Join 3,553 readers in helping fund MetaFilter (Hide)


Time : a flat circle :: Consciousness : a state of matter?
June 12, 2014 7:14 AM   Subscribe

"While the problem of consciousness is far from being solved, it is finally being formulated mathematically as a set of problems that researchers can understand, explore and discuss.

Today, Max Tegmark, a theoretical physicist at the Massachusetts Institute of Technology in Cambridge, sets out the fundamental problems that this new way of thinking raises. He shows how these problems can be formulated in terms of quantum mechanics and information theory. And he explains how thinking about consciousness in this way leads to precise questions about the nature of reality that the scientific process of experiment might help to tease apart.

Tegmark’s approach is to think of consciousness as a state of matter, like a solid, a liquid or a gas. 'I conjecture that consciousness can be understood as yet another state of matter. Just as there are many types of liquids, there are many types of consciousness,' he says."
posted by Strange Interlude (235 comments total) 48 users marked this as a favorite

 
It's quantum!
posted by Zarkonnen at 7:15 AM on June 12 [4 favorites]


Oh man this is exciting! Thanks, Strange Interlude!
posted by mondo dentro at 7:21 AM on June 12


It's dust. It flows in from space and settles on adults.
posted by bleep at 7:26 AM on June 12 [19 favorites]


This is wonderfully daft
posted by fallingbadgers at 7:31 AM on June 12 [8 favorites]


Philosophers hate physicists for this one weird trick...

I'm sure that philosophers will be happy to discover that "researchers" can now "understand, explore and discuss" problems about consciousness... Whew! It's about time!

tl;dr: consciousness-ons explain consciousness.
posted by Fists O'Fury at 7:31 AM on June 12 [12 favorites]


To a man with a particle accelerator, every problem is an electron.
posted by St. Peepsburg at 7:36 AM on June 12 [37 favorites]


Time is a flat circle. Otherwise it'd be a sphere.
posted by Beardman at 7:38 AM on June 12


I was at a live taping of Startalk and someone asked Neil Degrasse Tyson about whether consciousness was related to quantum mechanics. He shot the guy down in spectacular fashion by turning to the audience and saying that because quantum mechanics seems "magical", people love to use it as an explanation for everything they don't understand. While strange things are happening at the particle scale, at the usual matter scale, things are pretty stable. You exist in one place in space and time and that's that. 99% of the time, if someone suggests quantum mechanics as the reason for something happening on a scale larger than an atomic particle, it's a huge red flag.
posted by the jam at 7:40 AM on June 12 [22 favorites]


consciousness may just be an emergent property of a sufficiently complex electrical system. the internet may have achieved consciousness, but not skynet-style self-awareness yet, nor has it formulated an agenda for itself distinct from our own - when it does, look out.
posted by bruce at 7:41 AM on June 12 [4 favorites]


As far as anyone knows for sure, "consciousness" is "exhibited" only by a very specific configuration of matter.
posted by Jode at 7:44 AM on June 12 [4 favorites]


That's such an amateur mistake. Everyone knows time is a cube.
posted by kaibutsu at 7:48 AM on June 12 [31 favorites]


I'm not really even sure where to begin with this, but we might get some more mileage out of discussing the arxiv posting, which is not obviously as insane as the medium link. (Actually, I'm not sure that's really true. It just has a lot more math.) Even within disciplines that focus on aspects the mind/brain, the general feeling is that researchers who say they are working on consciousness are trying to sell you something, or actually working on something else but have chosen to use 'consciousness' as the big picture spin (and maybe still trying to sell you something). At best I suspect this falls into the latter category.

Just a few representative quote from the medium link... "For as long as the discipline has existed, physicists have been reluctant to discuss consciousness, considering it a topic for quacks and charlatans. ... Interestingly, the new approach to consciousness has come from outside the physics community, principally from neuroscientists..." ..skipping a lot.. "This leaves us with an integration paradox: why does the information content of our conscious experience appear to be vastly larger than 37 bits?” asks Tegmark. ... That’s a question that many scientists might end up pondering in detail. For Tegmark, this paradox suggests that his mathematical formulation of consciousness is missing a vital ingredient."

