Human beings learn or develop concepts through a process of induction; this is the process of generalizing from specific cases. Researchers differentiate two theories of induction: continuous and discontinuous.
Continuous induction is a mental process of applying increasing weights to recurring combinations of stimulus-response experiences. These weighting factors determine, in the subject’s mind, how the world works, and this is the essence of conceptualization. Some theorists disagree. “Concept learning is not a process of gradually strengthening associations at all, but rather a noncontinuous or discontinuous process of constructing and testing hypotheses until one works” (Mayer, 1992). Some research seems to indicate that this dispute is a false dichotomy; it seems that people use one or the other approach depending on the circumstances.
Despite the specific approach, it seems that people make inductions, not from their actual pool of experience, but instead from those experiences which are readily available to them in their memories. This is known as the availability heuristic and, to the extent that it happens, can be seen to be a major factor in how people conceptualize the world. Thus, someone who is making an induction will make it based on the experiences he can readily access – which means that he will make the induction based on the neural connections he happens to be making at the time. In a different context, he would induce different concepts.
Another result of research into induction is that most people test hypotheses by entertaining positive test cases. But, as Karl Popper has pointed out, no number of positive test cases allows a definitive judgment on the truth of a theory. However, this doesn’t stop people from doing so anyway. These inductions that we make may not always be correct or even reasonable, but we do them anyway and all the time.
It is fairly easy to trace how this proclivity toward induction was developed in Homo Sapiens. If, for example, a proto-human somewhere tried a fruit from a particular tree and died soon after, a reasonable (but not necessarily correct) conclusion might be that the fruit was poisonous. Given that there were other trees and fruit around, there was survival value in not going back to that tree to eat. From a logician’s point of view, this decision is called inducing from a single instance and is labeled invalid. And yet, perhaps, if these logicians had existed 2 million years ago, they might have all died from poisoning by now.
Lest one believe that the human race is now above this sort of error in thinking, consider the implied induction in the sentence: “That car dealership sold me a lemon – I’m never going back there again!” Thus there is a strong distinction between what logicians have come to label valid thinking and the thinking that we actually do.
How many test cases should it take to demonstrate the existence of God?
Postscript: I’ve been asked: How many test cases should it take to show God doesn’t exist? The very point of this essay is that an induction does not have to be mathematically valid to be useful. My question is that if you won’t go to a restaurant where you once got sick (a single data point), why won’t you accept evidence that God exists just by looking around? And, if you don’t happen to experience God, remember the adage absence of evidence is not evidence of absence.