Why did Socrates say "I know I know nothing?"
Assuming he ever did say it
I was once asked this question, by someone who hadn’t quite grasped what the sage was getting at.
It would have been most unlike Socrates, whose analytical technique (called elenchus) was based upon exposing contradictions in the statements of other people, to express himself in a contradiction. Obviously, anything that I know is something, therefore if I claim that I know I know nothing, I am contradicting myself.
If, however, I start from the assumption that I know nothing, I may ask anyone who claims knowledge to explain to me a particular aspect of what it is that they know.
Usually Socrates would invite his interlocutors to state a generally-applicable principle; then he would offer in reply a specific instance of the principle’s non-applicability. This would lead the interlocutor to redefine the principle more narrowly. The process would be repeated until it could be repeated no further, at which point Socrates would deem that they had arrived at a truth.
The process has considerable merit, but is inevitably limited in scope. Ironically, Donald Rumsfeld’s much-derided, but actually correct, categorisation of knowledge into known knowns, known unknowns and unknown unknowns helps us to understand this.
Most educated people are aware that there are limitations to their knowledge (or known knowns); for such people there exist known unknowns. So, for example, I am aware that there is a field of human knowledge called quantum mechanics, but I have no detailed understanding of what that field encompasses. It is, for me, a known unknown.
Unknown unknowns, his third, category, have been usefully defined as “phenomena which cannot be expected because there has been no prior experience or theoretical basis for expecting the phenomena”. In other words, I can’t know that I don’t know something, if I have no way of knowing that it even exists.
But in later life Rumsfeld came to acknowledge the existence of a further category, possibly the most problematical of all, that of unknown knowns. These he defined as "the things you think you know, that it turns out you did not".
Like Rick in Casablanca, we can be misinformed. Alternatively, having received no notice of change, we may be convinced that something we knew to be true many years ago is still the case today.
Sadly, we may act upon mistaken beliefs with all the intensity that would ideally be reserved for truth. This is often because we simply don’t know enough to realize that we don’t know enough. As they say, a little knowledge is a dangerous thing.
Truths may be much larger in scope than we commonly suppose, and it is even possible that two seemingly contradictory propositions would be seen to be complementary if only we knew enough about how they both fitted into a much greater whole.
Without adequate evidence, it seems, we nevertheless cannot countenance the possibility that we may be wrong. I seem to recall Oliver Cromwell saying to the Rump Parliament “I beseech you… think it possible that you may be mistaken.” This from a man who we know had pretty strong convictions himself.
So, if we return to Socrates, we find that even the little that we think we know may be just waiting for some counter-argument or evidence to turn up.
Unlike him, we should not regard it as the truth by default. We should reserve the term “truth” for proven propositions. That is to say, the conclusions of logical processes that are both valid and sound.