Thank you, no, I will not ponder this question.
posted by advil at 7:49 AM on June 12 [9 favorites]


Fingers still crossed for "Consciousness as illusion, and all our frantic thrashings in search of meaning as existential farce."
posted by Pirate-Bartender-Zombie-Monkey at 7:49 AM on June 12 [7 favorites]


"Consciousness as illusion, and all our frantic thrashings in search of meaning as existential farce."

This could only be stated by something that is conscious.
posted by Liquidwolf at 7:52 AM on June 12


My conciousness says this is ridiculous on many levels.
posted by blue_beetle at 7:52 AM on June 12 [2 favorites]


It isn't consciousness that is an illusion; it is unconstrained agency.
posted by grumpybear69 at 7:54 AM on June 12 [2 favorites]


"Doctor... we have this pet physicist and ...something's wrong. It keeps babbling about linguistics and neurology and climate science."

"He's a little old, isn't he?"

posted by leotrotsky at 7:57 AM on June 12 [19 favorites]


Photosynthesis and semiconductors are examples of macro scale quantum effects. Just saying.

That said, epistemology has first dibs on this stuff, physics can't just call cuts.
posted by Slap*Happy at 7:57 AM on June 12 [11 favorites]


Well, I think the main problem is that you can define *something* pretty exactly and then study its effects. But whether that is what people really mean when they say 'consciousness' is far from clear - and that's still the realm of philosophy.

That's not to say this course of study of information may not be useful and interesting, but whether it actually relates to consciousness is an unsettleable question until we know a lot more.
posted by Zalzidrax at 8:02 AM on June 12 [2 favorites]


The "computronium"/"perceptronium" thing is kind of part medieval and part comic book, like a Lex Luthor or Ozymandias project harnessing the phlogiston...

I feel like I should repost Eric Schwitzgebel's truly wonderful argument that If materialism is true, the United States is probably Conscious.
posted by batfish at 8:03 AM on June 12 [6 favorites]


Scott Aaronson had had a pretty good back & forth with Tononi and you can even find David Chalmers jumping into the blog comments
posted by crayz at 8:05 AM on June 12 [10 favorites]


Maybe I got out on the wrong side of the bed, but the process of defining something we don't understand into something we do understand, and then studying that definition, can lead to some pretty circular reasoning.

Don't get me wrong, there's a lot of value in working through metaphor, but I just can't see this as any sort of breakthrough... yet.
posted by elwoodwiles at 8:11 AM on June 12 [3 favorites]


I'm having a really hard time following this article's train of thought.
posted by Greg Nog at 8:12 AM on June 12


In a similar vein, how computers work:

Computers are based on logic gates. Using relays or transistors you can implement simple logical functions such as AND, OR, and NOT. Using two such gates you can implement a flip-flop, which is a storage element capable of remembering whether its state is zero or one. When you hook enough of these gates and flip-flops up in a certain manner, you get a computer.

Computers are not, however, a state of matter. Extensive research has revealed that they are created by forging a mixture of normal matter and smoke. As long as the smoke stays bound to the normal matter, the computer continues to function.
posted by localroger at 8:17 AM on June 12 [12 favorites]


Consciousness- something easily cured with alcohol.
posted by jonmc at 8:19 AM on June 12 [3 favorites]


1. Consciousness is everything we knowingly experience.
2. If we wish to "understand" consciousness, we must assume a subject who is outside of consciousness and able to observe it.
3. Good luck, scientists and philosophers!
posted by haricotvert at 8:21 AM on June 12 [1 favorite]


2. If we wish to "understand" consciousness, we must assume a subject who is outside of consciousness and able to observe it.

This is why spedometers only work outside of cars!
posted by mittens at 8:24 AM on June 12 [8 favorites]


Ah, but he's ignoring all the previous literature on the subject. This is all fine and well, Mr. Tegmark, but how does your theory account for the 4 corner simultaneous 24 hour days that occur within a single 4 corner rotation of Earth.?
posted by Mayor West at 8:26 AM on June 12 [1 favorite]


"why do conscious observers like us perceive the particular Hilbert space factorization corresponding to classical space (rather than Fourier space, say)"

Uh, maybe because classical 3d space admits much simpler models of the world that work just fine for all known living beings except the ones who need to kill each other with lasers and nuclear weapons? Why would a cow need to evolve direct perception of solid bodies in terms of Fourier modes (assuming that such a thing would be possible)?

