First of all. Yes, the title I’ve chosen makes me blush, but hey, I’m just following my own plan!
This post springs from a tweet by Keith R Laws (thanks!), that linked to Imre Lakatos commemorative pages, hosted by LSE. I did not know about Lakatos work and it has been a delightful discovery. It also sent me into confusion, because I found myself agreeing with the general points made by Lakatos, and consequently unsure whether I needed to reconsider my own position or not. This post is the result of my re-evaluation, prompted by the discovery of Lakatos’ argument.
First: what does he say? I will not explain his reasoning in detail (please read the transcript instead), but will quote his key points, to see if they challenge or agree with what I’ve written so far.
The problem of the demarcation between science and pseudoscience is not merely a problem of armchair philosophy: it is of vital social and political relevance.
I agree with all my heart: this is why this blog starts with all this epistemology, and also why I’m finding myself unable to move on and tackle more practical matters. I feel that I have to get this right, especially because what I’m planning to do has high risks of being pseudoscience.
Back to Lakatos, he convincingly explains why the solution offered by Popper doesn’t work, and concludes with:
Is Popper’s falsifiability criterion the solution to the problem of demarcating science from pseudoscience? No.
I can agree with this as well, without having to revisit anything I’ve written so far. The next step, naturally, is to see if Kuhn offers an answer, but (surprise, surprise!) he doesn’t:
But if Kuhn is right, then there is no explicit demarcation between science and pseudoscience, no distinction between scientific progress and intellectual decay, there is no objective standard of honesty.
Again, I have no problem with this conclusion, Kuhn’s theory indeed ultimately rejects the epistemological superiority of the scientific method. On the other hand, I have explicitly supported this superiority, and now face the need to substantiate my claim a little more. All I’ve said so far is that whatever belief is supported by observations (evidence), has to be considered more reliable than what isn’t supported. I have also argued that all our beliefs are ultimately the result of one or more purpose that animates us. In other words, what we regard as true is never true in a vacuum, we consider it true because this belief allows us to achieve something (from opening my front door, to understanding why saying this or that may or may not make my girlfriend angry).
Lakatos seem to disagree with my stance here, as he starts by explicitly saying that:
The cognitive value of a theory has nothing to do with its psychological influence on people’s minds. Belief, commitment, understanding are states of the human mind. But the objective, scientific value of a theory is independent of the human mind which creates it or understands it. Its scientific value depends only on what objective support these conjectures have in facts.
Now, this is apparently troublesome, as I clearly can’t support this statement. I am essentially denying the possibility of unequivocally establishing if a theory is supported by objective facts, all I claim is that it is possible to rate the usefulness of competing theories, depending on which one happens to better describe what we’re dealing with. Implicitly, I’m also saying that under different circumstances, this rating is expected to change: in some cases, theory A will be better than theory B, in some others, the other way round. The significant variable being what are you planning to use the theory for. I don’t need the laws of thermodynamics to cook a fried egg, but they come in handy if I’m trying to build a fridge.
Lakatos defines something similar to Kuhn’s paradigms (or does he?), and talks about Research Programmes:
Now, Newton’s theory of gravitation, Einstein’s relativity theory, quantum mechanics, Marxism, Freudism, are all research programmes, each with a characteristic hard core stubbornly defended, each with its more flexible protective belt and each with its elaborate problem-solving machinery. Each of them, at any stage of its development, has unsolved problems and undigested anomalies. All theories, in this sense, are born refuted and die refuted. But are they equally good? Until now I have been describing what research programmes are like. But how can one distinguish a scientific or progressive programme from a pseudoscientific or degenerating one?
And the solution is:
all the research programmes I admire have one characteristic in common. They all predict novel facts, facts which had been either undreamt of, or have indeed been contradicted by previous or rival programmes.
In degenerating programmes, however, theories are fabricated only in order to accommodate known facts.
