27 Comments

Evolution is not intelligent in the way humans are, but it is an optimization process. While the causal process may differ, evolution, like humans, makes things in a certain way because it fulfills a certain function. This is what we say when we mean that evolution designs things to do something.

In other words, evolution clumsily finds a local optimum for doing something.

Expand full comment

the chemicals it 'dumps' are, ultimately, released into synapses by axons

If we grant that only what's transmitted at the synapses is relevant, what would distinguish a signal processor from something else? It seems important to have a near-mode model of what would fail to function as a signal processor.

What about if the synapses work stochastically? Then (assuming determinism) the action at a synapse in a particular instance would be caused in part by conditions outside the synapse--that is, by causes that aren't also information.

Expand full comment

affect functions through neurons, just like cognition; the chemicals it 'dumps' are, ultimately, released into synapses by axons, the same mechanism used by cognition. Model all the neurons, and you've modeled both affect and cognition.

Actually, that seems right.

Expand full comment

My neuroanatomy knowledge is also limited, but my understanding is that affect functions through neurons, just like cognition; the chemicals it 'dumps' are, ultimately, released into synapses by axons, the same mechanism used by cognition. Model all the neurons, and you've modeled both affect and cognition.

Expand full comment

Isn't it time to retire the idea that a "tautology" is necessarily a trivial observation? 

Expand full comment

Yeahh... I think it is so difficult to do this.

Expand full comment

Two problems, one related to the blog post, another in some of the discussions:

a) The argument that the mind is like our current most advanced artifact is a "confusion of genres". From the fact that a mind can produce a device of category 'x' doesn't follow that a mind is a device of category 'x'. At best, it can be argued that it is at least as complex as devices of category 'x'. Whether they are more or not, though, is an open and unanswerable question, given that a future device category might be more complex than those in 'x', at which point said new device category becomes the lowest complexity estimation for a mind.

So: in the XVI century, clocks. Nowadays, computers. In future, something else.

b) Arguing against free will because hard sciences all show determinism isn't valid, for hard sciences are heavily dependent on math and formal logic, both of which are, inherently and by design, fully deterministic. This means that determinism is a precondition and formal premise of said sciences, and hence that determinism cannot be *concluded* from them, as doing so would be a clear cut case of begging the question.

It should be noted that a related incorrect reasoning is that of trying to reduce mind to some kind of logical language. Logic is a feature of minds, so the error, depending on how it's constructed, can be either a case of taking the part for the whole, or a fallacy of composition.

Expand full comment

That may the definitive flaw in the theory of evolution.There is more than one flaw in the theory, Another flaw in evolutionary theory is DNA. Trying to equate the elegance of DNA to brute aimless chance is clumsy and muddled.How can one accept the absolute speed of light and reject design?

Expand full comment

"All of which goes to show that signal processing theory is far from a tautology, even if every object can be seen as in some way processing signals."

"The medium is the message"

Expand full comment

Caplan once argued that libertarian politics mesh with rejecting Hobbes' belief in materialism/determinism, ethical subjectivism and "depravity".

Then Robin's argument, which depends on materialist thinking, begs the question for Caplan. (And while Caplan accuses Robin of having strange beliefs, Caplan's beliefs take the cake.)

I'm with Caplan against Hobbes only as regards "depravity." I'd say Robin is with Caplan only as regards believing in ethical objectivism (and occasionally Robin sounds agnostic on that). Most importantly, Robin is with Hobbes on depravity; at least I think homo hypocritus is depraved.

Expand full comment

Caplan once argued that libertarian politics mesh with rejecting Hobbes' belief in materialism/determinism, ethical subjectivism and "depravity". Oddly enough I sided with Hobbes on all those issues while identifying as a libertarian. I don't identify as a libertarian any more, not because I've changed my position on much over the years, but just to "keep my identity small".

Expand full comment

Objections to functional talk in biology quickly become tiresome.  If you can't stomach people saying that legs are for walking and eyes are for seeing, you need to get with the program.

All science outside physics is necessarily vague, and accordingly, function talk has proved indispensable in biology. But functions have been invoked to answer distinct questions, and for some, they aren't sound.

If you ask, why did the heart evolve? the response, "to circulate the blood," is illuminating despite being incomplete. But if you ask about some unknown characteristic of the heart (if any today remain), it's unsound to try to infer this characteristic from its optimality for circulating the blood.

It's illuminating to say the brain's function is signal processing, just as it is to say circulation is the heart's function, but the inference from optimal function to structure is unsound.

Expand full comment

 I can stomach that there is a non-trivial sense in which that is true [1], but design talk generally (and Robin_Hanson's post specifically) will lead you to think that the features are accomplishing their role in some optimal sense, rather than simply the best out of a local neighborhood.

If you think of the brain as designed *to* process signals, you will miss out on just how kludgey it is, and how much it diverges from what a (human-like) intelligence would do in making an optimal information processor, *and* how much it is influenced by having performed other roles and constrained to be compatible with previous iterations.

[1] the sense that those things *accomplish* that role, and that, among all their effects on the body, if it were fail in that role, it would have the most negative effect on the body's continued survival -- e.g. the heard is for pumping blood in that, if it stopped pumping blood, this would be more fatal than if it failed to dissipate heat to the body, which is another effect it has

Expand full comment

Objections to functional talk in biology quickly become tiresome.  If you can't stomach people saying that legs are for walking and eyes are for seeing, you need to get with the program.

Expand full comment

Congratulations, you may be the only person who has ever proved Bryan Caplan wrong.

Expand full comment

 Hmm.  I've found a number of interesting discussions of your point (some pro, some con) on the web.  Thanks for the suggestion to think in these terms.

Expand full comment