| Author: | |
| Website: | |
| Page title: | |
| URL: | |
| Published: | |
| Last revised: | |
| Accessed: |
The full title of our third assignment was "Alternatives to Inductivism: Falsificationism". Falsification is the term used to describe the principle that a legitimate scientific theory must be falsifiable. In other words, a theory can only be considered scientific if it can be falsified (proven wrong) by an empirical observation. In answering the questions posed in the assignment, we were asked to consider the work of philosophers Karl Popper1, Pierre Duhem2 and James Ladyman3 (see notes at the end of this article). As I have in previous articles in this series, I list below the questions posed in the assignment, followed by my answers:
Why did Popper think that merely having a large collection of confirming instances was insufficient for a hypothesis to be scientific?
"The 18th century saw huge advancements in the physical sciences which encouraged many to think that a scientific approach could be applied to the study of human behaviour - both that of individuals and of social groups. Prominent theories included those of Marx, whose achievements, at his funeral, were likened to those of Darwin, and Freud, who proclaimed his own achievements in the field of psychoanalysis to be akin to those of both Copernicus and Darwin.
Popper initially gave credence to the principles of both Marxism and psychoanalysis, but quickly came to regard them as being pseudoscience. He set himself the task of finding out what separated the theories he believed were truly scientific, such as those of physics, from those that he felt were not. He attempted to examine the ways in which both science and what he regarded as pseudoscience were practiced.
Popper came to feel that popular – and, it has to be said, successful - branches of what he considered to be pseudoscience were in no small measure a victim of their own success in the sense that the commitment and enthusiasm of their adherents had a negative impact on their objectivity. The problem, as Popper saw it, was that it was simply too easy to accumulate positive examples of cases that supported some theory, especially if that theory was somewhat broadly defined.
For example, the view that class struggle is the main driver for radical social change is central to Marxism, yet the fact that strenuous efforts were made during the nineteenth century to improve the lot of the worker did not undermine Marxist theory; rather it was seen as evidence that the ruling classes were attempting to placate the working classes because they had anticipated an uprising and were trying to forestall it. Similarly, psychoanalysts appeared to have an explanation for just about any kind of psychological response in a subject.
Popper was suspicious of the theories he came to regard as pseudoscientific not because they could not explain certain things, but because they could explain virtually any observation, even if it appeared at first to be contradict the theories concerned. He came to doubt, as well, that the simple accumulation of positive instances of theories and laws was sufficient reason to assume the reliability of a particular body of knowledge. In essence, he felt that the generality of the theories meant that any observations could be made to fit them, particularly if those observations were subject to somewhat less than objective scrutiny by the adherents of the theories."
According to Popper, what's at the heart of the scientific method?
"Popper found himself drawn to scientific theories that appeared to lay themselves open to being proved wrong. Perhaps the very boldness of such theories said something to him about the commitment of those who proposed them to seeking what he felt was genuine scientific knowledge. After all, to make specific claims of a scientific nature is to lay oneself open to criticism, possibly even ridicule, in the event that a claim turns out to be false.
Einstein's general theory of relativity was a case in point; it predicted that light passing close to the Sun would deviate from its course (normally, light travels in a straight line) due to the gravitational field generated by the Sun. This is a very specific statement about the behaviour of light that could very easily have been wrong. Yet it was this element of risk - the very real possibility that a theory could potentially be proved spectacularly wrong – that Popper seems to have found so compelling.
Many of the theories that inspired Popper were pushing the boundaries of science, often predicting something that was hitherto a completely unknown, like Mendeleev's prediction of the existence of the elements gallium and selenium, which was based on nothing more than simple extrapolation of the existing patterns he could see within the Periodic Table.
What also impressed Popper was the fact that most of the scientists who proposed such theories were also prepared to reject those theories if it turned out that observational or experimental data did not support them. This aspect of the scientific practices observed by Popper lent them a degree of intellectual respectability that he obviously found lacking in what he felt was pseudoscience. He essentially argued that observations that confirm a theory are only really meaningful if the possibility exists that the same observations could have falsified the theory.
To take Einstein's predictions as an example, if the data had come back showing that light does not in fact deviate from its existing path as it passes close to the Sun, there is no argument whatsoever that could make the facts fit the theory - unlike the theories of Marx or Freud. For example, the very nature of Marxist philosophy means that those challenging its veracity are open to accusations of doing so purely because they have a vested interest in maintaining the capitalist system (the issue then becomes even more clouded, because this could of course be true).
The predictions of what Popper considered genuine science, on the other hand, are precise, and can be tested through experimentation or observation, which makes them eminently falsifiable. Ultimately, there are no maybes involved; they are either right or wrong. For these reasons, Popper felt that it was ‘falsificationism' (as it became known), rather than confirmation, that lay at the heart of the scientific method."
