The Unicorn Hypthesis (redux)June 15, 2009 — Deacon Duncan
I hate to leave a loose end dangling, so just for the sake of completeness, let’s take one more look at the Loser’s Compromise and the Unicorn Hypothesis. The point of the original post was to demonstrate that we can’t justify our beliefs on the grounds that they are merely as consistent with the facts as some other hypothesis (or explanation or world view). We did this by setting up a scenario in which the consequences we would expect from one hypothesis (that world affairs are under the clandestine control of self-effacing magical unicorns) work out to be the same as the consequences we would expect from a contrary hypothesis (that humans are in control of their own governments).
I think we did that rather well, but one commenter disagrees.
No, what we’ve achieved is another silly, impertinent scenario, if nothing else, simply because a single sample is seldom sufficient. Here, you offer two hypotheses each with a single consequence that both permit. So, of course I agree with you that “there’s no reasonable basis for concluding, even provisionally, that we’re being secretly controlled by a one-horned oligarchy.”
It’s a classic example of misdirection and dodging the issue, so I thought it would be worth a bit of attention.
The commenter would like to disagree with the point I am making, but to do that he needs to deal directly with the facts I’ve presented. That’s a bit of a problem, because my facts don’t really give him any grounds for complaint. The Unicorn Hypothesis does indeed imply the same consequences as the Autonomy Hypothesis, and makes it very clear, even to him, that merely predicting a similar set of consequences gives us no grounds for calling the conclusion justified. Rather than admit that it’s a good example of the flaw, therefore, he tries to make it sound like a flawed example. He does this by making a stink about an issue that isn’t even part of the topic under discussion.
Your UH differs very, very significantly from your GH. Think about the GH for a moment – my intense distaste for it aside – its depth of scope and definition at least theoretically allows for a decent subset of nuanced consequences in more than one area of reality – you have lists of things we’d expect to see in several different areas of reality were the subject of your GH to exist in actuality. OTOH, think about this UH for moment. The only consequence we can reliably deduce is that if the subjects of your UH exist, we should see tensions and crises. Problem is, the singular prediction of the UH is an absurdly high-order abstraction, nothing near the nuanced, myriad abstractions of your GH in predictive or explanatory power.
We’ve been discussion what it takes for a conclusion to be “justified,” and in particular whether it’s reasonable to call a conclusion “justified” on the grounds that (in one’s opinion) the evidence is consistent with the belief. There has been no debate on the topic of whether there’s a requirement for the hypothesis to predict “nuanced” conclusions, or to make predictions in more than one area of reality, or to have “lists” of a certain minimum number of items. These are purely ad hoc, spur-of-the-moment requirements made up for the sole purpose of giving the Unicorn Hypothesis something to fail.
The reason I call this a classic example of misdirection is because the commenter is essentially arguing that, in order to lead to valid conclusions, a hypothesis needs to meet a certain number of other criteria besides or in addition to the criterion of c0nsistency with the evidence. In other words, the commenter is actually conceding the point that we can’t justify our conclusions solely on the grounds that we’ve managed to make them predict the same consequences as some other hypothesis. Yet, perversely, though he reinforces my point, he phrases his concession in such a way as to make it sound like he’s actually proved me wrong.
Notice, he calls it “another silly, impertinent scenario,” even though he has to grudgingly concede that “of course I agree with you.” He does not like the conclusion he is forced to agree with, so he tries to make it sound like an outlier, a special case that does not apply to any other circumstance. If you can’t beat it, isolate it! And, when we look at his arguments, it’s easy to see that he’s straining for pretexts on which to base his rejection.
Consider his initial claim that “a single sample is seldom sufficient.” That’s generally a true principle, but it applies to the actual evidence you collect in order to evaluate your hypothesis, not to the hypothesis itself. The scope of evidence for the Unicorn Hypothesis is the whole human history of world affairs, at the local, regional, national and international level. That’s way more than one sample: entire encyclopedias have been written on the history of politics. Even the relatively short history of the United States, from the 13 colonies to the present day, would take multiple semesters to cover in high school or college.
