Compare inductivism with falsificationism - Which is a more plausible account of how scientific knowledge is acquired? — UCL Philosophy of Science (2016)

David Hume (right) and I (left)

Note: Written in my first year at University College London (UCL)

The focus of this essay is to compare inductivism with falsificationism in terms of which is a better account of how scientific knowledge is acquired. I shall first state the problem of the essay, and outline the main alternative theories that attempt to deal with it. I shall consider inductivism in some detail; which I maintain is more plausible, then I will consider falsificationism. The arguments for why inductivism is a plausible account of how scientific knowledge is acquired can be explained by the principle of uniformity and the virtuous circle.

The key arguments for why falsificationism is not a plausible account of how scientific knowledge is acquired include: the Duhem thesis, the fact that scientists often ignore falsificationism and Ladyman’s query of future expectations. It is important to note that the question appears to assume that inductivism can be only compared in terms of falsificationism as a way to demonstrate its plausibility, and whereas this is true in part, there are a number of arguments in support of inductivism that do not inherently stem from the weaknesses with falsificationism. Similarly, there a number of criticisms that can be made towards falsificationism that do not inherently stem from the strengths of inductivism, but are nonetheless essential to our discussion of which of the two we should accept as a more plausible account of how scientific knowledge is acquired.

Before I begin the substantive body of my essay, I would like to define some key terms: inductivism (or the principle of induction) is defined as: (1) ‘the process of inferring a general law or principle from the observation of particular instances’; (2) ‘the greater the number of observed associations between two things (events, facts, phenomena) A and B, the greater is the probability that they will be associated in a fresh case’; (3) ‘The probability of the fresh case association approaches certainty as the number of observed cases approaches infinity’ (Russell, 1959, pg. 37) . The problem of induction can be defined as: ‘the problem of finding a way to prove that certain empirical generalizations which are derived from past experience will hold good also in the future’ (Ayer, 2012, pg. 61). Finally, a statement is ‘falsifiable’ if ‘it is possible to conceive of an observation or an argument which negates the statement in question.’ (Parsons, 2016).

The principle of uniformity suggests that inductivism is a plausible account of how scientific knowledge is acquired. This is a plausible view, for the following reasons.

Firstly, the principle of uniformity states that nature has always exhibited uniformity and will, therefore, continue to do so, which means that our ability to make associations between particular events, facts or phenomena can increase in probability as we observe more cases. For example, we tend to believe in the constant conjunctions of morning and sunrise or that even though some scientific theories and hypotheses offer only provisional answers and undergo modifications to account for exceptions or non-confirming instances, science habitually assumes that hypotheses which have exceptions can be replaced by hypotheses which have no exceptions (Russell, 1959, pg. 33), and as such, inductivism remains a viable way of converging on more accurate projections of future occurrences. Furthermore, Russell asserts that if, for example, the earth collided with a large celestial body that upset its rotation, one might choose to abandon their belief that the sun will rise tomorrow in spite of the fact that the laws of motion and the law of gravitation which govern its rotation would not be ‘infringed’ (Russell, 1959, pg. 34). In other words, even though scientific hypotheses which aim to give an accurate account of natural phenomena are relative to the observational data and a scientist’s interpretation of their findings, the natural laws that govern these phenomena for which these hypotheses give account are not, but instead remain fixed and are thus subject to inductive predictions derived from past experiences and background knowledge.

However, there are some reasons to be doubtful of whether the principle of uniformity justifies inductivism as a plausible account of how scientific knowledge is acquired. One of these reasons is that the principle of uniformity relies on the very assumption in question; namely that natural laws like the laws of motion and gravitation are uniform and will always be so. The rules of logic do not prevent the laws of nature from radically altering, and if such were to occur, we would be left with little to no frame of reference to draw up new theories or make adjustment to existing ones, which appears to supports the notion that inductivism is not a viable method of scientific inquiry. However, I would argue that assuming that nature is uniform has always worked in practice, so there is no clear reason to not believe it will continue to work in future. Furthermore, we seem to have no alternative

way of verifying what a ‘future futures’ (ibid.) would look like, that is, a future that does not resemble the past with which we are used to making associations, so this response fails to adequately refute inductivism in terms of its predictive power.

Yet, one might respond to my argument by asserting that my rejoinder is a meta-induction that is not justified. However, this counterclaim fails, because, like Hume, I would argue that inductivism is justified by the that fact human beings are hardwired to suppose that the future will resemble the past and that such is a habit of the mind of which there is no escape (Hume, 2012, pg. 10). Consequently, the old problem of induction is an insoluble and, therefore, a ‘fictitious’ problem for, as Ayer explains, ‘natural science is not impaired by the fact that some philosophers are puzzled by it’ (Ayer, 2012, pg. 63).

