Imagine being a scientist developing a new drug. You test it on over 1,000,000 patients of widespread backgrounds, and none experiences any side effects. You put it on the market and, once again, receive no notification of undesired outcomes. Imagine then prescribing it to one of your patients whose profile and medical history give you no reason to suppose him to be different from all other patients who have taken the drug so far. Could you assert with certainty that John will not experience any side effects?
Hume, Inductive Reasoning, and Causality
The above dilemma is known as the Problem of Induction. The first to uncover it was David Hume, an 18th-century Scottish philosopher, who was famously sceptical about the existence of causal relations, i.e., relations between an event and its effect. His disbelief in the notion of causation led him to be wary about inductive reasoning and to argue that the answer to the question in the paragraph above is that we have no grounds to assert that John will not experience any side effects following the consumption of the drug. Let’s step back and explore how Hume came to this conclusion.
A causal relation is a relationship between event A and event B, where event A precedes event B, and A represents the reason for which B is happening. Such a relationship is commonly described as A being the cause of B and B being the effect of A.
Let’s look at a practical example. Imagine having a pot of water and putting it on a lit cooker. After a few minutes, the water starts boiling and then evaporates. Knowledge of chemistry and physics aside, imagine repeating the experiment every day for the rest of the week, the month, and the entire year, obtaining each time an identical outcome. This might lead you to believe in the existence of a relation between heat and water such that heat causes water to turn into water vapour, through a process that can be called evaporation. Hence, heat is the cause of water evaporation and water evaporation is the effect of heat. This relationship between heat and water evaporation is defined as a causal relation.
Necessary Connections and Causal Relations
Once the causal relation between two events – such as heat and water evaporation – has been established, one would likely expect it to keep holding in the case of future instances of event A followed by event B. However, Hume questions whether there is any reason that might justify one’s belief that causal relations, once established, ought to keep holding for future instances of the same sequence of events, or, on the other hand, what looks like an established causal relation is nothing but a coincidence that is by no means bound to repeat itself forever.
Hume explores the problem of causal relations and the possibility of the existence of necessary connections between causes and effects in his An Enquiry Concerning Human Understanding. Within the text, he claims that “this idea of a necessary connection among events arises from several similar instances which occur of the constant conjunction of these events; nor can that idea ever be suggested by any one of these instances, surveyed in all possible lights and positions.”
Hence, Hume argues that the possibility of the existence of a necessary connection between events causally connected may easily arise from the observation of several instances of the events in question co-occurring, as if they were necessarily connected by a relation called causation. However, there is often nothing within the nature of the events that confirms, or even suggests the existence of such a relation. As for our example, there is nothing in the nature of heat that necessarily connects it with water evaporation, and there is nothing that entails that the two are in any sort of causal relationship with one another.
Can Causation be Disproven?
The preceding paragraph has provided an overview of Hume’s argument that one is not justified in claiming the existence of a connection between any two events, unless something within their actual nature, and not a series of independent observations attributable to mere coincidences, suggests it. Observing the co-occurrence of two or more events several times does not increase the probability of their future co-occurrence and adds absolutely nothing to the potential establishment of a relation of any kind between the events.
Looking at past experiences, however, it seems natural to derive that certain connections between events have always held across time, and there is no clear reason one ought not to rely on such connections given that they have never been disproven.
Imagine holding a tennis ball in your hand, and then releasing it. The tennis ball is going to fall on the nearest surface where it can rest – being that the floor, the grass, the table, or any other. This is due to a force known as gravity, and – on planet Earth – it has never once been seen a tennis ball being held and then released, that instead of falling on the ground would float up, left or right. Hence, for what reason would we have to doubt that the next tennis ball, once released, would fall on the nearest surface on which it can rest?
Does the Future Resemble the Past?
The reason to doubt the trajectory of the tennis ball in the example above – according to Hume – is that the assumption that the ball is going to fall downwards until it meets a surface where it can rest relies on another claim, which, similarly, cannot be supported. The claim in question is that the future has – as far as we know – always resembled the past, and for this reason, we ought not to doubt it always will.
In other words, the laws of Physics that we currently use to describe the phenomena observable in the world we live in have always held as they currently hold. Hence, we have no reason to assume that those laws could ever stop holding, resulting in a future which does not resemble the past. By arguing for the existence of a necessary connection of resemblance between the future and the past and claiming that such a connection between the two always held in the past and therefore ought to always hold in the future, one could also argue for the existence of causal relations between other events.
In fact, given the relation of resemblance between the future and the past, any other relation of causation could be expected to keep holding across time, on the basis that it has always held.
However, this would be nothing but a circular – failing – argument, as the fact that the resemblance between the future and past will always hold in the future, cannot technically be proven – if not through inductive reasoning, on the basis that it has always held so far, leading back to the original quest.
Giving Up: The Skeptical Solution
Induction is arguably the most efficient tool that science and philosophy alike can avail of for the development of new knowledge. In fact, at the heart of most groundbreaking scientific discoveries, there is the development of hypotheses that generalize over several individual and independent observations of phenomena, which consists of the process at the heart of inductive reasoning. It is inductive reasoning, in fact, that allows one to expand the realm of knowledge to any thinkable hypothesis rather than being limited to what is logically deductible from one’s current knowledge.
Additionally, it must be mentioned that – even if Hume were to be right in saying that there is nothing in the nature of two events, which happen independently from one another, that would suggest for them to be necessarily connected by a relation such as causation – relying on the existence of causal relations has so far always worked. Therefore, while this is not meant to suggest discarding the problem, one may be justified in deciding to live in reliance with causal relations despite the awareness that their nature remains philosophically uncertain.
It is on this line of reasoning that Hume himself accepts what is known as the sceptical solution—despite the absence of philosophical grounds for the acceptance of beliefs derived from induction, humans are naturally brought to rely on inductive reasoning and even its lack of philosophical grounds would not lead one to instinctively refrain from using it.
Hence, while there would be neither need nor possibility to conduct a life that refrains from relying on causal relations, one ought to continue investigating the problem of induction, with the goal of explaining what determines the – everlasting yet apparent – necessary connections between causes and effects.
