Thursday, September 08, 2016

Did Evolution Select Us For False Beliefs?

9 Comments:

Blogger The Mystic said...

Dude. I am shocked you even posted this. My level of shock is so severe, in fact, that I suspect I must be missing something big.

At first, I thought this guy was merely proposing that it is possible that evolution selects for a kind of fitness in creatures which does not require that the creatures perceive reality correctly. But no, he seems to be saying something much stronger:

"Given an arbitrary world and arbitrary fitness functions, an organism that sees reality as it is will never be more fit than an organism of equal complexity that sees none of reality but that is just tuned to fitness."

So he thinks that perceiving reality "as it is" will NEVER provide superiority in fitness over an organism which is merely equally complex and "just tuned to fitness"?

???

I mean, maybe he has this worked out way better than it sounds, but this sounds like total BS. First, if he means what I suspect he means by seeing reality "as it is" (i.e. correctly) then he's arguing that this capacity is incapable of providing a superiority over the creature which is "just tuned to fitness" without having the quality of correctly perceiving reality.

That seems obviously false. While it's theoretically possible that the capacity to perceive correctly is not required for being superior (in terms of fitness) to a creature which has the capacity, his theory seems to be arguing that perceiving reality correctly has no value whatsoever, since it can never bring superiority over another creature which is "just tuned to fitness."

Unless, I guess, the degree of that tuning is just ratcheted up in the reality-blind creature until the perception of the reality-aware creature is made irrelevant?

I really just can't even make sense of what this guy's trying to say.

If it's just the same sort of thing as "Hey, look; we could be brains in vats" where "we could be brains in vats" is replaced with "we could be reality-blind-yet-fit-nonetheless creatures" then whatever. But if he's trying to say that evolution requires that we be chosen for reality blindness, then he's a smacktard.

Also, his example in which he tries to demonstrate a reality-blind creature fails:

" ... an organism tuned to fitness might see small and large quantities of some resource as, say, red, to indicate low fitness, whereas they might see intermediate quantities as green, to indicate high fitness. Its perceptions will be tuned to fitness, but not to truth. It won't see any distinction between small and large — it only sees red — even though such a distinction exists in reality."

Actually, the perceptions are tuned to truth. They may be misleading in their representation of reality in some aspects, but the fact that the medium volume of water is good for the creature is properly perceived by the creature. Therefore, the perception is tuned to truth.

This guy sucks, does he not? I mean, he's talking about something really complicated and this is a short NPR article, but this comes across as pretty dumb to me.

9:41 AM  
Blogger Aa said...

Perhaps send a quick link to PZ Myers at Pharyngula, he is an evolutionary biologist after all. Just don't mention SJW stuff and hope he comments on it.

10:50 AM  
Anonymous Anonymous said...

If this guy is right, wouldn't science--and really, inquiry more broadly--be impossible? I'm not sure, as I didn't really give it a fair enough, in depth read, but it seems to me that Peirce is rather convincing concerning the idea that the chance of successful inquiry without man's attunement to nature is literally, statistically zero. One of the famous quotes on the matter (which I'm sure you've read many times):

"There is a reason, an interpretation, a logic, in the course of scientific advance, and this indisputably proves to him who has perceptions of rational or significant relations, that man's mind must have been attuned to the truth of things in order to discover what he has discovered. It is the very bedrock of logical truth." (CP 6.476)

11:17 AM  
Blogger The Mystic said...

Anonymous - I'm with you. Not only would inquiry be impossible, but it seems to me he'd have to be arguing for a theory that creatures are selected for by evolution in spite of utterly senseless cognition. We could be entirely detached from reality, intellectually/mentally, but it would mean that our survival in no way depends on our (conscious?) cognitive activity. We would basically just think whatever the hell we think while our independent-from-all-that fitness tuning carries us along in the reality which we cannot detect or understand.

It's a brain-in-vats analogue, as far as I can tell. I don't know how he intends to support this contention from an evolutionary biology perspective; at best, it seems to me he could argue that it's unlikely that our senses are selected for perfect objectivity, but rather that our senses represent to us data in manners conducive to our survival.

Which, of course, we already know. Where he seems to obviously fail, to me, is when he attempts to argue that this can somehow be the case while it is also true that our senses are absolutely unconnected to objective reality (e.g. the "way things are" or the truth). The example he provides, even, seems clearly to fail in that regard. The senses provide accurate data about the real world in his example; the real world just isn't necessarily exactly as the senses portray it.

