere stirs the pot (of chicken soup) to ask why alternative therapies are presumptively regarded as pseudo-science. The reflexive response of the quackbusters has been that alternative therapies fall on the wrong side of some bright line that divides what is scientific from what is not -- the…
Apparently, Martin Cothran believes that there is no life elsewhere in the universe, and that this unimaginably vast emptiness is evidence that a god created us. I don't understand the logic, but then I don't understand most of his weird leaps in this post on how life on other planets is like…

You're making a fundamental mistake: "falsifiable" does not mean the same thing as "false". Popper would not deny that a completely true theory can be scientific; he would only deny the scientific nature of a theory that could not be proven false by any logically possible experimental result. For instance, general relativity makes certain predictions about observable phenomena (the bending of light due to the presence of a gravitational field). If we observe these phenomena and find that the predictions are not fulfilled, then the theory will be proved false; thus, it is falsifiable by some result that we can imagine. That the predictions of the theory are fulfilled does not render it unscientific in a Popperian view.

Michael Clarkson's comment is right on. When Popper was talking about "falsifiability", he was referring to the way in which a theory was presented; a genuine scientific Theory (as opposed to just any old statement about how things work) has to be stated in such a way that there is some logically derivable and eventually possible physical test of some critical point of the theory: some event predicted by the theory must be observable, either to proceed the way the theory predicts that it will, or not to proceed in that way, which would falsify the theory. And a falsified theory is not necessarily a useless theory; its predictions about a great number of important events may have survived every attempt to falsify them, and we may have ample reason to trust our lives to the future predicted by the theory. The falsification of some part of it simply means that it is no longer as complete as it presented itself to be; it still needs work.

Richard

you might be surprised by how many scientists (at least in my field - cognitive science) take a more Kuhnian view of philosophy of science. perhaps this is a product of the fact that what we try to study is the nature of linguistic and conceptual representation itself (and which most cognitive scientists would argue is not like the naive-realist, common-sense perspective)...

that said, it's hard to actually DO science without taking a more pragmatic or Popperian attitude...

I am somewhat of a Popperian (lacking complete understanding of some of his work). It seems to me that science is the study of things we don't completely understand. I read someplace that geometric optics is competely understood, and therefore is no longer a part of science. Does this mean that we cannot picture finding error in geometric optics?

I think Kuhn is more of a sociologist of science, and less a philopsopher of science. The last comment of his I have seen was in a Chronicle of Higher Education some years ago, "We do not know how science progresses." I think his view of shifting paradigms has some verisimilitude. For example; fixed continents to plate tectonics during less than a decade. On the other hand I think the progress of genetics has been pretty seamless from the redicovery of Mendel to the present.

By Jim Thomerson (not verified) on 19 Jun 2008 #permalink

According to Popper, a science without limits or imperfections or fixable flaws isn't science: it's metaphysics.

This is not that strong a criticism of Popper, as others have mentioned - in the case of the perfect theory, the limits, imperfections and flaws of the science simply correlate with the limits, imperfections, and flaws of reality.

The really cutting criticisms of Popper have to do, typically, with how his theory works in the real world. Popper's falsificationism, historically, was pre-Deleuze and Derrida. His theory was more a reaction against logical positivism (operational definitions and all that). It has been argued by Stove that the postmodern view of science is a direct consequence of Popper's thought, in fact.

And Popper's philosophy of science has been shown to have some massive problems:

1) In practice, one observation never does falsify a theory - theorists can always point to variables that the falsifying experiment didn't control and say that they caused the falsifier's results (especially in psychology where there are a million things to control).

2) Psychology and neuroscience, in particular, use statistical probabilities to determine the truthfulness of a hypothesis; thus, in these sciences, Popperian falsificationism cannot falsify theories. They can only prove them improbable. This very much weakens the power of falsificationism.

3) The most famous examples of theories which Popper argued were unfalsifiable (Darwin and Freud) are quite poor. Aspects of both Darwin's theory of natural selection and Freudian psychoanalysis have been either falsified or could have been falsified - e.g., DNA would look very different without natural selection, and if psychoanalysis were true, it would be a much more effective treatment method than, say, CBT.

It is unclear, therefore, what is actually falsifiable, and it is unclear what is unfalsifiable. This is quite a big problem for Popper, I would say.

By Tim Byron (not verified) on 19 Jun 2008 #permalink

As a modernist piece, Les Demoiselles may not have been viewed as beautiful or true to a connoisseur during the high Renaissance. The figures depicted, ladies of the night, are obviously not cherubic or Rubenesque, neither are they realistically portrayed as in, say, Manet's Olympia.

Artistic magnificence may be timeless in some sense, but I'm not sure that's equivalent to saying its truth is immutable. One might quite rightly assert, for example, that the Ptolemaic system of cosmology is magnificent--while recognizing that it does not reflect our truer understanding of the cosmos. Perhaps scientific ideas that are no longer "true" possess instrinsic aesthetic value.

I just only disturst Feynman´s quote, philosphy and science lie in a continuum. The philosopher Alva Noë says that scientific questions are philosphical in origin and philosophical questions need a scientific treatment.

