In this essay I will attempt to resolve the difficulty that Hume’s problem of induction (POI henceforth) poses to everyday activity by looking to Karl Popper’s theory of falsification. I will first outline Hume’s POI, showing how it affects all of us and not just the philosophers. I will then set up a framework within which I will work. I will then briefly outline Popper’s theory of falsification for the purposes of daily application. I will consider objections to my proposal and respond to them. I will conclude that to use Popper’s theory would require assuming Hume’s POI which we are initially trying to resolve. However, as inductive inferences are important for everyday activity we should adopt a “reflective equilibrium”.
Hume thinks there are two objects of human enquiry: relation of ideas and matters of fact – Hume’s fork (Millican, 2008, xxxv). Relation of ideas are intuitively or deductively certain; they can be known a priori (discoverable without experience). ‘2+2=4′ is gotten by intuition for it is by our direct intellectual “ grasp” (Millican, 2008, xxxvi). To know something a priori does not mean that these truths are ‘innate’ but rather that once we have discovered such a truth we do not need experience to justify it (Millican). The truth that all bachelors are unmarried is known deductively, not in the strict sense of being limited to a formal logical system (Millican, 2008, xxxviii), as is the case for Euclid’s geometrical system. Further, the denial of a relation of idea would lead to a contradiction, such as ‘2+2=5′. Matters of fact describe the world through experience; they are known a posteriori. Their denials are always possible because they are conceivable. For example, drinking water quenches my thirst is a matter of fact for it is entirely conceivable that drinking water doesn’t clench my thirst because I may be allergic to water.
To learn from our experiences, we must infer from past events to future events. For example, drinking water has quenched my thirst in the past and so I think it will do so again. Hume thinks that such beliefs are based on the idea of causation; it is from the experiencing of similar causes being followed by similar effects. Hume thinks that we assume a principle of uniformity: the future will resemble the past (2008, 27). Hume thinks that we cannot justify such inferences from a priori reasoning alone for we can conceive of a cause as having a different effect (2008, 24); for example the water does not clench my thirst. However, we also cannot rely on matters of fact to provide justification for inductive inferences for we would be assuming what exactly we are trying to justify, namely: that we cannot make any inferences from past events without thinking that the future will conform to the past (2008, 27). Thus, by accepting Hume’s fork, we are not justified in making inductive inferences.
Hume finds induction problematic not merely because we cannot justify it, but rather that our everyday practices and beliefs are questioned by the rejection of the validity of induction. Induction takes us above and beyond our evidence of the senses and memory (Hume, 2008, 31). I am thus no longer justified in believing that feeding my family will keep them from starvation; whether I continue to do so is another question. Though Hume does not think that our everyday activities are affected by his problem (2008, 32), it concerns both philosophers and common people alike. The philosopher, like the common people, needs to survive, putting the former in a similar position of the latter.
I will now briefly outline Karl Popper’s theory of falsification. Although Popper is concerned with the scientific method, he thinks that it can be, and is, applied to everyday activities (Thornton, 2016). Popper does not believe that one can go from single observations to universal laws (2005, 33). For example, because ‘every table I have seen in the past had four legs’ I infer that ‘every table must have four legs’. We cannot logically verify universal statements, for the next table that I come across may conceivably have only three legs. However, what we can do is falsify statement of the kind ‘all tables have four legs’, deductively through experience (Popper, 2005, 48).
Let us suppose that that I am concerned with how I can lead a healthier lifestyle. I come up with a theory of human physiology. As Popper thinks that science is about the progression of theories (Thornton, 2016), whether my theory is ‘naive’ does not matter. I will assume for now that the theory I have proposed is internally consistent, it can be subject to empirical testing and that in its proposal I am advancing the thought surrounding healthy living. Now, I will derive a prediction from my theory: a paradigmatic example might be ‘eating 50 grams of walnuts I will reduce cholesterol by 5%’. This example is not very scientific but will suffice for my purposes. The point of making such a prediction is to see how the theory can stand up to rigorous observational scrutiny.
Through practical application and experimentation we attempt our best to falsify this prediction. If through one observation we find that eating walnuts does not decrease my cholesterol by 5%, then my theory has been falsified and I must go in search for another. One may object that one observation is not sufficient to disprove a theory. For now, I will add that it should be revised and re-examined; to specify up to what point is beyond the scope of this essay. However, if the prediction is confirmed, the theory is not confirmed but rather it has been “corroborated” (Popper, 2005, 39) by past experience. We are never able to say whether a given theory is true by the verification of a prediction (Popper, 2005, 11).
To compare competing theories Popper uses the notion of “verisimilitude" (Popper, 2005, 120) or “truth-likeness” (Thornton, 2016) to distinguish better from worse theories. Suppose we have two theories: T1 and T2. The number of non-falsified predictions derived from T1 must be greater than those in T2 and the number of falsified predictions from T1 must be less than those in T2. Even if the theory with greater verisimilitude is later falsified, we can think of it as the better theory for now because such theories are closer to the ‘truth’ (Thornton, 2016). One might think that in order to know that a theory is closer to the truth than another, we have some idea of the truth. Rather what is being said is that we take such a theory to be closer to the truth than suggest that it actually is.
Let us assume that a ‘super-human’ can indeed live ‘falsifiably’; but what then should we say for children and animals? How are they supposed to adopt such a method? If a child puts her hand over a flame multiple times to find herself burnt, we would find it appropriate that she try to stay away from getting burnt regardless of Hume’s POI. Maybe we just bite-the-bullet and say that falsification isn’t a method applicable for children and animals. However, what basis does the ‘super-human’ have for relying on the verisimilitude of theories, which themselves are supported by past experiences, for future activity? To rely on the verisimilitude of theories would be circular.
One could respond that because we have nothing else to go with, we may as well pick the theory which is most the ‘truth-like’ (Salmon, 1981); that which has withstood rigorous testing. However, such a response would still be based on experience of the past, thus resulting in another circular argument. In light of the objection I have considered above and the points for further research I have mentioned throughout this essay, it would be premature to end on a definite position in regard to the question. Thus, I think it appropriate to adopt a “reflective equilibrium” (Daniels, 2016) which is the end of a thought process whose conclusion we are not sure about and feel the need for further revision of our beliefs. Such further deliberation could help us to decide how we should act in light of Hume’s POI. An equilibrium will be reached once we have coherence amongst our beliefs.
In conclusion, I have attempted to show the applicability of Karl Popper’s theory of falsification as a response to Hume’s POI. The objection, in particular, of the circularity of Popper’s theory makes it difficult to adopt his position fully; however, as inductive inferences are important in our everyday activity I think we should adopt a reflective equilibrium whereby with more time we can decide what we ought to do.
Bibliography
Hume, D. (2008) An Enquiry Concerning Human Understanding, Oxford: Oxford University Press.
Millican, P. (2008) Introduction. In: Hume, D An Enquiry Concerning Human Understanding. Oxford: Oxford University Press. ix-lvii.
Popper, K. (2005) The Logic of Scientific Discovery, London: Taylor & Francis Group.
Thornton, S. (2016) Karl Popper [Online]. The Stanford Encyclopaedia of Philosophy. Available from: http://plato.stanford.edu/archives/win2016/entries/popper [Accessed 25/10/2016]
Daniels, N. (2016) Reflective Equilibrium [Online]. The Stanford Encyclopaedia of Philosophy. Available from: http://plato.stanford.edu/archives/win2016/entries/reflective-equilibrium [Accessed 25/10/2016].
Salmon, W. (1981) Rational Prediction, The British Journal for the Philosophy of Science,. Vol 32(No.2), 115-125.