Karl Popper essentially falsified the inductive method of determining what is so.
Induction is a now debunked methodology, which once held sway as orthodoxy in the 19th century, comprising the idea that knowledge about something can be obtained from past observations (or events) of it. Famously, Popper put a stop to that idea by using the example of what we once thought we knew about swans. If the only swans that had ever been observed by those classifying them were white was it correct to say that all swans are white? No. Because after black swans were discovered in Australasia those previously unimagined swans overturned existing knowledge. Popper essentially demonstrated what is today known as the fallacy of induction.
The impact of the improbable, based on Popper’s (1959) fallacy of induction, is a theme that has been very successfully exploited by many authors. One, whose best selling book I am currently reading, is Nassim Nicholas Taleb.
Taleb’s book, The Black Swan: The Impact of the Highly Improbable (Allen Lane/Penguin Books 2007), develops the black swan example in many areas of life to assert that social science and economic experts essentially know little more about their subject than the man in the street. The reason for this is that they fail to take account of chaos and uncertainty in the world.
In the wider social affairs of man there are currently too many possible and complex ‘known unknowns’ and ‘unknown unknowns’ for us to accurately predict what will happen next. Yet, surely, it seems fair to argue that if you cannot say what something - that is not completely subject to unknowable random forces - will do next then you really don’t understand it at all. This means that a lack of genuine expertise is masquerading everywhere. Terrorism experts, for example, failed to predict 9/11. Economists failed to predict the economic crash. And in my own field we criminologists all failed to predict the 15 year crime drop in the western industrialized world.
The point of this blog post is that I have an issue with Popper’s fallacy of induction. One that the self-admitted black swan obsessive Taleb has, as far as I can tell, failed to see. My issue is that Taleb, among other Popperians, does not allow for the possibility that the entire fallacy of induction would be falsified itself if a methodological mega black swan event came along. Something, for example, as seemingly improbable today, as progress in quantum computing serendipitously making it possible to accurately predict future ‘black swan-type events’ in the affairs of man by analyzing past events.
My argument here is that the fallacy of induction is in fact, ironically, based on induction because the falsification of induction is itself based on past knowledge of the failure of the inductive method to know the present and predict the future.
If a new discovery does overturn existing methodological orthodoxy so that induction becomes a good method of knowing the future we should be prepared to name such an event: The First Coming of the Antiswan.
I would like to thank Professor Michael Smithson for kindly sharing his thoughts on my ideas on this issue: Click here to see our brief discussion about this issue on his blog.
Popper, K. R. (1959) The Logic of Scientific Discovery. London. Routledge.
Taleb, N. N. (2007) The Black Swan: The Impact of the Highly Improbable (Allen Lane/Penguin Books.
Thank you for your most helpful additional information.
Regarding whether or not Popper's refutation of induction was based on its past failures. But surely, had induction been a successful method of predicting the future and, consequently, for knowledge development then Popper could not have refuted it. So it must - ultimately - have been refuted for its past poor performance. And the swan analogy is testament to that.
The fact that we cannot know with certainty that induction might not one day turn out (in some way that is currently unimaginable) to be something more than a 'confirmation report' surely means that Popper's refutation of induction was merely a confirmation report that induction is merely a confirmation report? Is that not inductive?
August 31, 2011 at 9:20 am
Popper's response might be...
Any so called inductive knowledge must be based on some presuppositions which organize the framework in which the induction takes place. For example, the induction based upon the observation of swans relies on assumptions about biological species, the presence of observable objects on the planet, etc. For this reason an inductive proof is really just a confirmation report: the conclusion 'all swans are white' which is reached at the end of an inductive proof is really just an explanation which is confirmed by the many instances cited in the premises. Past observations (represented here by the premises of the induction) are compatible with an infinite number of possible explanations. Induction leaves out the role of assumptions (ie. concepts or background knowledge), expectations and selection which guides the development of knowledge.
Popper does not base his theory on the past failures of induction. Induction is rejected because it is a bad explanation for the development of knowledge (it is also logically problematic but we'll ignore that for now). Induction had be conjectured, and Popper refuted it. There is no backhanded, hidden induction lurking here. Certainly past evidence plays a role in the formulation of this non-inductive theory. But that does not necessarily imply that the fallacy of induction is itself based on induction.
The reality is that people are unlikely to abandon decision making based on the past behaviour of systems just because it is logically flawed. Having said that I find it very difficult to say any given event is impossible any more. However, I am quite prepared to risk my life on the assumption that said event has a vanishingly small probability. And we probably all do everyday. It allows us to get by - and has a certain ecological validity!
Dear Dennis, Many thanks for the feedback. Yes I agree with you. And the point has been made by psychologists writing about irrationality and prior expectation - such as the late Stuart Sutherland. As you point out, we need to refer to past events in order to imagine what might happen next given the data currently available to us.
That said, I think that the danger – both in knowledge progression and in our everyday lives - comes in using such past data (in this case personal memories or held stereo-types) to believe that we are predicting the future - or interpreting the present with a fair degree of accuracy - say (as a guess) above 75 per cent accurate in order to know what rough probabilities should guide you. Staking our lives on past events might even turn out to be worse than using our imagination to capitalise on safer known unknowns - without being knowingly reckless. Who knows?
How much safer is it, statistically over the course of a year for example, for a middle class person to walk up a dark alleyway frequented by muggers while swaggering along in a thick leather coat - with a pistol or machete tucked away - than it is for him to walk home along a well-lit footpath in a suit coat unarmed? Who really knows?
And, similarly, walking up the 'safer' side of the street to avoid such rough looking types and panhandlers on the other side might in fact increase your unknown and, anyway, uncalculated risk of being hit by a furiously pedalling reckless middle class cyclist coming downhill on the pavement from behind you just as you suddenly 'unexpectedly’ step to the left to avoid stepping on a risky banana skin.