We have just seen how inductivism is justified by the principle of uniformity, and that according to Hume and Ayer, the principle of induction is habitual and the old problem of induction is a pseudo-problem. I will now present another two arguments in support of inductivism.

My second argument is that the practice of inductivism is justified by the very rules which govern it which are themselves justified by the practice. This is known as the virtuous circle (Goodman, 1983, pg. 9). For example, a scientist might observe a tree and make inductive statements regarding its biology, biochemistry, chemistry and physics and will be unwilling to infer, for example, that a tree which is shedding its leaves is a coniferous tree since this would violate the definition and usage of the latter term. Instead, the scientist will rely on their definitions and background assumptions to infer that the tree is deciduous, which is supported by their observation and conforms to the usage of the term. Like Goodman, I would argue that much of scientific inquiry is characterised by the ‘dual adjustment between definition and usage’ whereby the usage of a term like coniferous tree, that is, trees that do not shed their leaves and have fine needles, ‘informs the definition, which in turn guides the extension of the usage’ (ibid, pg. 64).

One might respond to my argument by saying that this so-called virtuous circle is a prime example of circular reasoning and that merely describing the practice of induction does not in turn justify induction as a means of gaining certain knowledge of the future. However, this response fails, since the old riddle of induction which assumes the integrity of the virtuous circle is less concerned about claiming whether a certain prediction will turn out correct since this simply is not possible in a logical sense (ibid, pg. 62). Instead, philosophers are more concerned about being able to find a precise method of ‘distinguishing antecedently between true and false predictions’ (ibid, pg. 62), that is, choosing between one theory over another, using the canons of induction. This is known as the new problem of induction. Also, I would argue that there is not a distinction between the justification of inductivism and the practice thereof because the two mutually support one another without the need of further justification from elsewhere.

We have just seen how inductivism is justified by the virtuous circle and by the fact that the old problem of induction is more concerned about theory-choice. I will now present three arguments that falsificationism is not a plausible account of how scientific knowledge is acquired.

My first argument is that falsificationism attributes the failure of a hypothesis to yield confirming results exclusively to the hypothesis when the problem could very well be with the observational evidence or equipment. The Duhem thesis states that ‘an experiment in physics can never condemn an isolated hypothesis but only a whole theoretical group.’ (Duhem, 1906), which means that when there is a conflict between a scientific hypothesis and an observation made a scientist, the scientist can choose either to reject the hypothesis or they can choose to reject their observation since the fault can be located in either of the two. As James Ladyman explained, if, for example, a comet fails to adhere to the path Newtonian mechanics predicts, we can assign the problem to the law of gravitation, one of Newton’s other laws or to the values of the mass of the other bodies (Ladyman, pp.77–78). What this example suggests is that falsification in practice is no more conclusive than inductivism when it comes to accepting a theory as we are still stuck between choosing to reject the hypothesis or falsifying evidence.

My second argument is that there are many scientists in the history of science who chose not to abandon their theories in spite of observational data that appeared to contradict them. For example, Isaac Newton’s theory of gravity implied that, given the attractive forces between stars and other celestial objects, the universe should implode. Even though Newton saw this as a problem, he concluded in an ad hoc fashion that God was responsible for providing the gravitational counterbalance (De Lagemaat, 2011, pg. 238). Another example that illustrates my point is that of Dimitri Mendeleyev’s process of designing the periodic table, which he did by arranging the elements in accordance with their atomic weights, and not all elements fit the model. Instead of abandoning his theory, Mendeleyev concluded that the ‘anomalous weights’ were a result of ‘experimental error’ (ibid). Both examples suggest that there is something inherently impractical about debunking an entire theory as a result of one or more non-confirming instance, particularly if it is a promising theory like Newtonian mechanics which continues to be of use today in the construction of bridges and calculating the movement of planets (ibid, pg. 181) even if it was replaced by Einstein’s theory of relativity. The fact is auxiliary hypotheses can rescue a falsified theory, which in the case Newton’s theory of gravity, the universe does not collapse in on itself due to ‘the speed at which the stars are moving away from each other which counteracts gravity’, and the anomalous weights of some of Mendeleyev’s elements were due to the ‘presence of various isotopes’ (ibid).