So we’re far from dealing with a mere “single sample” in the case of the Unicorn Hypothesis. There’s a huge body of evidence that it’s consistent with. But our commenter has an alternative way to try and make the same objection apply: he says the Unicorn Hypothesis is capable of only a single prediction. “The only consequence we can reliably deduce is that if the subjects of your UH exist, we should see tensions and crises,” he writes. But once again, this is a spurious objection.
First of all, there’s no rule that says a valid hypothesis has to make more than one prediction. Let’s compare the Uniform Acceleration Hypothesis (which says that gravity accelerates all masses equally) with the Proportional Acceleration Hypothesis (which says that gravity accelerates objects in proportion to their mass). The Uniform Hypothesis makes one prediction: that a heavy object and a light object will fall at the same speed (apart from friction or air resistance) and the Proportional Hypothesis makes the equally singular prediction that the heavier object will fall faster than the lighter one (again, apart from friction and air resistance).
You don’t need for either hypothesis to make more than one prediction in order to test which of the two is closer to the truth. I remember watching a film in my high school science class showing a feather and a hardball being dropped inside a giant, airless glass cylinder. They both fell at the same speed, consistent with the Uniform Acceleration Hypothesis (and incidently confirming Galileo’s experiment on the Tower of Pisa). A single predicted consequence is just fine, provided it predicts different results for different hypotheses.
Secondly, the Unicorn Hypothesis does indeed make several “nuanced” predictions. For example, in the hypothesis, the unicorns do not hate us, nor have any particular desire to inflict undue suffering on us. They simply want to keep us too distracted to look for unicorns. This implies that world affairs will sometimes produce results that benefit mankind, like emancipating and enfranchising slaves and women. Too much bad news might arouse human suspicions, you see. But they have no particular love for mankind either, so they have no qualms about letting us endure atrocities and disasters, as long as it keeps our minds preoccupied with world affairs.
Another consequence is that the unicorns can’t be too obvious about how they exert their control, so the real causes for wars and conflicts will have to be complex and subtle, with humans believing that they are freely exercising their own sovereignty and control. The best institution for creating this kind of situation is democracy, where people can be influenced in subtle ways, and manipulated politically, all while wholeheartedly believing that they are free and are governing themselves. Thus, we can predict, based on the Unicorn Hypothesis, that democracy will tend to spread and gain power (at least in areas that are less subject to traditional, tyrannical forms of rule).
Further examples are left as an exercise for the reader, but the point is that our commenter accused the Unicorn Hypothesis of implying only a single consequence simply because he didn’t want to find more than one. There are plenty there to be found if we’re willing to look, but what he was looking for was an excuse to reject the whole scenario. It led to a conclusion he could not adequately deny, so he sought to find some pretext for discrediting it.
But the real genius of this argument is the way it opens up whole new cans of worms for us to argue about, like whether or not the Unicorn Hypothesis is really similar enough to the Gospel Hypothesis to make this example applicable. How similar does it need to be? How do we measure similarity? What constitutes a “nuanced” prediction? How much nuance is needed? How many different areas of reality need to be covered? What constitutes “covering” an area of reality? What constitutes an “area” of reality?
We could argue for months, or years, about these and a host of other tangential topics, and never get back to the main point of the post, which was that we can’t justify our beliefs solely on the grounds that they seem “consistent” (to us) with the evidence. And that, I think, was at least the subconscious goal of the tactic of “agreeing” while simultaneously trying to make it sound like he was disproving my point.
My point is that rationalization—the art of convincing ourselves that the evidence is consistent with our beliefs—is not a valid means of justifying what we believe in. It is too easy for us to be satisfied with low expectations regarding what supports our beliefs, and unreasonably high standards regarding what contradicts our beliefs. We’re to prone to believe, with little justification, that the evidence matches what we want to believe in. We need to test our beliefs, using techniques like comparing the Myth Hypothesis with competing hypotheses to see which matches the facts better than the alternatives.
We can know the truth, but we have to want to know the truth. If we don’t want to know the truth, then the Loser’s Compromise is waiting to serve us. And we can make whatever bogus objections we like in order to defend it.