OTOH, aren't color vision and hearing more or less perceptions of Fourier modes?

tl;dr: "Tegmark does not have an answer."
posted by mubba at 8:26 AM on June 12 [4 favorites]


So, to understand life we need to assume a subject that is outside of life, too?

By, jove, we've got a new proof for God! (or not).
posted by oddman at 8:27 AM on June 12


If we wish to "understand" consciousness, we must assume a subject who is outside of consciousness and able to observe it.

Or we could assume that everything is conscious, and then pursue the understanding of consciousness by comparing the different forms that it takes.

According to some, Tegmark is in fact proposing a kind of panpsychism:
Regarding the ‘hard problem’, Koch, Tononi and their physicist colleague Max Tegmark have embraced a form of panpsychism in which consciousness is a property of matter. Simple particles are conscious in a simple way, whereas such particles, when integrated in complex computation, become fully conscious (the ‘combination problem’ in panpsychism philosophy). Tegmark has termed conscious matter ‘perceptronium’, and his alliance with Koch and Tononi is Crick’s legacy and a major force in the present-day science of consciousness. Their view of neurons as fundamental units whose complex synaptic interactions account for consciousness, also supports widely-publicized, and well-funded ‘connectome’ and ‘brain mapping’ projects hoping to capture brain function in neuronal network architecture.--"‘Collision Course’ in the Science of Consciousness: Grand theories to clash at Tucson conference"
The mistake that Tegmark makes is to construe consciousness as a state of matter. It is more accurate to say that matter is a means by which consciousness construes itself.
posted by No Robots at 8:29 AM on June 12 [1 favorite]


Fingers still crossed for "Consciousness as illusion, and all our frantic thrashings in search of meaning as existential farce."

If consciousness is an illusion what is being deluded?

Or, to put it more succinctly, cogito ergo sum.
posted by vorpal bunny at 8:30 AM on June 12 [1 favorite]


So, to understand life we need to assume a subject that is outside of life, too?


To understand life as an object of consciousness, you would need a subject aware of that object, yes.
Unless you've got another way of doing it!
posted by haricotvert at 8:31 AM on June 12


This is why spedometers only work outside of cars!


This is actually a surprisingly common error. A speedometer is not an observer. It's an object that an observer can use as an instrument for measurement. The speedometer and the car would both be objects of consciousness. If the speedometer somehow became conscious, and then attempted to figure out what consciousness was on an objective level, the poor little speedometer would have the same problem as anyone else. But it would be adorable!
posted by haricotvert at 8:35 AM on June 12 [6 favorites]


(ctrl-F banana)

(okay then)

Time flies like an arrow.
Fruit flies like a banana.
posted by Curious Artificer at 8:37 AM on June 12 [3 favorites]


The more I think about the problem of consciousness, the more I think dualism is unavoidable, ontologically.

I was just at a conference where a scientist asked during a presentation on 'quantum dot' display technology, "Do you know what color is? How would you explain it to a 12 year old?" He mentioned there are some great educational videos, one of which did just that - explaining, "our brains have evolved to use light wavelength to interpret our surroundings to help us find food or whatever." He seemed very satisfied the explanation, though later seemed to say again that actually we don't know what color is.

I find it really sad that we teach our children how to explain away really cool problems we don't understand. And the mind-body problem is way more interesting than just so evolutionary stories anyway.
posted by Golden Eternity at 8:41 AM on June 12 [3 favorites]


I think of consciousness not as an attribute or a state of being, but as an action.
posted by rebent at 8:41 AM on June 12 [3 favorites]


okay, it goes like this.

Individual consciousness in all its ever evolving multiplicity is best understood as a tiny seed which grows into a great and infinitely complex tree (or perhaps vine) that spreads its branches, leaves, tendrils (whatever) off and out through the ever expanding reaches of the universe, and like a tree (or perhaps vine) there's all manner of twisting and turning, curving and convoluting. Except in fact, it's all actually happening the other way, from the outside in -- we're not the seed at all, we're the furthest tendrils.

At least that's how it was explained to me one night by a friend after a long and particularly BIG (for him) acid trip.
posted by philip-random at 8:43 AM on June 12 [4 favorites]


If we wish to "understand" consciousness, we must assume a subject who is outside of consciousness and able to observe it.

Or we could assume that everything is conscious, and then pursue the understanding of consciousness by comparing the different forms that it takes.