Agreed, there must be some philosophical quest underlying scientific endeavour. We ask questions that are important to ourselves and hopefully to mankind. It may not necessarily correspond with a homogeneous worldview across different fields, or even the same ultimate goal, but they share the philosophical mode of questionin. Its certainly not done for the financial gain...

By Mike Woods (not verified) on 20 Jun 2008 #permalink

"...I'm a little worried about what [Popper's model of science] means for the distant future of science. ...if our knowledge of everything emerges from criticism, from the falsification of yesterday's truth, then the scientific process can only keep going by proving itself wrong..."

Remember though, the scientific process also includes *building* theories. These are whittled down by the process that Popper most focuses on, but the building is still important; it's just that he - perhaps rightly - leaves people free to use whatever method of initial theory construction that takes their fancy, and only attaches important criteria to the "whittling away" aspect. Nonetheless the construction of useful models allowing us to predict the future and prepare to deal with it, or to choose action adequately suited to the environment, or even to update our knowledge efficiently, is still the core operation.

"If our truths were ever perfect - if science could ever come up with a theory of everything that wasn't flawed or couldn't be improved upon - then those truths wouldn't be scientific, for they wouldn't be falsifiable. According to Popper, a science without limits or imperfections or fixable flaws isn't science: it's metaphysics. In this sense, the end goal of science - the construction of a perfect mirror to reality - isn't just unrealistic: it's also unscientific."

Much of philosophy - indeed science, any knowledge, or even life - is about clarifying concepts, and this is an excellent example. It's only theories that are intrinsically untestable by their very nature that are unscientific; theories are not untestable simply because they have not yet been refuted. If we seek some category of theory with some kind of perfection beyond simply surviving refutation, it's beyond anything dreamt of in Popper's philosophy, so Popperians need not worry.

"This isn't directly related, but I love this quote from Richard: "Philosophy of science is about as useful to scientists as ornithology is to birds." I happen to think that isn't true - philosophers can actually be quite useful - but it's still pretty pithy."

You love it? I hate it! As with: "How much does David Beckham know about aerodynamics?" it gives useful material for both sides of the argument, but most of all it gives thoughtless disparagers a licence. "I don't understand gravity but I still don't fly off into space, so physics is useless." A personal philosophy is simply styles and patterns of thinking; you can study them and try to improve them but you don't escape having a philosophy simply by trying to denying it. Don't forget, physics and maths are special cases within which to theorise. Unfortunately, the people most famously considered "clever" tend to be physicists, and their pronouncements on philos. of sci. get listened to, even though they are limited in scope. Contact with biological or historical sciences soon helps one realise that Hawking's claim to be a positivist (and not, presumably, a Popperian) need not matter, any more than Hoyle and Wickramasinghe's views on Archaeopteryx... or Feynman's on ornithology.

Great blog.

PS - Re Steve's comment - doesn't Kuhn say more about what scientists do or tend to do, whereas Popper said what they should do?

Re Jim Thomerson - Kuhn may be right, in the sense that there's still lots we don't know about how scientists do science (or about how they should do it.)

Re Tim Byron - yes, the Popperian approach has to contact the world via probability. This adds huge complexity, but natural cognitive and even just genetic systems prove this extra layer doesn't make things intractable.

By Strangetruther (not verified) on 26 Jun 2008 #permalink

The map is not the territory nor can ever represent in a 1:1 scale (if so it would be the territory itself, not an abstraction, not a model, not a formulation, not a map).

This doesn't mean I am Popperian even if agree with him in this. For what I've read, I prefer Deleuze to Popper because I find the Austrian too plain, naive and conformist.

In fact I suspect this issue of the intrisecal impossibility of a totalistic model actually unites Popper with Derrida's différance: they are both talking about what is not, not what it is; because what it is exactly, in all its dimensions is unspeakable, unthinkable, uncomprehensible. No computer or brain could ever understand all in a single fraction of reality, much less in all Reality. That's why our attention gets focused or deactivated, why our senses are limited, why our memory is fragile: there is no way to capture it all much less understand it, so better keep focused in specific areas of interest, either by nature, need or choice.

Whether or not Jonah was exactly right on in his characterizations of Popper's theory is beside the point--the idea at the end of the post is the crux of the issue:

According to Popper, a science without limits or imperfections or fixable flaws isn't science: it's metaphysics. In this sense, the end goal of science - the construction of a perfect mirror to reality - isn't just unrealistic: it's also unscientific.

Ive often thought about this. Sciences sheer pursuit of the perfect mirror of reality including pursuing reality down to its smallest, imperceptible pieces (DNA, atoms, etc) is actually a kind of self-cannibalism. If science were to achieve absolute accuracy in describing the world, it would eliminate the need for its own existence. And, as Jonah points out, the kind of truth it would create would no longer be a scientific truth, because it would be absolute. Scientific truth can be seen as intermediary--truth that leads to further truth. But absolute truth leads no further than itself. Its a little like the old riddle about God: Could God create a boulder so heavy He could not move it? Could Science discover a truth so great It could not improve it?