Why does Popper hold that the more falsifiable a theory is, the ‘better' it is? Do you think this is a good idea? Why or why not?
"Popper argues that the problem of induction does not impact negatively on scientific knowledge because scientific knowledge does not depend on induction for its justification. Indeed, he argues that no matter how many confirming instances occur of some generalisation we have derived from observational data, there will always be the possibility that the next instance that occurs will falsify it. Ladyman cites the ‘all swans are white' theory as a good example of just how this may occur.
Popper holds that science is, in fact, not about confirming our theories, but in trying to prove them wrong. Once a falsifying instance has occurred, induction becomes redundant, since it can then be inferred that the theory is false purely as a matter of deduction; no further evidence is required. Once the Titanic has sunk (possibly not the best example, but I think it kind of illustrates the point), there can be no more debate over whether or not she is "unsinkable".
Popper maintained that any theory that could not be falsified through experience must be unscientific. Any scientific theory worthy of the name must be falsifiable, in the sense that observations could be made that would falsify it; thus, if we found a metal that did not expand when heated, it would quite obviously falsify the theory that all metals expand when heated. He argued that science proceeds, not by testing a theory in order to accumulate observations that support it, but by attempting to disprove it.
Popper asserts that any theory that makes precise predictions about a range of phenomena must be highly falsifiable, because it only takes a single case that does not conform to the expected outcome to disprove the theory. Thus, the theory that all metals expand when heated is far more falsifiable than the theory that copper expands when heated, simply because there is a far greater range of observations that could falsify it. Popper basically turns things around, and says that what we think of as confirmation should really be thought of as unsuccessful falsification.
Popper saw good scientific theories as ones that can make very precise predictions about a broad range of phenomena. By definition, this makes them more falsifiable; there is simply more scope for testing such a theory, and therefore more opportunities to prove it wrong. I would certainly agree with Popper to the extent that some of the greatest advances in science have come about as the result of bold conjecture.
Certainly, mistakes have been made and we have often had to revise some of our most cherished scientific theories, but this surely implies that we have found some new and (presumably) better theory. I like Popper's proposition that all knowledge is provisional and subject to revision (what he calls fallibilism). Discredited theories are, and should be, abandoned in favour of new theories that will often be more precisely stated (and thus more falsifiable). Popper likens the process to a kind of scientific version of natural selection - one in which only the fittest theories survive."
Explain the difference between the ‘context of discovery' and the ‘context of justification'. Why is it useful to demarcate these two?
"We have seen that naïve Inductivism suggests both a scientific method for testing theories and a method of generating those theories in the first place. Indeed, the thinking, for a long time in the history of science, was that the laws of science should be entirely derived from experimental data. This idea was often supported by the testimony of scientists.
Newton himself claimed to have derived his laws of mechanics purely from observational data, and that he did not indulge in speculation. This seems somewhat unlikely, however, particularly in view of the fact that he is known to have had a lifelong interest in alchemy and an obsessive interest in certain parts of the Bible found in the Old Testament. There is also the small matter of Newton's contributions to the advancement of mathematics and the physical sciences. These came from the mind of a genius; they were not merely the result of a mechanical process of data analysis.
Popper certainly didn't think radical new ideas could be generated from observational data alone. In fact, he defined two contexts within which the history of science in general, and the development of particular theories in particular, should be examined. One of these was the context of discovery. He believed that hypotheses could have any number of causal origins. The inspiration for a particular hypothesis, for example, could stem from a dream or a vision, or from particular religious teachings.
Just to illustrate this point, between the ninth and thirteenth centuries AD, Islamic scholars made significant contributions in the fields of mathematics and astronomy. The development of algebra as a branch of mathematics in its own right, for example, is due to the work of the ninth century Persian mathematician Muhammad ibn Mūsā al-Khwārizmī. Islamic interest in astronomy, however, was at least in part driven by the rapid spread of the Islamic empire, and the consequent need to be able to calculate the direction in which Mecca lay for the purposes of prayer.
The credibility of a scientific theory, according to Popper, was to be regarded as being independent of the source of its inspiration, or indeed the identity of the individuals from whom it stemmed. In this sense, the context of justification was the other side of the coin. Whereas science may often have found common ground with the arts in terms of its inspiration, it differed from the arts in the sense that its theories would necessarily be subjected to empirical testing.
It was this aspect of science with which, according to Popper, the philosophy of science should concern itself, rather than any concerns over how scientific theories were arrived at in the first place. There is certainly a case for arguing that any evaluation of the validity of a hypothesis should not be influenced by the origins of that hypothesis. The question of what precisely inspired a hypothesis, or by whom it was conceived, should be kept entirely separate from the process of confirming the validity of the hypothesis in order to ensure the objectivity of the conclusions reached."