Leaving aside the question of whether inanimate objects have consciousness, there is clearly some level of consciousness in all living organism. "Level" is an important part of that, but there has not been much work on defining those levels, or figuring out how some organisms have "more" or "deeper" consciousness than others. My feeling is that consciousness results from a feedback mechanism, or a whole set of layered feedback mechanisms (which are designed to monitor and correct a variety of functions, many of which are biomechanically understood), the accumulation of which in "higher" organisms results in a consciousness that becomes increasingly self-aware. The error-correcting mechanisms described in the article sound similar to this. In other words, your conscious experience results from the coordinated effect of a huge number of error-correcting mechanism occurring in your body (and not just in your brain). Anyway, that's my theory.
posted by beagle at 8:45 AM on June 12


I find it really sad that we teach our children how to explain away really cool problems we don't understand. And the mind-body problem is way more interesting than just so evolutionary stories anyway.

There's nothing particularly wrong with a just so evolutionary explanation for anything as long as the teller doesn't get his back up when the proverbial twelve year old counters it with a good old-fashioned "but what if".
posted by philip-random at 8:46 AM on June 12 [1 favorite]


The "computronium"/"perceptronium" thing is kind of part medieval and part comic book, like a Lex Luthor or Ozymandias project harnessing the phlogiston...

Everything old is new again. William James summarizing (before demolishing) the theory of "mind dust":

If evolution is to work smoothly, consciousness in some shape must have been present at the very origin of things. Accordingly we find that the more clear-sighted evolutionary philosophers are beginning to posit it there. Each atom of the nebula, they suppose, must have had an aboriginal atom of consciousness linked with it; and, just as the material atoms have formed bodies and brains by massing themselves together, so the mental atoms, by an analogous process of aggregation, have fused into those larger consciousnesses which we know in ourselves and suppose to exist in our fellow-animals. Some such doctrine of atomistic hylozoism as this is an indispensable part of a thorough-going philosophy of evolution. According to it there must be an infinite number of degrees of conscious- [p.150] ness, following the degrees of complication and aggregation of the primordial mind-dust.
posted by shivohum at 8:46 AM on June 12 [5 favorites]


Consciousness clearly isn't a state of matter like how liquids or solids are. Our brains are composed of liquids and solids, and exceedingly complex arrangements of those don't imply new matter-states any more than a computer chip could be regarded as a new matter-state of silicon.

My impression is that the hard problem is figuring out how increasing neural complexity yields subjectivity. I don't see how this conception would help explain that.
posted by clockzero at 8:48 AM on June 12 [2 favorites]


golden eternity, of course we know what color is. the physicist in your comment dropped the ball on your foot and it rolled out of bounds. whether individuals perceive color the same or differently is an entirely different question.
posted by bruce at 8:56 AM on June 12


Tegmark seems to like to posit ontological status to things in weird ways. His last book states that math is real. Real like rocks. Why? As it "explains" how things happen it must be what is happening. When he was on NPR awhile back I called in and asked him if this was just confusing the map for the territory. I didn't really get an answer. It seems to me that some scientists need a serious course in both the history and content of philosophy. It might save us from flashy stuff like this.
posted by njohnson23 at 9:02 AM on June 12 [5 favorites]


Tegmark previously on MetaFilter.

Does anyone actually take Max Tegmark seriously? I mean, I guess we can assume he's actually qualified to talk about physics, though after reading enough of his stuff one does begin to have one's doubts; but he's such an obvious crank when it comes to philosophy that his stuff only seems to get traction on account of the cultural prestige of SCIENCE.
posted by RogerB at 9:10 AM on June 12 [5 favorites]


We know what colour is, for almost all uses of the word. The bit we don't understand isn't bound to colour per se, it's how we perceive perceptions - and even here, a lot of the mechanisms are being mapped out The final switch into a conscious impression is tied to what consciousness is, that's still mysterious.

Having spent a lot of my aware life wondering about this, I'm increasingly of the opinion that consciousness is largely illusion - or rather, we're running a model of the world and ourselves on a very capable machine, and that model includes aspects of itself.

This is simply what it is to be that model.

We exalt consciousness because it seems very separate to anything else we can observe; our own is very intimate, and other people's is very inaccessible, but expecting to find something very special within it that justifies our exalted expectations may be as otiose as vitalism.

