The conclusion seems pretty obvious; we only see light in a certain (relevant) spectrum, we only taste certain (relevant) chemicals, we only hear sounds in certain (relevant) frequencies, etc., because it doesn't help us (and in fact distracts us) to perceive more information.
Also, it seems like a big aspect he's missing is that within each relevant spectra, we've gotten pretty well tuned to reality. And that's simply because oftentimes, fitness is highly correlated with perceiving reality. For example, seeing all physical objects realistically is important for survival, since if you can't see perceive certain physical objects, then predators/prey may take advantage of the failure (e.g., zebras blending in with tall grass).
I seemed to get a bit more from his TED talk than I did from the article. I think the main thrust is an attempt to show that not only do we not see reality (which I think many people would agree is probably true), but that in the vast majority of evolutionary simulation models, perceiving reality is actually a disadvantage. The new angle seemed to be the experimental data from computer models.
"By convention sweet is sweet, bitter is bitter, hot is hot, cold is cold, color is color; but in truth there are only atoms and the void."
or:
"We know nothing accurately in reality, but [only] as it changes according to the bodily condition, and the constitution of those things that flow upon [the body] and impinge upon it."
Nothing like a little LSD to give a little insight to the world.
The mind isn't a learning device, it is a filtering device. Much along the same lines of "The web is a collection of everything we know. Still you'll be lost if you don't use a search engine to filter everything you don't want out"
>"We see things not as they are. We see things as we are." attributed to Anaïs Nin
No. We see things as they are observable through our senses. We're working with noisy and incomplete information, but still information that ultimately comes from the world rather than from us.
I asked the same question. My gut reaction is that this is just something that seems philosophically appetizing on the surface but breaks down when you start asking questions.
"We should view genes, not organisms, as the fundamental unit of propagation. They optimize for their own survival, not necessarily the organisms." --- this is an insight.
A testable hypothesis being, "altruistic behavior that seems detrimental to organisms clearly exists, detrimental to the organisms survival. Perhaps this behavior propagates because genes that encourage this type of behavior ensure survival of other copies of themselves within other organisms in the population."
I'm having a hard time discerning A) the exact insight here and B) whether or not there's a hypothesis to go along with it.
> "We should view genes, not organisms, as the fundamental unit of propagation. They optimize for their own survival, not necessarily the organisms." --- this is an insight.
Richard Dawkins wrote a book on just that. "The Selfish Gene" (~1970). However, recent epigenetics discounts just the gene as the sole 'unit' of propogation.
Yeah, that's why I brought it up. Should have made the reference more clear. :) Dawkins had an insight and shared it with the world. I'm not sure this pop-sci article is the same magnitude of revelation or what, if anything, is actually there.
>And that's simply because oftentimes, fitness is highly correlated with perceiving reality.
It's more than that. Building a whole range of fit animals whose brain-design is approximately correct in-general ends up being much simpler and easier than custom-building a highly fit but generally wrong brain for each animal.
I think what you're missing (if anything at all) is that Hoffman is playing word games.
Reality and perception are heavily overloaded words. For example, you interpret perception in terms of excitations from spectral stimuli. For Hoffman's theories to make sense (i.e. applying the Principle of Charity), we need to define perception in terms of a mental model of the cause+effect relationship among physical things.
Both of are you admitting that our modes of perception discard relevant information, which could cause us to make wrong deductions. But we need Hoffman's more encompassing definition to allow us to say that perception is intrinsically and irredeemably at odds with reality. The problem is that were he strictly correct, Hoffman would never be able to make the statement in the first place. The false reality isn't absolute if you can deduce that there exists another reality. And like with cryptography, that little chink in the armor plus applied logic is often all you need to tear down the wall.
Clearly humans are capable of seeing past the immediate false realities our brains construct, permitting us to examine indirect realities (which may or may not be the same thing as "objective reality", whatever that's supposed to mean). Which suggests that there isn't a single "reality" that we perceive as humans; there are different levels and dimensions of perception and "reality" which we could identify with similar rigor, and clearly Hoffman's assertion cannot apply to them all equally. And the assertions are obvious trivialities if he's referring to the basic mental models we construct in our daily lives.
To put it more concretely, science and math allow us to see past our immediate physical senses and mental modeling, at least in so far as they allow us to make predictions which we can and do subsequently verify. That much is clear, and that destroys Hoffman's argument in as much as it relies on handwaving about genetics and evolution. The rest of the argument would appear to just rest on ground well trodden by various philosophies of epistemology, semiotics, etc. And that's a quagmire I'd personally prefer to stay clear of because it's not at all clear we understand what the relevant questions are.
