Astronomy Cast Ep. 390: Occam’s Razor and the Problem with Probabilities

I’m not saying it’s aliens, but it’s aliens. Actually, it’s almost certainly not aliens, or a wormhole, or a multiverse. When scientists discover something unusual, they make guesses about what’s happening. But Occam’s Razor encourages us to consider the probabilities of different events before making any concrete predictions.


Visit the Astronomy Cast Page to subscribe to the audio podcast!

We record Astronomy Cast as a live Google+ Hangout on Air every Monday at 12:00 pm Pacific / 3:00 pm Eastern. You can watch here on Universe Today or from the Astronomy Cast Google+ page.

8 Replies to “Astronomy Cast Ep. 390: Occam’s Razor and the Problem with Probabilities”

  1. The episode touches on some important points, but those points could do with elaboration.

    People often misunderstand Occam’s Razor as favouring the simplest hypothesis, but as the episode says, it is better understood as favouring the hypothesis that minimises assumptions.

    You can see the difference most dramatically in domains involving human behaviour — psychology, linguistics, politics, etc. If you apply Occam’s Razor naiively, you might be seduced by the idea that all of human behaviour in a given domain can be accounted for by some particular cognitive algorithm. After all, that’s simple. But that “simple” hypothesis requires huge assumptions — firstly, that there exist aspects of human behaviour which can be fully accounted for by a single mechanism (even though there is no precedent for that), and secondly, that you’re looking at such a case. Whereas the hypothesis that human behaviour in a given domain is complex and multifaceted requires much more modest assumptions. We know human behaviour is complex and multifaceted; we don’t have to assume it. So in such a case, Occam’s Razor favours the nominally more complicated explanation.

    Also touched on is that Occam’s Razor is only an objective metric when comparing two very similar hypotheses, because when unrelated hypotheses are compared there is no objective way to say that One Big Assumption in Hypothesis A outweights Two Smaller Assumptions in Hypothesis B. (The nearest you can get is to apply the Bayesian principle, but that’s only applicable if you have the data.) In other words, Occam’s Razor is almost always subjective in practise. But that’s a good thing, because if different scientists have different opinions on which hypothesis is worth pursuing, different scientists will pursue different hypotheses, to the benefit of us all.

    The third point (not addressed in the episode until the questions) is the rebuttal to those who argue (usually facetiously) that the explanation “God did it” only makes one assumption, and is therefore favoured by Occam’s Razor. The rebuttal, of course, is “Only if you don’t ask any more questions”.

  2. Since Adrian covered all the juicy bits, I’m left only with:

    1) Aren’t our Dads’ workshops/basements/garages just the absolute best places? So cool.

    2) Love your hair, Ms. Pamela! *^_^*

    2a) Please don’t hurt me, Mr. Pamela!!!

      1. Um… sorry?

        As Dr. Gay is married, some might have found a comment on her pleasing appearance from another man somewhat improper in its own right. With that in mind, I was attempting a bit of a jest to defuse that a bit, a la, Dr. Who’s habit of referring to Rory as “Mr. Pond.”

        Apparently, it fell a bit flat. =/ Apologies all around.

      2. I don’t want this issue to dominate the comment thread, which is why I kept my response short and discrete, but the point is that when directing comments at a woman, it is offensive to imply that your main concern is whether or not a man approves of them.

        Now on with the discussion of philosophy, astronomy and workshops. (Also, for me, sleep. See you in the morning.)

      3. I got it and appreciated it as I use that joke often, myself. Same could go the other way complementing a male’s looks. IMO, your taste remains fine. And that’s the point… IT’S a JOKE! 😉

Comments are closed.