How Do I Know What I Know?

Posted by

When I was first getting into philosophy, my favorite author was Immanuel Kant, and as I started reading his Prolegomena to Any Future Metaphysics, I stumbled upon this passage:

[T]he remembrance of David Hume was the very thing that many years ago first interrupted my dogmatic slumber and gave a completely different direction to my researches in the field of speculative philosophy.

Upon reading this, I wondered what monumental problem could have inspired a philosopher as influential as Immanuel Kant, so I decided to do some reading about David Hume. What I found was a question central to the study of knowledge: How can we determine if the things we believe are actually true?

When humans take in new information, this process is often separated into two categories: deduction and induction. Deduction involves taking concepts we know to be true in some form and applying pure logic to determine some other conclusion. Induction involves observing the physical world and making predictions based on those observations. Most agree that the process of deduction produces valid knowledge, but many question the process of induction. These questions about induction lead to inductive skepticism, the idea that induction is not a valid process of producing knowledge.

David Hume’s influence on Immanuel Kant comes from his contribution to this idea of inductive skepticism. His line of argument concludes that trying to infer knowledge from induction can never be valid. Hume begins with the statement that since many beliefs are based on this kind of inductive inference, we should have some justification for the validity of induction. However, Hume argues, neither deduction nor induction itself can be used to justify knowledge based on induction.

The reason that deduction cannot justify induction is that deduction deals with necessary knowledge (propositions which must be true). In contrast, induction deals with contingent knowledge (propositions that are true in some possible worlds but false in others). One frequently-used example of necessary knowledge is the statement that ‘a bachelor is unmarried.’ Since a bachelor is definitionally an unmarried person, there is no possible world or situation where someone who falls into the category of a ‘bachelor’ is married. An example of contingent knowledge is the statement that ‘the walls of my room are painted blue.’ There is nothing intrinsic about the nature of the walls in my room which demands or necessitates that they should be painted blue; it is simply a contingent occurrence. Deductions can only ever deal with knowledge that is necessarily true because they use only pure logic, so induction (which deals with a lot of contingent knowledge about the physical world) does not work on the same level as the process of deduction. Since these processes deal with very distinct areas, they cannot be used to justify each other.

If instead, we argue that our inductive inferences in the past have proven true (making induction valid), this fails. To begin with, this argument is circular as it uses induction to prove that induction is valid. Furthermore, a person skeptical of inference based on induction would likely deny that our past inductions have produced true conclusions at all.

It is for these reasons that David Hume concludes that induction can never be a valid method for obtaining true knowledge. Nevertheless, we continue to live our lives observing things and making predictions. If everything works now with induction, then why does this kind of skepticism even matter? Well, inductive skepticism creates many problems with the very important concept of cause and effect, and that’s what I’ll be exploring next.