To: Edoardo S.
From: Geoffrey Klempner
Subject: Scientific method: indisputable truth or real hoax?
Date: 29 January 2008 12:11

Dear Edoardo,

Thank you for your email of 17 January, with your essay towards the Associate Award, 'The scientific method: an indisputable truth or a real hoax?'

I enjoyed reading this essay. It's main strength is that it explains the development of ideas about scientific method in the twentieth century in a clear and simple way, which newcomers to philosophy, or the philosophy of science, would be able to understand and appreciate.

I did, however, have some difficulty in determining exactly what question you are trying to answer. The title of the essay is obviously one clue, but even here there seem to be different ways of understanding the question.

There are three issues -- maybe more, but I will focus on three:

1. What is the correct scientific method? This is a normative question, which assumes that there is, in principle, a way of distinguishing 'correct' science from 'incorrect' science, or 'good' science from 'bad' science. As such, it belongs to epistemology.

If you are answering this question, then Francis Bacon is a figure who has to be considered, because Popper's work is largely in opposition to Bacon's view that the correct method consists in formulating general propositions on the basis of induction from observed instances.

2. How has science, in fact, progressed, in, say, the last 300 years? This is a question for the history of science and the history of ideas.

Kuhn's argument in 'Structure of Scientific Revolutions' is largely based on the observed discrepancy between, e.g. Popper's falsifiability model and the actual historical progression of science.

It would still be possible for Popper to assert, in response to Kuhn, that, according to his falsifiability theory, the majority of scientists have done science in an 'incorrect' way. In other words, the debate here is over the significance of these historical facts. Do they show that Popper's view of how science should ideally be conducted is wrong? Or do they show that Popper is right, but that much of science fails to fully live up to this ideal?

Feyerabend merely takes Kuhn's ideas one step further. It is a waste of time even trying to formulate the 'correct method'. All that matters is the results. Is this a reductio ad absurdum of Kuhn's approach? A critic of Kuhn would argue that 'lucky' successes in science (i.e. discoveries which did not arise out of the 'correct' use of scientific method) do not show that scientific method is useless, any more than beginner's luck in Poker shows that it is unnecessary to be a skilled Poker player if you ultimately want to succeed as a Poker player.

3. A third question, which comes into your essay when you consider Heisenberg's Uncertainty Principle, is the question of the nature of truth, as such, as distinguished from what human beings can know through scientific investigation.

Popper is a full-blooded realist about truth. He believes that there is a 'third world' out there which is the way things really are, but we cannot make direct epistemic contact with this world. The only way that we can approach the truth is indirectly, through the method of science, which only yields conjectures which stand the test of falsifiability. We never attain the 'truth as such'.

Much of what Popper says would remain unchanged if we was an anti-realist about truth -- indeed, the same applies to Feyerabend and Kuhn. So there is an issue here about the relevance of raising the question of realism in the context of an examination of scientific method.

Here are two further points, one minor and one major:

On page 2, you accidentally switched 'deduction' and 'induction'. Induction passes from the particular to the general. E.g. I observe several white swans and conclude, by induction, that 'all swans are white'. Deduction passes from the general to particular. E.g. All men are mortal. Socrates is a man. Therefore, Socrates is mortal.

The latter is not a good definition of deduction, however, as there are many cases of deduction which do not fit this model. E.g. P or Q. Not P. Therefore Q.

There is also a serious unclarity in your account of the verification principle and Popper's falsifiability criterion.

The verification principle is a criterion of meaningfulness. Any proposition which fails the test of verifiability is meaningless. By contrast, Popper's falsifiability criterion is a criterion for whether a proposition is scientific.

You unsuccessfully try to cover up the gap by saying, 'From a scientific point of view, a theory which cannot be contradicted is devoid of any sense.' However, this is clearly wrong. Astrology is a prime example of a belief system which is not scientific, but it is not meaningless. I agree that if you take the side of science then Astrology is a 'senseless' thing to believe, but astrological beliefs are not, literally, devoid of sense. Rather, they are irrational, and (to our best knowledge) false.

I should also mention that a complicating factor is that there are two versions of 'verificationism':

a. Verifiability principle, which we have been discussing.

b. Verification theory of meaning, according to which the meaning of a proposition can be given in terms of its verification conditions.

For the purposes of the present discussion, version b. can be ignored. In the 1970's the idea of a verification theory of meaning underwent a revival through the work of Michael Dummett and his 'anti-realist' account of the form of a theory of meaning.

All the best,

Geoffrey