My third argument is that traces of inductivism can be found in falsificationism, which threatens the basic integrity of the latter. In Conjectures and Refutations, Popper asserts that it is rational to trust in theories which are ‘well-corroborated’, ‘bold’, that is, in contrast with background knowledge and dogmatic tradition; ‘novel’, that is, yielding unexpected predictions (Popper, 1973, pg. 203). The problem here is that although Popper does not explicitly say so, it seems to me that he is making a reference to (future) expectations which is based on background knowledge, and as a falsificationist who advocates the notion that one does not need to rely on background assumptions since this is inductive in nature and that that there is no ‘mechanical way’ of coming up with testable hypotheses (ibid, pg. 37), this seems like a flagrant contradiction.

Popper might respond to my arguments in several ways. For instance, Popper might object by claiming that falsificationism’s emphasis on a theory being bold and novel allows us to demarcate true science, such as Einstein’s theory of relativity, from pseudo-science, such as Marxism and psychoanalysis (Thornton, 1997), and that theories which put themselves at risk whilst making novel predictions should be treated as genuine science. However, this objection does not succeed since Einstein’s theory of relativity intended to be more comprehensive than Newton’s gravitational theory, which was ‘a first approximation by specializing general relativity theory’s equations’ (Rivadulla, 2004). What this suggests is that if Einstein had not relied upon Newton’s preceding approximation, albeit a limiting case, when developing his theory of relativity, which a falsificationist would forbid by virtue of it being part of the inductive method, it is possible that Einstein would not have accomplished his theory of relativity. So we have seen that none of Popper’s replies to my argument that an adherence to background knowledge, which is typical of the inductive method, succeed. Hence, we should reject Popper’s claim that there is no logic to discovery and that background assumptions are a hindrance to formulating bold and novel theories.

In conclusion, I have argued that falsificationism does not succeed as an account of scientific inquiry and that inductivism is more plausible. I have established that scientists readily accept nature as being uniform and construct their theories using background knowledge in order to guide their definition and usage of scientific terms, also known as the virtuous circle. In addition, I highlighted that contrary to Popper’s account of falsificationism, inductivism succeeds by positing that there is indeed a logic to discovery since scientists naturally ground their formulation of scientific theories in past experiences and preceding knowledge, and although there might not be any way of proving its validity in the mathematical sense, the fact that it has always worked in practice would lead one to believe that it is our only viable option. Even though the new problem of induction suggests that we cannot conclusively verify a theory, we cannot conclusively falsify a theory either, and Popper does not adequately address whether we ought to choose to reject an observation rather than an entire theory.

Word count: 2,365

References

1. Ayer, A. J. (2012). Language, Truth and Logic. : Dover Publications. Retrieved from http://www.ebrary.com
2. Chalmers, A. (2007). What is this thing called science?. 1st ed. Maidenhead, Berkshire: Open University Press, Chapters 5–7.
3. De Lagemaat, R. (2011). Theory of Knowledge for the IB Diploma. 1st ed. Cambridge: Cambridge University Press, “The Natural Sciences” (Chapter 8).
4. Duhem, P. (1906). 1954. The aim and structure of physical theory. Trans. Philip P. Wiener. Princeton: Princeton University Press.
5. Goodman, N. (1983). Fact, fiction, and forecast. 1st ed. Cambridge, Mass.: Harvard University Press, “The New Riddle of Induction” (Chapter 3).
6. Hume, D. (2012). A treatise of human nature. Courier Corporation.
7. Ladyman, J. (2002). Understanding philosophy of science. Psychology Press.
8. Lakatos, I., Worrall, J., & Currie, G. (1980). The methodology of scientific research programmes: Volume 1: Philosophical papers (Vol. 1). Cambridge university press.
9. Parsons, K. (2016). Popper on Falsifiability. The Secular Outpost. Retrieved 4 January 2017, from http://www.patheos.com/blogs/secularoutpost/2016/03/03/popper-on-falsifiability/
10. Popper, K. (1963). Conjectures and refutations. 1st ed. London: Routledge & K. Paul, “Science, Conjectures and Refutations” (Sections 1, 2, 8).
11. Rivadulla, A. (2004). The Newtonian Limit of Relativity Theory and the Rationality of Theory Change. 1st ed. [ebook] Available at: https://www.ucm.es/data/cont/docs/481-2013-10-14-TheNewtonianLimit.pdf [Accessed 4 Jan. 2017].
12. Russell, B. (1959). The problems of philosophy (1st ed., p. “On Induction” (Chapter 6). New York: Oxford University Press.
13. Thornton, S. (1997). Karl Popper. Plato.stanford.edu. Retrieved 4 January 2017, from https://plato.stanford.edu/entries/popper/