Why does Duhem think that ‘an experiment in physics can never condemn an isolated hypothesis but only a whole theoretical group'?
"Duhem's reasoning is that a hypothesis cannot exist in isolation. It exists as part of a system of hypotheses in which there is often a high degree of interdependence between individual hypotheses. To falsify one hypothesis therefore calls into question the validity of the whole system of hypotheses, or at the very least raises questions about the experimental or observational conditions under which the hypothesis was tested, and generates the need to examine all of the relevant background details.
For example, the statement that all metals expand on heating implies the statement that a particular sample of copper will expand on heating. If the latter should prove to be false, it would seem to invalidate the former, i.e. it would invalidate the theory that all metals expand on heating - a deductively valid argument, but is that all there is to it?
Duhem contends that it is simply not possible to deduce what will be observed from a single hypothesis, because hypotheses must be evaluated in the context of assumptions made about things like background conditions, the reliability of any measurements taken, and so on.
Ladyman cites the example of testing Newton's gravitational theory by observing the path of a comet. The laws of gravitation are only one factor that governs the path the comet will take. We need to know, for example, the mass of the comet and the various planetary bodies involved, together with their relative positions and trajectories. We need to establish a gravitational constant, and we need to take into account Newton's other laws of motion. Even the instrument we use to observe the motion of the comet – a telescope - contributes to this complexity, since we must rely on it working in the way we think it should.
If the observational data in such a case do not support the hypothesis, where does the problem lie exactly? Maybe we miscalculated the planetary masses, or the velocity of the planetary bodies. Maybe one of the supporting hypotheses was, in fact, incorrect. Maybe a mistake occurred with the observations. Maybe the observational data was OK, but somebody got the maths wrong.
There are, potentially, numerous factors that can skew the expected outcome of an experiment, so the principle of falsification of a hypothesis based on observation alone is not as straightforward as it might at first seem. Duhem realised this, but felt that many people, even scientists, failed to take it into account. He asserts that 'science is a system that must be taken as a whole . . . one part cannot be made to function except when the parts that are most remote from it are called into play . . .'.
The implication seems to be that, although observation statements may appear to refute what was hitherto an accepted and successful theory, we should think very carefully before abandoning that theory altogether. This is borne out by (relatively) recent scientific developments. For example, when scientists discovered that the atom, which they had previously thought of as indivisible, was in fact composed of a number of different kinds of particle, they did not abandon the concept of the atom altogether, but instead revised their theories concerning its precise nature."
How is the ‘sophisticated inductivism' Ladyman describes at the end of his chapter different from Naïve Inductivism? And how is it different from Popper's falsificationism?
" "Ladyman starts the chapter by reiterating the need for a theory of scientific method – particularly the need to determine whether scientific knowledge is justified, to establish its limits, and to decide whether a theory or discipline really is scientific or not. He acknowledges that naïve inductivism fails to deliver in this respect; he infers that the problem of induction undermines our capacity to justify our scientific knowledge, and states the need for a precise theory of confirmation and a more plausible account of how scientific theories are actually developed. He does not reject the principles of inductivism outright, however.
Ladyman appears to agree with much of what Popper says, particularly with respect to the idea that ‘ad hocness and novel prediction' must play a large part in providing the inspiration for new scientific hypotheses. This, he says, together with corroboration, ‘must surely play a part in explaining the difference between right and wrong reasoning in science'. He maintains, however, that while many scientists appear to concur with Popper that theories must be falsifiable, and that actively trying to falsify them is an important and productive process, we nevertheless have to fall back on some form of inductivism in order to adequately explain the scientific method and to justify scientific knowledge.
Ladyman states what many believe to be the case, which is that science must embrace both confirmation and falsification. Popper's ideas, he says, are valid, but not exclusively so. He feels that they should be used together with the principles of naïve Inductivism to generate a more sophisticated form of Inductivism. The ‘sophisticated inductivist', according to Ladyman, uses the distinction between the context of justification and that of discovery to keep the question of how scientific theories are developed separate from the question of how they should be tested.
Unlike naïve inductivism, therefore, sophisticated inductivism does not insist that a theory cannot be proposed before there is empirical data to support it. It does not even require ‘probable cause' to exist before the formulation of a hypothesis can occur. Describing a model he calls ‘hypothetico-deductivism', Ladyman essentially says that theories can be produced by any means necessary; their subsequent confirmation (or falsification) is purely decided on the evidence found to support them (or refute them)."
Notes: