Trying to Define Skepticism
The publication Skeptic has an editorial that attempts to define skepticism. It is a crisp, interesting article with a brief reference to ancient Greek skepticism and a conclusion with a rephrasing of Descartes’ famous metaphysical equation.
However, the article really caught my attention because, it defines skepticism as a methodology. The editorial states “Skepticism is a method leading to provisional assent”. And that is a fine definition; unfortunately it isn’t true.
Skepticism is not a methodology, at the very least not in the singular sense of the word. People using skepticism do not address uniform problems. The method one would take to assess the statement “Some cows are black and white” look very different from assessing “I remember a past life as a ship captain”, which in turn looks different from trying to determine whether a Chi-square analysis was adequate, or a matched-pairs design was sufficiently well controlled.
It is not simply the complexity, that makes these examples different, it is the nature of the hoops we have to jump through, and the level of certainty we can have after investigation. If anything, it would have to be many smaller methodologies specific to each type of problem.
I would argue that even that, is not enough, you need to have logic is your assessment. I would define logic as:
The valid rationalizations by which we could potentially withhold provisional assent.
And methods as:
The controls one places in a trial in an attempt to see if we can withhold provisional assent.
Now maybe some of you are wondering why my language was worded so strangely. I didn’t say “give provisional assent” or “prove”. That is because in research (or in group based statistical research) we don’t try to prove an effect; arguably we can’t prove an effect. So, instead we try to see if there is no relationship, we call this the null hypothesis. We are attempting to show that there is not a relationship, and to do this we have to make our ideas potentially falsifiable. And that is why; we do not prove the hypothesis, but reject the null hypothesis.
Again, I want to stress why logic and methods must go hand in hand. I will provide examples below:
1) I hypothesize that 1 + 1 equals 2. Logically I know that I should be able to add 2 objects together to test this hypothesis. I choose to do this using water. I add 1 drop, then a second, but I end up with only 1 larger water drop. My logic was impeccable, but my methods, were not well suited to the task.
2) I hypothesize that the ability to sling webs out of one’s hands comes from watching television. I compare the number of reported cases of people with web slinging abilities in New York to the average amount of television viewed by New Yorkers. My logic does not allow me to determine a relationship here, even if my statistical methods were appropriate and rigorous.
3) I opine that a particular Polar Bear is white. I observe the bear in the sunlight and I notice that it is white; I later observe it by accident in the shade and notice that it was distinctly brown. In this case both the methods and the logic need some work.