Popper on non-inductive science

Working with the Scientific Method and Hume’s work on induction can reveal some difficulties. This post outlines  Popper’s response to Hume’s problem, and some issues with his attempt. Which will lead to a discussion of Lakatos’ refinement of Popper, where he tries to evade the problems that Popper encounters.

Firstly, Hume’s problem of induction highlights that there is no good rational grounds for accepting the conclusions of inductive arguments. If we take the claim that “the sun will rise tomorrow”, we can’t justify it deductively, as it is not a logical contradiction to say that the sun wont rise tomorrow, to try and justify the claim inductively relies on the Uniformity Principle, which is the idea that the future will resemble the past, as it always has done, but this is inductive in itself, so Hume argues that this is a circular argument, and of course these a very poor arguments.

Therefore, if we cant really on induction this has implications for science, as science uses induction a huge amount. However, we think it is rational to treat illnesses in a certain way because our knowledge of the things related to the treatment leads us to reason that it will cure the illness. However, Hume says that there is no justification for inductive practices it is all just habit.

There are responses to Hume, I like the one that Hume as a naïve view of induction when it comes to issues in science, by which I mean, it may be the case that everyday generalisations made by the average person who knows only the basics of science may be hard to justify, but perhaps we could say that scientific induction is based off of deep scientific factors, that make it rational?

However the main reply that I will talk about is from Karl Popper, he wanted to argue that we can evade, note not solve, the problem of induction when it comes to science by suggesting that science isn’t actually based on induction at all.  Known as Popper’s Falsifiactionsim, it is a non-inductive approach to science. The basic idea is that no number of positive instances provides rational support for a generalisation, but a counter example can refute it. So for example, all swans are white, seeing a load of white swans does not prove anything, but observing a black swan does refute the generalisation. For Popper, a theory is only scientific if and only if it is falsifiable, it has to be able to be refuted by experience.

So the falsifiactionist solution to induction would be that we could propose a hypothesis and then test it. We get either one of these results

  1. It is falsified, so we reject the hypothesis
  2. Hypothesis is borne out, so the theory is supported, but note, that supported is not the same as confirmed.


  1. A worry about practical predictions: do you really want the people that make aeroplanes to build things that are only supported by tests and never said to be confirmed, is this taking a risk? To use only corroborated things as a guide for the future.
  2. History of science refutes Popper: if we look at major theories in the past, we find that they always face falsifying instances right from the outset.
  3. Duhem-Quine: Falsification can happen as a result of just one auxiliary (outer hypothesis) being wrong, but logic doesn’t tell us where the blame lies. Therefore if we want to test hypothesis then we must have auxiliaries that can be confirmed.


He wants to get around these problems, specifically I will talk about 2 and 3. So when we look over the history of science we see lots of theories that confronted anomalies, yet they were not refuted. This shows that many scientists are not falsificationists, rather as Khun suggests scientists hang on to core principles of their scientific field and try to find different ways of resolving the anomaly, and as long as their changes are not too ad hoc this can still be a scientific approach.

So now we can talk about Lakatos’ scientific method. What he aimed to do was develop Poppers ideas in a way that would allow for continuity in the face of anomalies without IMMEDIATE falsification, so we can see Lakatos’ as blending Khun and Popper to get a new theory. He wanted to address Duhem-Quine and also develop scientific investigation so it was not just a theory but also a research programme.

What is a Research Programme? It is a sequence of theories which is made up of

  1. A hard core which are the core theoretical features
  2. A protective belt – these are the auxiliary features which can be altered
  3. Heuristics. We have negative ones, which is an injunction not to change the core and positive ones which means a plan to alter the auxiliaries and turn anomalies into victories

The protective belt is altered because a research programme will make unrealistic assumptions in its early stages. It means that if a research programme makes false experimental predictions it can be protected from immediate Popperian falsification. Don’t throw the baby out with the bath water type thinking. The idea is that science progresses by following heuristics not immediate falsification.

Falsification is a part of Lakatos’ idea, as we can get rid of degrading research programme if it doesn’t predict new things and any alterations to the protective belt are just ad hoc in an attempt to explain things. Where as a progressive research programme is one that does the opposite. A revolution is where we trade in a bad research programme for a good one.


Popper faces the Duhem-Quine problem, as we can’t know what caused the falsificationsim, thus we can never really do experiments that rely on auxiliaries that have to be confirmed. How Lakatos helped with this with his research programme style is that we can look to the hypothesis and auxiliaries and initial conditions and we would then decide what aspect to reject based on the core of our research programme.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s