Like peeling away an onion, you'll never find some special onion-ness at the heart of it all. When we've mapped the modelling system, and then when we've solved the stimulus->model element mapping function, we may not have consciousnessinos floating around or a set of simple equations, but we'll understand what's going on enough to create analogues and perhaps get new ways to manipulate our own experiences and capabilities that come with the train set.

And there may be nothing else to it than that.

Along the way, there's plenty to find out - and who's going to rule out macro-quantum systems akin to the photosynthetic energy pathway optimisation? Not I. But that's a very long way from conflating consciousness with quanto-woo, and not far from arse-about-facedness.

In the end, there may be no there there. Perhaps I've done too much acid. Perhaps, not enough.
posted by Devonian at 9:40 AM on June 12 [15 favorites]


Oh, yes, please, let's dress up our woo as physics. That makes it so much more plausible(-sounding).

Here's the thing I keep getting stopped at with all this bullshit: It's basically just religion is scientistic clothing. There is absolutely no reason to suppose that any of this might be true. It's a very attractive-sounding set of ideas, but it has no more basis in evidence than any other theology.

Could we please stop pretending that the word 'consciousness' actually means something in the context of quantum mechanics?
posted by lodurr at 9:40 AM on June 12 [5 favorites]


My impression is that the hard problem is figuring out how increasing neural complexity yields subjectivity.

I've never understood why that's perceived as a hard problem. To me it's always seemed more like a non-problem. We only care about it because we think we're conscious. We have no way to know when consciousness emerges, if it even makes sense to say that it does, but we have increasingly strong evidence that there are a lot of non-human animals that do stuff we'd previously reserved as falling within the territory of "consciousness." Which suggest to me in turn that whatever it is that makes us 'special' is probably not all that special.
posted by lodurr at 9:44 AM on June 12


I'm very prepared to believe this is all nuts, but I also don't get the sense that any of the glib dismissals in this thread are based on any kind of good-faith effort to work through his actual arguments. This all feels like the equivalent of "there's this nutjob called Einstein who says that energy and matter are like the same thing--what a maroon!" Not to say that Tegmark is Einstein, of course, but "haw haw, this sounds silly!" isn't much of an engagement with an unfamiliar theory.
posted by yoink at 9:45 AM on June 12 [7 favorites]


Unless his argument starts with some reason to believe that there's a connection, I'm not interested in hearing it. All the expositions I've ever heard of quantum consciousness theories basically start by assuming the consequent. This is really no different. It's theology, not physics. It's barely even math.
posted by lodurr at 9:49 AM on June 12 [1 favorite]


METAFILTER: Perhaps I've done too much acid. Perhaps, not enough.
posted by philip-random at 9:54 AM on June 12 [8 favorites]


Unless his argument starts with some reason to believe that there's a connection

Er, what? "Unless the argument starts by showing me that its conclusions are sound, I won't listen to the argument for why its conclusions are sound"? I think you might not have quite figured out how this whole "argument" thing works.
posted by yoink at 10:00 AM on June 12 [1 favorite]


"A photon can behave as a wave and a particle, because two separate experiments demonstrate both aspects. Here are the details..." - that's a perfectly good way to start an argument, yoink, especially in physics. "Some reason" does not equate to "prove the conclusion".
posted by Devonian at 10:09 AM on June 12 [1 favorite]


Has anyone here read the arxiv paper? (I can't get on medium at work so I can't comment on their take on it.) This sure as hell isn't my field, half the definitions were new to me and I only got about a third of the way through. However, the paper seemed to be addressing the question 'If we assume that consciousness is a thing that matter does sometimes, what properties does the matter need to have to do it'. Which seems reasonable and fairly interesting, and sidesteps any of the ontological woo that people are accusing him of. For example, the most accessible conclusion he comes up with is that any object that produces consciousness needs to be more ordered than a liquid but cannot be as ordered as a solid, which sounds fine to me.
posted by Ned G at 10:14 AM on June 12 [2 favorites]


We only care about it because we think we're conscious.

..the glib dismissals in this thread are based on any kind of good-faith effort to work through his actual arguments.


Many of the arguments against consciousness are against its existence. Why do we need to listen to Tegmark's or Krick's argument for unicorns?

I don't understand why there is so much confusion about the terms, 'consciousness,' 'perception,' 'color.' When I look at a sunset, or whatever, I have the conscious experience of seeing or perceiving color. When someone says, "no actually you just think you do ... it is actually an illusion ... in reality you are not seeing, you are not conscious," I have no idea what they are talking about. It's fascinating how we can see the same problem so differently.

We understand a lot about the mapping between light wavelength and color experience, but we don't have a scientific, materialist explanation for the experience itself, imo. As such, we don't know what color is.
posted by Golden Eternity at 10:18 AM on June 12


njohnson23: "Tegmark seems to like to posit ontological status to things in weird ways. His last book states that math is real. Real like rocks. Why? As it "explains" how things happen it must be what is happening. When he was on NPR awhile back I called in and asked him if this was just confusing the map for the territory. I didn't really get an answer. It seems to me that some scientists need a serious course in both the history and content of philosophy. It might save us from flashy stuff like this."

This is something I've mentioned on the blue before... I don't think it's a problem to ask these sorts of questions. I mean, if you're going to be stuck into the materialist/reductionist mode of thought, then sure, abandon any such queries.

Of course, one could argue that science *has* to be materialist, and any sort of metaphsical discussion such as this from a scientist is absurd, but is it? (Don't get me wrong, I consider myself primarily a materialist; generally systems theory, emergent behavior. I personally have a feeling that Gravity is an emergent phenomenon, and I don't see why that isn't taken more seriously as a potential explanation -- but then you get into Ontological questions... and we're right back where we started)...

But I always call this the "Prescriptive" vs "Descriptive" science (or math) issue.

Is math, as a language, merely descriptive of the universe around us? Is it an emergent phenomenon from Consciousness (which is interesting, considering we're discussing the idea of consciousness as emergent from matter)?
Or is it a descriptive language, a Logos, a Platonic sort of ideal that gives shape and form to the substance/Ousia.

Metaphysically it sounds a bit too... metaphysical. Too ... Godlike. Hence a certain apprehension amongst the more strong materialists.

Of course, further - is Math SEPARATE from Substance? Is Substance (Space, Time, Energy, Matter) emergent from Math itself? Does it exist separately formless and void and given meaning via math casting itself into material?

Is it a feedback loop? Does math "evolve" over time as it embeds itself into the material structure of existence, to idealize itself via consciousness and somehow perpetuate itself in a higher form... a sort of Aeonization of existence in some strange demiurgic manner?

At some point, this does get a bit ridiculous, yes. But I don't think we should cast all questioning aside on this issue if it fulfills two criteria:

1) Logical Consistency
2) Falsifiability

Of course that last one is what is going to make this a scientific question vs a philosophical question, and I leave that those who devise the theories. I'll just sit back and enjoy the holographic projection from 3d surface to 4d spacetime.
posted by symbioid at 10:26 AM on June 12 [2 favorites]


"This leaves us with an integration paradox: why does the information content of our conscious experience appear to be vastly larger than 37 bits?” asks Tegmark. ... For Tegmark, this paradox suggests that his mathematical formulation of consciousness is missing a vital ingredient.

That's putting it mildly.

This morning I, personally was experiencing consciousness at, like, over 1,000 bits, easily. And that was before coffee.
posted by haricotvert at 10:29 AM on June 12


> "A photon can behave as a wave and a particle, because two separate experiments demonstrate both aspects. Here are the details..."

But that's not how it all started. If I remember correctly the first experiments didn't occur until one or two centuries after the theory was proposed.

This is certainly not my wheelhouse, but people seem overly eager to misinterpret or dismiss this argument. For example, the article credits neuroscience--not Tegmark--with the idea of consciousness being quantifiable matter. Tegmark admittedly stretched out the concept far too early, but it seems like a reasonable foundation for an argument.

At the very least, it's interesting food for thought. Like the argument that love is made of dark matter. It may ultimately be silly, but because our relevant knowledge is so extremely lacking (and on preview, what symbioid mentions about falsifiability), I'm inclined to say "that's, like, your opinion" in debates like these.
posted by Johann Georg Faust at 10:37 AM on June 12


This is simply what it is to be that model.