I never studied semiotics, but I think we could all do better to keep in mind that words are funny things, and just like the way we perceive physical reality, we tend to subconsciously fill in the gaps (or assume they're filled in) to construct reasonable interpretations of theories. Reading summaries of Hoffman's ideas my first thought was that he's relying on special definitions of overloaded words. Which suggests 1) his theory, if sound, only makes sense within his unique conceptual framework and thus the conclusions don't mean what they superficially mean, and 2) there are a lot of obscured, hidden premises that are either wrong or speculative.
Pedantism is a virtue, not a vice. When someone says that being overly pedantic is not constructive, my first thought that is that 1) yes, that's a good assessment, but 2) if it's possible to be overly pedantic (but still reasonable) then the fact it's being unproductive is more a reflection of a lack of substance in a discussion. If people can point at meaningful problems with interpretation of words and phrases, that suggests the discussion is built on sand. Similarly, unless a hypothesis is testable and falsifiable, people should put little stock into it. Hoffman's notions are so abstract (literally, semantically) that my instinct is to turn and run. Which isn't to say he might not be on to something, but untangling the mess is probably more effort than discovering anew whatever useful insight lay beneath it all.
Okay, looking at his publications, a slightly more charitable view is that we have a man who has produced some useful science related to the mechanics and neurobiology of perception, and related selective pressures. But he also has a penchant for imbibing in speculative, philosophical extrapolation. Again, it all comes down to playing word games. His science is more concrete; the more abstract theorizing (i.e. the TED Talk-level stuff) rely on the nuance of language and surreptitiously redefining the context of the discussion.
His point is really that perceiving reality may not be correlated with fitness that well. It's a shame this work has been taken up by a lot of pop press philosophizing, because the actual research is quite interesting, and the philosophical surface layer is mostly not Hoffman's fault.
I wonder how many rationalists on HN this post will piss off before it falls off the main page.
"Reality", "science", "fact", and "logic": these are all arbitrary concepts and disciplines that are stuck in a limited worldview. There is no true objectivity we can experience as humans.
Just because we don't experience something or it doesn't fit with what we consider rational thought, it doesn't mean that that thing cannot exist.
However, we learn this truth from science itself- in seeing how other living things experience life and react and how it is so different than how we experience it- how we aren't made to be objective.
The means of showing us truth we've relied on is flawed. At this point, the rationalist understands why Plato divides into thing and form- because form is the only ideal that is an anchor when you realize that our experience is unreliable:
https://www.northampton.edu/Documents/Subsites/HaroldWeiss/I...
But, then when you accept form as the ideal and reject things, you have rejected everything we have to understand form. So, you fail to have anything dependably rational left.
So the options are the world is rational and understandable, or there is no truth, everything is a lie and we're living in a Lovecraftian horror where we are physically incapable of understanding reality.
Presented with the two options I'm inclined to believe the prior.
- If the latter is true and I believe the prior no vice.
- If the prior is true and you believe the latter however no virtue. There's no sense in making an attempt to understand the world. You're missing out on everything. Just giving up on understanding.
The set of all possible "Gods" includes deities who punish worship, who lie to their worshippers, who are completely two-faced in their dealings with mortals, who have completely random intentional urgings, who put everyone through the torture of infinite fractal realities, and/or anything else that throws the wrench in trying to rationally deduce the "goodness" of "God".
This is also compared with the possibility that humans are faggots who will talk about absolute bullshit to make themselves feel better. No offense to gay people but I'm talking about the mass of humanity that seeks to "feel" more using ass-backwards strategies because they trust their idiotic culture too much.
I'd prefer to remain silent until overwhelming proof forces me to consider "a being of superior influence on this reality who resides outside of visible reality". It could definitely exist (along with the things from the set of known unknowns and unknown unknowns). But I'm betting more on humans being faggots.
5) God does exist but the reason there is literally no evidence is that he wants to test us and will punish those who believe in something with no evidence and with send to heaven all those who don't believe in him.
Seriously. Pascal's wager is the dumbest wager I've ever heard.
> So the options are the world is rational and understandable, or there is no truth, everything is a lie and we're living in a Lovecraftian horror where we are physically incapable of understanding reality
This sounds like a restatement of the warm/cold reality thing. Most religious people are not actually deficient in capacity to reason but basically can't accept the cold universe where we pop screaming out of our mothers, live pointlessly and die in pain knowing that whatever we accomplished is transitory and amounts to nothing.
As opposed to, we're playing some kind of game where we get points for kind acts, piety, being nice to old folk, making money, etc, then we get rewarded for all of this in the afterlife by our loving creator.
>So the options are the world is rational and understandable, or there is no truth, everything is a lie and we're living in a Lovecraftian horror where we are physically incapable of understanding reality.
Random chaos is way too staticky for Lovecraft. The snow on old TVs when they've got no signal is random noise. Lovecraftianness requires distinct things with tentacles and little bubble-eyes everywhere. Totally different aesthetics ;-).
I feel like this works from a misinformed or co-opted definition of "rationalism." A rationalist is someone who tries to think clearly in order to win. Science, logic, the teachings of Plato; all of that stuff can inform a rationalist as they try to achieve their desired outcomes in the world.
This strongly reminds me of things touched on in Ian Hacking's Representing and Intervening. Early on he mentions that done medieval philosophers thought that crabs emerge out of tree trunks and later transform into ducks. Obviously that's total nonsense from our perspective but it might be a good enough heuristic like the world being flat is a really good assumption for everyday use.
At least you stated your premise up front. Everything after that is unnecessary. You may as well paste that as the entirety of every comment you ever make. Saying anything more is simply an exercise in creative writing. Indeed, one wonders why you would bother to say anything, period, if you cannot admit to the conceit of objectivity.
This is all very nice and abstract. But let's make it concrete.
Multiple studies have found that the more realistic we are about our prospects, the more likely we are to be depressed. The more optimistic we are, the better we are likely to do even though we are wrong.
As a result we have a constant bias towards believing that the world is likely to turn out better than objective evidence would show. And on this evidence we are biased to explore new things, have babies, and so on. Which improves our evolutionary fitness. Even though we perceive and predict the world less accurately.
A similar cognitive bias is towards seeing patterns where there is little evidence that they exist. That is because if something is random or a pattern, there is little cost to acting on random data if it isn't there, and a concrete benefit to following it if it is. So we are pattern seeking animals at the cost of potentially becoming convinced by astrology, ghosts, religion, and a wide variety of other superstitions.
Wanting to survive and reproduce, promotes survival and reproduction. What is rational at the gene-kin level is not necessarily optimal at an individual level.
This isn't really much different from Kant's own critique of pure reason, where he, following Hume, agrees that we cannot know the in-itself, and redefines metaphysics as more epistemological than ontological. More recently, but still long before this guy's TED talk, Plantinga had a very similar argument.
Claiming that "Evolution Bred Reality Out of Us" is just a weird way to phrase a logical problem, and it sneaks in the same naive realism that it claims to critique.
Kant and Hume are assuming a mind-body dualism as in Descartes.
A great many important philosophers in the last century or so have rejected this. This includes Whitehead, Husserl, Heidegger, Merleau-Ponty, the later Wittgenstein, Strawson, Putnam, and all the Pragmatists. And of course, Aristotle in classic Greek philosophy.
Oh, and by the way, Plantinga's argument is for the purpose of promoting Christianity, but radical mind-body dualism is quite contrary to traditional Christian belief.
This is kind of trivially obvious, confirmed by every human sense having a different response curve than other animals, different species being able to deal with different abstractions (eg, certain dogs being able to recognize conditionals but not predicates) and so on.
I have to wonder if he's doing a Straussian thing and "really" talking about selected-for responses to certain obvious truths particularly relevant to evolutionary biologists.
Try Googling for "Chrysippus' dog". That, at least, brings up hits like http://my.fit.edu/~aberdein/DogLogic.pdf. Just reading that paper made me realize that there were at least two dimensions in which the statement
"different species being able to deal with different abstractions (eg, certain dogs being able to recognize conditionals but not predicates)"
could be be misinterpreted. Assuming you were looking (as was I) for experimental evidence supporting that specific statement, I'm guessing we'll both be disappointed. But maybe not!
Hoffman's argument makes sense when you are talking about bacteria. But who thinks bacteria have a conscious understanding of reality?
More complex animals, and especially humans, collect an immense amount of sense data, from multiple sources, and use it to construct a general view of reality. It has to be accurate because they don't have a few, simple, stereotyped response patterns. Rather they have extremely complex patterns, highly customized to particular, unique circumstances, and the only way that can work is if their image of the world is itself complex and fits reality.
Here is an example that is immensely more simple than what humans do in their living: how could you design a software program to play chess if it was just all of the form "when x, do y"? The only way you can do it is if you have a representation of a chessboard and use it to plot out moves, and what the opponent might do in response. Now think of how much more you have to know about reality in order to, for instance, go shopping. Or run a software startup.
Hoffman goes wrong partly because he is violating some fundamental scientific principles.
According to modern science, the world is organized in levels of complexity, and for each higher level we get new phenomena that require new labels that don't apply to the lower ones. So for instance, an election consists of activity by human beings made of atoms, but to describe it in any meaningful way we have to use terms that are not from physics, like candidates and ballots.
In a similar way, when we move from the sensory and response mechanisms of bacteria to first multiple cell organisms, and then ones with nerves and a central nervous system, and then human beings, at each level we get phenomena for which we need new, quite valid concepts, including accurate understanding of reality.
To make this a little clearer, I am saying the phenomena of having a perception of reality requires very complex mechanisms that bacteria simply don't have. Therefore to note that bacteria don't have a perception of reality is simply irrelevant, and doesn't in any way prove humans don't. It's like saying atoms don't reproduce, therefore animals don't either.
By the way, if, as Hoffman claims, human beings don't have a perception of reality, then it would be meaningless to say that sometimes it is accurate and sometimes it is inaccurate, and certainly not something that could be studied scientifically. Does that seem really right?
An important point here is that bacterial sensing is tightly efficient, with the bacteria having only the absolute minimum needed to direct a simple response at the right time.
Human perception, on the other hand, has a vast amount of excess. So for instance, when you see a tree, you see hundreds of leaves, each with a fairly complex shape, even though all of this is usually of no use. Our senses and brains do this because with humans, you never know when some new situation will arise where perception that is normally useless will be useful, and so it makes sense to have vast excess in our reality perception capacities.
With bacteria, on the other hand, there are no novel situations, only a few stereotyped ones arising over and over again, so excess capacity would be useless and run counter to efficient survival.
One more point. Hoffman claims he is making a scientific argument, and it is based on scientific concepts like bacteria, sensing, and evolution. But if humans can't know reality, then these concepts, developed by humans, are false, and so his whole argument collapses. That is, if his argument is true, then it is false.
Actually, he is really making a philosophical argument, one based in the radical skepticism that has plagued many (but not all) Western philosophical schools since Descartes, but dressing it up in some quite poor science.
> What Hoffman's theorem says is the fitness-tuned critter will — almost always — win the evolution game.
It doesn't from the explanation in the article at all. It just states the reality critter can't beat the fitness-tuned critter.
Evolution doesn't 'care' if you have extra useless things, you won't lose them evolutionary. (Although if there is a energy or something cost it might 'care' about this negative)
Let's assume for a moment that someone believes that human beings can "perceive reality". Adding the axiom that "evolutionary theory is mostly correct" (which doesn't seem too bad), one concludes that because "perceiving reality" would require a large expenditure of energy, it must be necessary in order for optimal fitness, since if we are to believe it is with us, we must have evolved to do it. Similarly, walking is "necessary". NB. "Perceiving reality" seems to be defined by the person writing the question, which is a conflict of interest.
(I leave out the interesting but undesirably complex hypothesis that some strategies exist which do not involve perceiving reality but are not accessible by evolution for unspecified reasons)
This argument takes off by constructing a fitness function where perceiving reality is unnecessary. It then concludes by construction that the perception of reality won't evolve. However, there is no real "theorem" here: all that has happened is that it is now encoded in mathematical language that this belief could be true or false.
I have a phobia of insects(not if they are small enough for me to notice their details), crabs, prawns etc. All the not-animal-like-life, to be honest.
One day, at a supermarket, a friend of mine teased me with a crab in her hand. Later that day, we were discussing about this incident, and we realized that both of us look at insects very differently.
She looks at insects etc as "things that move", ie, lacking of will, like a toy or something.
I look at them as creatures who I cannot empathize with, and hence, whose motives I cannot understand. So they freak me out.
That conversation kind of felt like interfacing at the boundaries of our consciousness
I think that the point of view depicted in the article completely misses the fact that we are active part of a reality we live in.
It looks like the article (I am not entitled to talk about the original research, since I am not familiar with it by any means) unquestionably assumes that there exists some "real" (platonic, mystical, static, eternal, fundamental?) reality that we are unable to observe due to bounds of evolution. What's the point of this assumption?
When I look around, I see things made by humans for humans using our ideas about the human reality—is this reality "worse" or "less real" than some other reality? Yes, we are far removed from, e.g., a reality of existence of a plant or a bacteria, and the evolution drives us further from it.
Maybe it's the evolution process that creates very realities?
One of the neat things (i think) is that the visual and tactile systems in our bodies seem to always agree on what is perceived 'reality'. This on the surface seems like an argument against this proposal. But of course we can easily be fooling ourselves.
I once saw some experiments where they showed people some movies where the sound was offset from the picture by some time shift, either too slow or too fast. Within a surprisingly large range the mind could easily meld the two out of sync signals together so that it appeared that they were synchronized. Of course beyond the threshold that illusion fell apart, but still was pretty interesting.
The weaving of still pictures of a movie into ... a movie is also an interesting concept.
Maybe vision and tactile works the same way? Could one influence the other?
It's amazing how well our brains are built to pretend.
I remember seeing a report of a study in which a fake arm, cleverly manipulated within the subjects' fields of view, often succeeded in overriding both tactile and proprioceptive senses - some of the subjects thought the fake arm was their arm, because they saw it doing what they expected their own arm to do. As I recall, some subjects reported significant confusion and some discomfort on recognizing the actual nature of events, and I should think that's no wonder.
I don't have a cite to hand. I'll look for one if you like, but it's been so long that you'd probably be able to find it at least as quickly as I could. Fascinating stuff, in any case, and perhaps somewhat germane here.
Maybe not the same source, but V.S. Ramachandran's book, Phantoms in the Brain, covers his use of a set-up like that to help patients recover from phantom limb syndrome. Shockingly effective for what is, essentially, just an illusionist's trick.
It makes sense to grasp concepts like "huge/small amount of something" and it might even be less effort than hard wiring much and little red to the same neuron.
Also seeing "too much red" and "too little red" as the same, is no misconception and still seeing the truth. You won't be fooled it no thinking there is no red, when there is a lot, you would probably perceive too much or too little red as "in the danger zone" which is a true statement either way (given that too much and too little red is most often dangerous for your fitness).
The paper “Objects of consciousness” by Donald D. Hoffman[0][1] is an interesting read—if only for its attempt to build a formal model of consciousness and define what it means to “observe”, in the quantum sense.
They said the same thing of physicist, biologists, geologists, chemists, etc. before, that is until they found themselves irrelevant. I suspect in time, it will happen again.
in part this is because whenever something relevant comes up, it gets a name that is not philosophy.
Also because we're still shaking off perception of philosophers as useless navel-gazing victorians; post WWI, the field has become substantially more akin to formal math and logic.
The problem here is that the concept presented isn't scientifically rigorous as applied to human consciousness and perception. One, fitness simulations are reductive by design and trying to draw analogies from them to explain complex subjective phenomena is a stretch. Two, we need a testable hypothesis - I haven't seen one presented here.
It's an interesting concept, but the author provides us an example - an organism responding to a fitness function optimizing for Y resource where X < Y < Z doesn't distinguish between X and Z, just that X and Z are both "bad" - reducing complex quantities to binary outcomes. To then suggest physical reality could be "hidden" in these sorts of reductions without an analogy or testable example just seems philosophical.
The best supportive example I can come up with is - we obviously don't see "reality." Visible light, for example, is only a tiny fraction of the entire electromagnetic spectrum. Our ability to perceive and differentiate wavelengths in this range has been selected for. But language and formal logic --- abstract relationships, mathematics --- seems to be something we've discovered that describes a Universe of relationships we can't possibly entirely perceive. But we still utilize this reality, as we've incidentally discovered the rest of the EM spectrum by its interactions with other aspects of reality we can measure and the ability to describe abstract relationships (language).
So I guess I'm a bit confused, and I admit ignorance here --- my 2c here is an off-the-cuff response to the article provided without digging deeper. Is the argument there are rules and laws of the Universe we can't possibly understand or perceive (physical constants, relationships)? If so --- that's plausibly true, but if something doesn't interact with the Universe in a way we can measure, it may as well not exist. Or is the argument that the way we describe and interpret the Universe (language, logic) is inherently flawed because it's been developed or "discovered" based on selection for the ability to use language between extreme quantities of "language resources" we can't possibly measure? The latter is interesting, because it's perhaps a unique perspective on mental illness and drug hallucinations / delusions (schizophrenia, LSD) but still not testable.
Just musings. Neat philosophically. Would like to see something concrete here, but seems to be a prime candidate for circular reasoning - we can never test this hypothesis because we can never know what is, by definition, unknowable to us.
>Is the argument there are rules and laws of the Universe we can't possibly understand or perceive (physical constants, relationships)?
No, probably not. But instead that our filter biases make us miss most of that invisible universe. Which brings us to the point, what else are we missing? We did not know of most of the universe until we invented the scientific method to overcome the biases of our own mind.
> The latter is interesting, because it's perhaps a unique perspective on mental illness and drug hallucinations / delusions (schizophrenia, LSD) but still not testable.
>"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."
The water example showcases the germ of the idea pretty well. Stimuli don't need to be in 1-1 or even linear correspondence with physiological responses to those stimuli. Or maybe I'm misunderstanding his argument.
I've only read the linked article, not the more in-depth source mentioned there. But based on just that information, I'm deeply unimpressed by Hoffman's work. From where I stand, his idea is both poorly informed (as in, it seems like he's not made a credible effort to examine his premises) and, ironically, proven wrong by reality. That's a pretty contentious statement for a non-expert to make about a presumed expert, so I'll try to explain myself.
First, like some other commenters, I'm happy to concede that at a trivial level H. is quite correct. The world around us as perceived by the unaided human is mapped inside his brain to a vague, sometimes exaggerated, sometimes obscured and very often distorted image of reality. There are entire books on optical illusions; the trade of stage magic and various kinds of crime rely on systematic human misperceptions.
There's an obvious, perfectly good reason for this: the human system of vision is simply not a pixel-perfect 3D camera connected to petabytes of fast digital storage, and the same applies to our other senses. Given a perfect recording of the world at least in our vicinities, abundant energy and sufficient time, we could come up with highly effective survival strategies. But the real world doesn't afford us these luxuries, so evolution crafted us into organisms tuned for a reasonable approximation to an optimal compromise of this ideal. Thus, our mental model of the world is a crude abstraction, with survival-relevant information emphasized and other details brushed over. This is not a survival-optimized transformation of reality but a constraint-enforced one. A highly sophisticated system built on the shoestring budget that nature affords us. It's proven to be superior at survival to many competing models but I don't agree it works better for misrepresenting reality. Rather, it works _at all_ by necessarily sacrificing detail and accuracy in representing reality.
But none of this supports H.'s contention that we are blithely unaware of reality, or unable to apprehend it. There is a reality out there, and the depth and accuracy of our model is a function of how much time and energy we're willing to expend on mapping it. A given beach, sharply defined, has a finite and very countable set of grains of sand. If we really, really cared to know, we could build machines to count them for us. Similarly, we can or could know the shape of every coastline of every continent. Some day, humanity may have high-quality reality mappings of every planet within X light-years of our solar system. We can in principle understand the function of every gene in our genomes. We don't have to talk about how we perceive colors, because spectroscopes can tell us the exact, reproducible wavelength of every beam of light emitted by a given object. We could exchange this information with aliens having completely different bodies and brains, should we discover them, and if their science is as advanced as ours and we're careful to define our terms and measurements on observable nature, we'd have a common understanding of that reality.
But how do we know that our reality is real? How do we ascertain truth? I say we can base a pretty solid epistemology on a confluence of observed phenomena. If we encounter an obstacle we can't see through, if it's grey in color, weighs about 6 tons, stands on 4 legs, has a long nose, occasionally moves around and eats bananas by the bushel, then we can safely assume we've found an elephant. If it's a chunk of some yellow shiny solid that displaces 18 grams of water per cc, and samples drilled from arbitrary locations in it uniformly have atomic weights of X (?), melting points of Y degrees, fail to react with sulphuric acid and show a chromatographic signature consistent with that of gold, then by golly, it's a chunk of gold!
As humanity, not as individual naked humans, we've amassed a large and ever growing body of knowledge about the world around us, and (fortunately for our sanity and our ability to make sense of reality) the properties of objects and phenomena in our environment are consistent and convergent. There are no 1 gram elephants, there is no sodium that doesn't react violently with water, there are no snowflakes whose basic structure isn't hexagonal. We know that our image of reality is good because we're able to extrapolate from what we know and observe to what we haven't observed yet, to make predictions about what we'll observe and have those predictions prove mostly true.
The author and his (perhaps coincidental and unintended) idol Plantinga fail to acknowledge humanity's ability to create models of reality of whose accuracy (within limits) we can be confident because they're part of a huge network of mutually supporting sub-models with excellent predictive power. And, more importantly, that our ability to create such mappings is a human ability that we have evolved to have. A goodly part of this evolution is cultural rather than biological, and a goodly part of our senses are mechanical and external rather than built into our wetware, but our evolution and that of our apparatus is quite natural insofar as everything that we humans, natural beings in a natural world, are natural too and a part of nature.
Hoffman's conjecture is completely, utterly wrong: Evolution has in fact "bred" in humans the ability to discover reality, and this ability has incidentally given us dominion, at least in the short term and for whatever that's worth, over all other species on the planet, including our own ancestors and close cousins. Our ability to apprehend reality has made us so fit that, barring various possible disasters, we could survive the death of the Sun and Earth.
If Hoffman wants to support his claim that a creature who views too much and too little water as similar instances of "bad amounts of water" would display a higher degree of evolutionary fitness than us, I feel he has his work cut out for him.
there are some concrete examples of this though. we see the visible spectrum, a particular subset of EM spectrum that's a pretty good enough abstraction to get around in the world, it's a heuristic for the spectrum that's "good enough". we hear a subset of the audio spectrum, it's "good enough". etc etc...
meh. These are not new ideas, and probably seem rather obvious to those who think about evolution and fitness functions. I suppose however, there is some good in making this point more salient to cognitive science theoriticians however.
and why was that worth -2 moderation? I AM a cognitive neuroscientist; I certainly am not denegrating them. Cognitive scientists are used to contrasting percepts with objective reality, hence all the facinating illusions they love to demonstrate. Evolutionary theorists are used to counterintuitive ideas that result from putting fitness functions at the forefront. I would not have thought that putting the two together is a new idea.
Because not everyone had time to think deeply about every field that everyone on HN specializes in and it's sort of pointless to complain that an article about your specialty doesn't teach you personally anything new.
To be specific, I was not conplaining about the article. It is an interesting subject. I was providing context for those reading the popularized article which tends to oversell these things. but whatevs
I'd love to see any demonstration of a model of an intelligent organism with no sense of reality at all but that still functions properly. Barring any evidence, the whole article is simply "if you assume the impossible, then magical unicorn rainbows". (or else it's just extrapolating claims about the intelligence of amoebas way too literally)
> I'd love to see any demonstration of a model of an intelligent organism with no sense of reality at all but that still functions properly
That is a strawman argument. Hoffman's actual theory is that "under evolution by natural selection, 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 is just tuned to fitness."
This is taken from the more in-depth in the Quanta interview [0] linked in the submitted article.
This seems to be begging the question of whether it's possible to have a well-tuned fitness function that doesn't map in some significant way to the underlying reality.
Sure, maybe there's no point in us knowing the difference between hot and cold, as long as we avoid contact with both of them. Only wait, there is a point - if we know the difference, we can take different actions in response to each in order to respond in a more survivable manner to both. And this is indeed what we see - as well as abstracted signals (like 'pain') we have signals which measure our environment more directly (like 'hot' and 'cold'). These signals raise our fitness because they let us know not only that we should take action, but in what direction that action should be taken.
For those who have actually looked at Hoffman's work (rather than just reading a popularized account):
What happens when the fitness function is not static (as it often is in the real world)? My gut instinct is that an organism that perceives actual reality (or some close analog thereof) will be more adaptable to a changing environment than one that is optimized for static conditions.
Perhaps Hoffman has taken this into account (I don't know -- that's why I'm asking).
Which lacks predictive power. Humans have a blind spot because they can operate just fine while having a blind spot. However other organisms don't have a blind spot thus blind spots are not selected for.
The missing piece is something where getting the wrong information has value. But, that's harder than simply incorrectly interpreting information.
> Humans have a blind spot because they can operate just fine while having a blind spot. However other organisms don't have a blind spot thus blind spots are not selected for.
I'm not quite clear on your point of view, can you clarify? The phrases "blind spots" and "other organisms" are very vague. I'll try to rephrase it into something manageable. This is what I think you are saying:
Humans [cannot see behind their heads] because they can operate just fine while
[not seeing behind their heads]. However, [other species] [can see behind their
heads], thus [seeing behind one's head] is not selected for.
The "other organisms" are other species, as logically, the "other organisms" don't have "a blind spot" and therefore could not be human, which do have "a blind spot".
Do let know if this is correct, and I'll get back to you.
> The missing piece is something where getting the wrong information has value
I'm unclear where this disagrees. That's exactly the point of Hoffman's theory: an (sometimes) incorrect view of the world may have more value than the actual (correct) view of the world.
Squid for example use a different arrangement without that hole. So, while people are not getting accurate information, the gap is arbitrary and not selected for by evolution.
"his conclusion is a dramatic one: The world presented to us by our perceptions is nothing like reality"
If both he and the press are hyperbolic about the absurdity of his claims, I feel justified in saying that the claims he's making in these articles are on-their-face absurd.
Still the color argument is pretty strong. What if the energy emitted/reflected by the universe was totally in the non visible light spectrum. No one would see anything. To a blind person this is a normality, and so perhaps a person who is blind would have a better feel for what this means, but to the rest of us, color is so ingrained as "color is reality" it's hard to imagine.
Now also realize that we touch absolutely nothing. And yet we perceive solid things, as if they were solid. But again they are not. We just feel electro-magnetic force.
Following with QC, there isn't really anything such as particles.. the world is just fields of various types. Only when we go and look at the fields do we see particles, but they are just excitations in the fields.
I sometimes walk around and think what if i shut off all the perceptions that i know are only in my body/head, and thought about what was left. There really wouldn't be anything at all what we think there is.
I'm not concluding that this guy is right or wrong, but I think we have an almost impossibility in comprehending that this could actually be the case. Given that we are 90% fooled, why couldn't we be 100% fooled?
You still have to wonder why the brain created color to differentiate wavelengths. If we could not see and were speculating on how you could discern wavelengths, maybe you would guess that there would be some sort of frequency pattern that differentiated. But no... color? What is color, really? I'm not sure i even know the answer to that.
What would it be like to touch "something", rather than mere electro-magnetic forces? Isn't this just a game of words? When we touch electro-magnetic forces, we are in fact touching solid things.
Imagine you had a force field generation machine, that could produce fields in any size, shape, density, etc.
If you made a force field in a tubular shape about 8 inches long, with a smooth rubber tactile force, that was squishable, but still somewhat firm, .. would it be a banana? If you moved that force field close to your mouth would it provide sustenance (whatever that means)?
Of course you would be touching a real thing. What else could you touch?
However, on closer inspection (such as taking a bite) you would be able to tell that it wasn't a banana: the theory that the object is a banana is falsifiable. If, on the other hand, the machine can produce a simulated banana that is indistinguishable from a real one in all respects, then we can say that the object _is_ a banana.
Your point can be generalized to Descartes' evil demon, or the simulation argument. Imagine a machine that can provide arbitrary sensory inputs to the brain by simulating your body as well as the universe it inhabits.
The simulator has complete control over all reality, so that all perceptions are illusions; there is no "physical" reality, only a simulated one.
To me, this is a reduction to absurdity. If a simulated universe is indistinguishable from a real one, then we (as inhabitants of the simulated universe) should say that it is a real universe, as the simulation theory is unfalsifiable.
(If we discovered that we are in fact living inside a simulation because the simulation code was buggy and we hacked it from inside, then a Simuluation Argument Believer could just say that the "real" universe is also just another level of the simulation. This is why the _metaphysical_ Simulation Argument is unfalsifiable.)
Similarly, if a machine can produce a perfect copy of a real banana, then the copy is also actually a banana. And if a banana feels solid, then it actually is a solid object, and we aren't touching "nothing". It is irrelevant whether the feeling is produced by electromagnetic forces, or some special Platonic "solid" surface that doesn't actually exist.
Not to put too fine a point on it, but this reaction seems "tuned" rather than "reality-based". That is, one reaction to a novel theory is to consider an exaggerated caricature of the theory, because that's easier to dismiss than the theory itself. Dismissing novel theories leaves more time for foraging. b^)
Having observed the behavior of prey animals, one might find oneself more open to this theory. Prey animals like deer or turkey largely ignore other, larger prey animals like elk or moose. The presence of a single small dog, however, will set them totally on edge. It's difficult to say that a moose is in any "real" sense less dangerous than a small dog. However, it is sensible that evolutionary tuning would focus more on dog-like animals.
I don't think they are talking about having NO sense of reality. They are saying that having filters that 'simplify', or make these signals more 'user friendly' bring evolutionary benefits in comparison to experiencing reality as is. In other words, we experience an altered reality sculpted by our cognitive engine.
This reminds me of Aldous Huxley's "The Doors of Perception", by the way.
It's not saying that intelligent organisms have no sense of reality, rather that they only see as much of reality as is advantageous to their reproduction and continued existence.
Evolution does not fine tune things that closely. Sure, we can't tell which specific frequency's something reflects. That's a well known limitation. It's also kind of boring.
If we had gaps in our hearing range that may support the theory, but a more accurate assessment is detecting the world has costs. Organisms only collect as much sensory information as they need to survive. But, the same is true of running speed. To be predictive you would have to find a new and active distortion in sensing based on some prediction.
Actually yes, these things and many beyond that are knowable, yet you don't know them. Therefore you only have a sense for a subset of reality, just as you can't see the whole spectrum of light or hear every sound.
But this extreme example is not really relevant for this discussion, at least that's my humble opinion.
You also can't detect the Higgs field, or neutrinos or dark matter and energy, or gravity waves, or really vast amounts of forces and particles that are nonetheless very real. Most of 'reality' is totally invisible to us because it doesn't help us survive.
If bananas reflected yellow light, but we saw them as blue instead by some purely neuropsychological phenomenon, would this cause any significant problems?
That's actually a question of qualia and as long as you see the same frequency or collection of frequencies as the same color consistently, you can swap them around arbitrarily, but if you can't tell the difference between, for example, red and green, that is a real disadvantage.
Is that true? Yellow/Orange/Red are perceived as 'danger' colors. Is that just a made up association in our brains, or do we see yellow/orange/red more readily? If you flashed a bright orange billboard in front of a chimpanzee, would you get a different reaction than if you flashed one that was dark blue?
Orange and dark blue are names we give to perceived frequencies of light.
Dark blue isn't really any 'darker' in terms of energy content spectrum than orange is, its our brains that have evolved to make this association that X spectrum is dark and Y spectrum is bright.
The conclusion seems pretty obvious; we only see light in a certain (relevant) spectrum, we only taste certain (relevant) chemicals, we only hear sounds in certain (relevant) frequencies, etc., because it doesn't help us (and in fact distracts us) to perceive more information.
Also, it seems like a big aspect he's missing is that within each relevant spectra, we've gotten pretty well tuned to reality. And that's simply because oftentimes, fitness is highly correlated with perceiving reality. For example, seeing all physical objects realistically is important for survival, since if you can't see perceive certain physical objects, then predators/prey may take advantage of the failure (e.g., zebras blending in with tall grass).
Is there some greater point I'm missing?