He doesn’t talk about purposes, but I believe the concept is implicit: a good new theory allows to make new predictions, and new predictions will be useful to someone, somewhere, facilitating the adoption of the new theory. Lakatos calls a programme that has this ability to produce new predictions “progressive”, and crucially adds:
As opposed to Popper the methodology of scientific research programmes does not offer instant rationality. One must treat budding programmes leniently: programmes may take decades before they get off the ground and become empirically progressive. Criticism is not a Popperian quick kill, by refutation. Important criticism is always constructive: there is no refutation without a better theory.
I couldn’t agree more, and will spend a few more words to show how my own contributions fit in Lakatos view.
In the previous posts, I’ve defined the scientific attitude (see also this post), and added two concepts: the first one is that a correct scientific attitude requires to be ready to throw away a theory if needed, no matter how solid it is, and that when working within a given theory, there needs to be a clear and indisputable drive toward solidification. The second concept is equivalent to Lakatos’ dichotomy between degenerating and progressive programmes, I don’t think I need to discuss this similarity at length: a degenerating program is one that is losing its potential to improve, it simply can’t be solidified much more. A progressing programme is generating new predictions, and further developments are consistently expanding its explanatory powers. The more correct predictions are made, the more solid the programme becomes (and paradoxically, the less interesting it gets). What I haven’t been able to explain follows Lakatos contribution, and revolves around when it is appropriate to abandon a given theory, the details are below.
Conclusions, according to both Lakatos and me:
- Albeit we started from very different positions, Lakatos and I fundamentally agree on what distinguishes science from pseudoscience. To use my own definitions, pseudoscience happens when someone sticks to a given theory even if:
- Another already developed theory is demonstrably better at describing the observable phenomena (can make more accurate predictions).
- And/or, despite all efforts, no new useful predictions are being generated via the old theory, while there may be evidence out there that another competing theory is emerging.
- On the other hand, speculating with the intent of generating a new theory is risky (as it may fail), but is not inherently unscientific.
I also may add my own contributions:
- There is nothing wrong in applying an old, surpassed theory to make predictions, if and only if these are the predictions that solidified said theory. In this sense, my notion of purpose justifies the survival of theories that are shown to be imperfect (the purpose drives the choice of the appropriate theory). This is important for my discourse because I postulate that all theories (I normally call them Models) are imperfect, always and without exception.
- The whole explanation also resonates with my concept of cognitive attraction: Lakatos explains how a programme has some core ideas (strong attractors) and a “vast ‘protective belt’ of auxiliary hypotheses” (weaker attractors that depend on the strong ones). A degenerating theory will react to refutations by changing, adapting and strengthening the protection belt, leaving the strong attractors in place. My point being that this process is irrational, but entirely natural and predictable.
- Together with cognitive attractions, my focus on purpose also allows to attempt an elaborate rescue of Kuhn (if I may): he observed and highlighted the fundamental irrationality associated with the onset of “scientific revolutions”. Lakatos refutes this view, explaining why after all, changes happen for rational (?) reasons. My addendum here is that (pretending to accept that we can agree on what is rational), I can produce a more precise model. Say that theory A is solid, and that N “A” scientists successfully work with it, pay their bills and enjoy well deserved pints. Along comes theory B, that is still less solid than A, but seems to produce some better predictions. Trouble is, theory B is built on new maths, or some other concept that the “A” scientists are unfamiliar with. What should the “A” people do? Their purposes include: pay their bills, drink their pints, and avoid unnecessary efforts (because of cognitive attraction, adopting theory B requires a lot of effort), so they will resist the change. However, the change will probably happen despite the resistance of the “A” camp. With time, new scientists, familiar with B, will be formed and will eventually take over; while this happens, certain (probably young) supporters of “A” will certainly do the jump as well. That’s an account of the (apparently) irrational resistance (shown by Kuhn), along with the explanation of why and under what circumstances the shift may happen nevertheless (following Lakatos). Yes, it’s obvious, but someone still needs to write it down.
I’m left with an open question: is this whole post a sign that my theory is already degenerating? I have my own idea, but I’ll leave the answer to you.