Which, again, is a fact of which we have long been aware.

2:20 PM  
Blogger Winston Smith said...

Wow you guys are all over this...

I'll admit, I read it incautiously and in a very charitable mood.

He says this:

"Given an arbitrary world and arbitrary fitness functions, an organism that sees reality as it is will never be more fit than an organism of equal complexity that sees none of reality but that is just tuned to fitness."

I think the Mystic's right that this has roughly the status of a skeptical hypothesis. Doesn't it mean that, if we select any organism that is built to have (mostly) true beliefs, there is always another possible organism that is just as fit but is built to believe falsehoods? (roughly. I'm being sloppy here.)

So anyway...does he end up mistaking that with a stronger thesis? I'll go back and actually read it this time.

Anyway, that's why I said "maybe," but the thesis does invite/demand a more critical response than that...

And yeah, A, I'm totally down with the Peirce quote/thought!

2:39 PM  
Blogger The Mystic said...

Yeah, that is what he says, and no, that doesn't mean what you describe. What you describe is what I said I initially thought he was arguing; it's a boring skeptical hypothesis that says it's possible that there are reality-blind creatures which are tuned to fitness in a manner superior to reality-aware creatures.

Yes, that is possible.

But what he actually says in that very quote is:

"Given an arbitrary world and arbitrary fitness functions, an organism that sees reality as it is will never be more fit than an organism of equal complexity that sees none of reality but that is just tuned to fitness."

What the hell is that? Are we just precluding the idea that seeing reality as it is could contribute positively to fitness tuning? If not, then how the hell will this possibly make sense? He's saying something like "When comparing two creatures' fitness, a reality-aware creature will always be inferior to a reality-blind creature." (note: that's not all of what he says, the remainder of which I will get to below because adding it makes the whole thing borderline-incoherent)

That is stupid. You can't possibly establish a universal rule like that without establishing that being aware of reality is catastrophically detrimental to an organism's fitness.

And that's not even the entire stupidity of that mess of a thesis; we're talking about which creature is better tuned to fitness and his thesis is that the one which is tuned to fitness is better than the reality-aware creature. He must be saying that regardless of the fitness of the reality-aware creature, any given reality-blind creature which merely satisfies the requirements of "equal complexity" and being "tuned to fitness" is superior.

wat.

If he's just outright excluding the quality of being reality-aware from possibly positively contributing to fitness, then this remains at least possible, I guess, but then one is left to ask what this supposition about complexity is doing inserted in here.. It's like some sort of biological coherentism or some crap; does he not think a simpler organism could be more fit than a more complex one? Ugh, I don't know wtf he is talking about.

Anyway, if he's not outright excluding the possibility that being reality-aware contributes positively to an organism's fitness, then this basically makes no sense, since he's alleging that a set of fitness qualities which contains a particular positive quality will always be inferior to an equally complex [for whatever reason] set of qualities which merely lacks that particular positive quality.

And to confirm that he is, indeed, arguing for such a radical position and not the boring brain-in-a-vat analogue, you can consult the article and find things like:

"... an organism tuned to fitness might see small and large quantities of some resource as, say, red, to indicate low fitness, whereas they might see intermediate quantities as green, to indicate high fitness. Its perceptions will be tuned to fitness, but not to truth. It won't see any distinction between small and large — it only sees red — even though such a distinction exists in reality."

Which I already quoted because it's SO STUPID.

He's not crafting a hypothetical scenario in which the organism he describes cannot detect truth; the organism is still clearly required to accurately detect facts about reality (e.g. the quantity of a resource necessary to its survival). He's just pointing out that what might really be a difference in size or quantity could be represented to the organism symbolically by a color. Then he's all like "but the organism doesn't see any distinction between small and large"

EXCEPT: YES IT DOES; IT SEES THOSE DISTINCTIONS IN TERMS OF COLOR. YOU SAID IT DOES IN THE SENTENCE THAT PRECEDES THAT STUPID ONE.

He's stupid. That article is stupid.

Stupid.

4:17 PM  
Anonymous John Plato said...

Watch the TED Talk linked at the end of the NPR article. Hoffman does a better job of explaining himself in it and addresses some of your questions there.

What I see him saying is:

Perceiving reality "as it is" means perceiving and processing a lot of useless data, and doing this comes at a major evolutionary cost.

