#64 On the degrees of skyhookery

Dear Jamie,

Your letter reminded me of one of the most mind-boggling experiences I had. I watched Derren Brown’s show Miracle. In Miracle, before the segment started, he told the audience that he will use the tricks of religious healers to heal them and that there was nothing magic going on. He then conjured something that certainly looked to everyone like faith-healing. As the show ended, a family member of mine said “I don’t know… I’m still skeptical”. I was bewildered, “What?” I thought, “Skeptical that he healed them after admitting it was a trick? Or skeptical of his admission of a trick, it was a double trick and actually magic?”

The magic comparison is useful, Deutsch uses it a lot. He says that reality itself should be seen as one big conjuring trick, all the time. We have only seen behind the curtains once we’ve explained how something is done, not when we can predict what will be pulled out of the hat. But that’s enough foreshadowing.

Thank you for taking the time to specify the skyhooks we’re talking about. We agree. Skyhooks like consciousness and creativity are the ones we butt heads on. I’ll add to this. We even agree on what we call a skyhook and what we don’t. However we disagree how bad it is to be a skyhook. You condemn all skyhooks, while I think they’re necessary and there are degrees of good and bad skyhookery.

Here I explain why. This is going to be a bit more involved than a normal letter, it’s my Christmas gift to me, I mean you… you, gift for you. As you read please ignore the content of the theories and whether I’m accurately representing them, that doesn’t matter. What I’m talking about is how we make progress moving from theory to theory, concept to concept.

If you believe progress can be made in explaining the world, then that means some theories improve on other theories. Let’s say Popper improved on the major incumbent, justification from induction (remember, it doesn’t matter if this was indeed the transition that occured, I’m talking about any claimed improvement from any theory to any theory.) The incumbent theory was something like:

Knowledge is accumulated through observation of reality.

Then Popper came along with his theory:

We invent conjectures, then test them against reality and ditch what is refuted. What remains is knowledge.

Let’s say we’re convinced that the second theory is superior to the first, and that the reason we are convinced was because we determined you can’t confirm theories, you can only falsify them. So really only one bit makes this theory superior, and that’s the “test them against reality and ditch what is refuted” bit. I’ll refer to just this bit as C&R (Conjecture and Refutation).

However, C&R alone isn’t a theory of knowledge. For it to be actually explain how we make progress in knowedge, it needs other new concepts to fit around it. So bits like “we creatively invent” are complementary to C&R as a mechinism in a theory of knowledge. True, we’ve replaced one theory with its problems (what is “observation”? how does it “accumulate”?) with another theory with new problems (what is this spooky “creativity”? how is a conjecture actually “refuted”? is really all that remains, “knoweldge”?). The new theory however has a nugget of improvement that makes it a superior theory. New problems can be an indicator of progress.

These new problems beg for answers. Let’s focus on the old favourite. “Creativity” is a placeholder for where some mystery has been consolidated after the C&R part of the puzzle of knowledge has been accepted. We know its rough shape owing to the needs of C&R, and the reason the theory is around in the first place, but not much else. It’s very much mysterious.

What I propose we do in this situation is to continue to hold all aspects of this (Creativity, Knoweldge, C&R itself, etc) tentative, and try to solve each problem with reference to the concepts we invented. What I think you’re wrong to do is assume the answer to the new problem of creativity is “it comes ex nihilo”. Having no evidence of where it comes from is not evidence that it comes from nowhere.

So I suppose what I’m saying is that a skyhook (like creativity) is as good as the good theories (like C&R) that give rise to its necessity.

You can argue with the specific truth of C&R, and the unreality of epistemology, but my argument here is about how progress is made in any science. To prove this I’m going to show how this happened with our precious theory of evolution by natural selection.

Let’s position ourselves in 1860s. The problem of the day was why living things speciate. One incumbent was Lamarkism:

Creatures morph as they live and turn into new species over time.

Darwin changed the world with

The offspring of creatures are born with variations, and the environment selects the most fit.

This theory contains real progress. The nugget here is natural selection. But, like any discovery, its inclusion pushed mystery to new concepts. What are these “variations”? Just what is “selection”? How do you define “fit”?*

Those things in double quotes are each skyhooks (to us still in the 1860s). Mr Darwin do you expect me to believe creatures don’t change but somehow “vary” as they give birth? What is this magic “selection” occuring from up high? Who defines what “fit” means? And those critics who assumed variation was “ex nihilo” were condemned to spin their wheels forever because those who embraced the skyhook of variation as worth studying later discovered a new theory to explain “variation” in the mutation of DNA. It was beautiful, for one because it’s a clearly reliable crane, but also because it has led to wonderful new questions, replete with its own skyhooks. How do genes “encode” traits? Is there anything that can’t be “expressed”? Could we make artificial genes?

And back here in 2018, “selection” still remains a skyhook. Originally necessitated by 24 Karat natural selection, it continues to be reinforced into undeniability through excellent explanations of mutation, DNA and information, despite never itself being explained, yet.

Let’s indeed look at it like a bet. It sounds like you want to bet on an outcome, i.e. guessing the next bird out of the magician’s hat will be a crane. But that’s the wrong kind of bird. Whatever the next theory, its appearance will be a white swan (a positive Talebian black swan). Knowledge is prone to this precisely because scientists are too confident in their old paradigms. We need a skyhook that floats free of traditional beliefs just enough to escape the old guard. So long as we remain vigilant in demanding good theories and testing them, science’s integrity can survive the flight of fancy.** My bet will take a page from Taleb’s risk strategy for an unknown future. Instead of betting on a specific outcome, we should instead bet on our exposure. What processes and rules help us overcome our cognitive baises, loosen our unquestioned assumptions and avoid dead-end dogmas? My money’s on C&R.

Desiring exposure,

Mat

* C&R and natural selection are both selection-based arguments. It’s unfortunate, I think it muddles my point. Ignore selection. The only point I’m making is how skyhooks inheret the ”goodness” of the explanations that give rise to them.

** I’m willing to say most, if not all progress involves some kind of skyhook. The best skyhooks of the past were the stepping stones to real knowledge. Our best skyhooks now are still that. Corollory: Our most fundamental theories are themselves skyhooks.


Originally published at unlamed.