There are cheats and hacks that allow creatures to obtain similar evolutionary results at a far lower energy investment, and evolution favors perceptions which build on these cheats and hacks. Creatures that use them out-compete the ones that don't.

This means the perceptual landscapes of evolutionary survivors are likely to be useful but also wrong in important ways.

The cheaper the evolutionary shortcut, the greater the fitness advantage to the organism, and the more profoundly wrong its perceptual landscape likely is.

In the TED Talk, Hoffman gives the example of the Australian jewel beetle that almost went extinct because the males were mating with discarded beer bottles that resembled female jewel beetles to them.

The males used cheap evolutionary hacks to recognize females as anything dimpled, glossy and the right shade of brown, but these hacks weren't actually giving them a true picture of the reality around (or underneath) them.

* * *

To me, this is interesting to think about and play with. We know our perceptions are flawed, but we tend to think of our flaws as rare exceptions that somehow snuck past the blind watchmaker of evolution. We look at an optical illusion like the checker shadow illusion as an outlier, a tiny glitch in an otherwise amazingly accurate system of visual perception.

But Hoffman wants us to conclude that the entire visual system is likely just as useful and just as deceptive as that Checkerboard shadow illusion, and for the same reasons: Because our perceptions evolved not to represent all of reality but only the parts of it that were meaningful to our procreation, and those parts only as cheaply as possible.

Because of this, we should expect our entire perceptual matrix to be wrong in important ways, to omit or misrepresent aspects of reality not crucial to our survival, and to use shortcuts and heuristics that may be good enough to keep us alive, but inaccurate at showing us the full picture of reality.

There's some sense in which this is obviously true. We don't see in ultraviolet, for instance, so we don't see the complex designs on flowers that pollinators do. When we look at a flower, we don't see the full reality of it.

But Hoffman means this in a much vaster way. He believes we may not even be the kind of beings we think we are, or live in the kind of universe we think we do, victims of the same kind of cosmic confusion as the Australian jewel beetles, with our senses telling us reality is one thing when it is something else entirely, something much more complex.

Hoffman suggests that our perceptions may be something like a computer desktop, a simplified interface meant to shield us from the deep complexities of reality and to guide adaptive behavior.

In the TED talk, he suggests that this deeper reality could be "a vast machine" or "a vast, interacting network of conscious agents," but he doesn't claim to know, and only says that these are ideas that are worth exploring.

It would be a shame if these few suggestions overshadowed his much more interesting argument, that evolution is a pathologically stingy gift-giver, and we should be suspect of anything it has given us, no matter how solid or shiny it seems.

2:21 AM  
Blogger Winston Smith said...

I still haven't even actually given the article a respectable read, much less seen the whole talk.

But yeah, I was going to say, by way of pointing to perceptual shortcuts: we don't perceive objects as swarms of subatomic particles...and sure as hell don't perceive subatomic particles as probability distributions (whatever that would mean...)

OTOH, science seems to be a way of getting around this stuff. *Maybe* nature made us so deluded that it's impossible for us to see our way clear of it...but that's a stretch. Evolution isn't an Evil Genius dedicated to deceiving us...just a cheapskate looking for a workaround that satisfices rather than optimizes.

8:54 AM  
Blogger The Mystic said...

The problem is; none of this plausible/semi-interesting stuff is as radical as his claims are.

He is making a categorical distinction between perceiving the truth and fitness-tuned perceptions when there cannot be such a distinction unless perceptions become utterly irrelevant to fitness. There could be hypothetical creatures, little idioblobs floating in nutrient pools, which are quite fit for survival without needing to perceive the world at all and which have rich inner lives that are utterly devoid of practical value, being entirely detached from reality, but that's not what he talks about. Instead, he talks about creatures whose perceptions are critical to their survival because they provide symbolic representations of facts about the real world about which the creature is then capable of making decisions which are critical to its survival.

And in spite of his contention that this is an example of being "tuned to fitness" but not "tuned to truth," his example does not present such a distinction (unless the tuning is some sort of inclusive spectrum, and he's saying it leans more towards fitness than truth, but that fails to support or be relevant to his thesis, as well). It is an example of successfully perceiving facts about the world; it is merely true that the creature could make improper inferences about its perceptions (e.g. color doesn't represent quantity).

And this is all stuff we already know, and none of it supports his radical thesis.

Lame.

LAME LAME LAME.

12:59 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home