I think so too.
posted by flabdablet at 10:44 AM on June 12


It's not that you can't start an argument with no reason to believe it, but that it affects its status as something to be taken seriously. Especially when it's in an area marked by a lot of arguments that haven't stood up to scrutiny. It's a matter of probabilities. Atomism has been around for millennia, and was remarkably prescient, but it had nothing to mark itself out as being particularly (ho ho) special and it was a rational and sensible choice to dismiss it until more evidence turned up. Rational, sensible and wrong... but we're faced with very many daft theories and it's best to concentrate on those that seem to have a chance of being useful.

And just because the perception of consciousness itself as something special and distinct requiring special and distinct basic mechanisms may be an illusion, does not mean that consciousness is not a fact. Illusions themselves are facts, and very often worth investigating - you learn a lot about the visual system from optical illusions. What they appear to depict is not what they are, but they do have existence. I like thinking about God and find such thoughts useful in my life, and that God exists as a concept is a fact; none of this is contradicted by my conclusion that there's nothing outside our skulls to which the concept strongly corresponds.
posted by Devonian at 10:49 AM on June 12


Essentially, all models are wrong, but some are useful.
-George E. P. Box
posted by klarck at 10:49 AM on June 12 [2 favorites]


In 2008, Tononi proposed that a system demonstrating consciousness must have two specific traits. First, the system must be able to store and process large amounts of information. In other words consciousness is essentially a phenomenon of information.

So far so good.

And second, this information must be integrated in a unified whole so that it is impossible to divide into independent parts. That reflects the experience that each instance of consciousness is a unified whole that cannot be decomposed into separate components.

Definitely not enough experience with psychedelics.
posted by flabdablet at 10:50 AM on June 12


Just so we're clear on my intent, I posted this mostly to get MeFi's read on the basic concepts, since I have just enough of a background in quantum physics to make jokes about cats in boxes and understand Robert Anton Wilson novels, and I was wary of the inherent potential for metaphysical woo here. But I thought that Tegmark's ideas were laid out in a lucid enough fashion in the original paper that I was curious to see how the MeFi hivemind's bullshit detector responded, because I wasn't getting the Time Cube crackpot vibes I'd expected when I saw the article title.
posted by Strange Interlude at 11:10 AM on June 12


Tegmark is a right proper physicist, so I take his ponderings with a bit more respect than I would a lot of people who use the word "quantum". That doesn't mean he's correct, but it does mean he at least knows what the fuck he's talking about when it comes to the term "quantum" vs a lot of the new-age woo-hucksters.

I have a feeling that using the term "Quantum" in conjunction with "Consciousness" has turned off some people in this thread, assuming he's just another one of "those people". I could be wrong. But I hope that they don't just discount his thesis because he joins the two concepts.

But that's me. That's how I roll. There certainly deserves to be a lot of caution when someone tries to claim a source of consciousness or an understanding that seems to step outside the current scientific understanding. I don't know if his work integrates the current understanding and builds upon it (i.e. neuronal patterns, etc...) I watched him talk about this on Youtube once (there's a lot of good videos of him up there if you ever want to watch him... I like the guy, personally)... I don't recall the details. Anyways, if it's "throw out the baby with the bathwater" then that takes a harder pill to swallow than if it just builds upon our understanding and adds in something else to explain it.
posted by symbioid at 11:24 AM on June 12


I would never argue that consciousness doesn't exist. But I would argue that there's no evidence it has an essence.
posted by lodurr at 11:56 AM on June 12 [2 favorites]


"Unless the argument starts by showing me that its conclusions are sound, I won't listen to the argument for why its conclusions are sound" isn't even vaguely close in meaning to "Unless his argument starts with some reason to believe that there's a connection". So please feel free to try again, yoink.
posted by lodurr at 11:59 AM on June 12


To me, trying to explain consciousness is like trying to touch your finger with your finger, or directly see your eye with your eye. You're trying to make awareness itself into an object of awareness -- consciousness into an object of consciousness. But that's impossible. You can be conscious of a model that represents consciousness, as Klarck points out. But not the thing itself. And models have only practical, not absolute, value. Even if you were clever enough to cook up a model capable of creating a conscious computer, you would have advanced not one iota your understanding of what it's like to BE that computer.
posted by haricotvert at 12:30 PM on June 12 [1 favorite]


I would never argue that consciousness doesn't exist. But I would argue that there's no evidence it has an essence.

If we allow for forms of consciousness other than human, then we do have to posit principles of distinction, ie. distinct essences.
posted by No Robots at 1:35 PM on June 12


Why would the 'principles of distinction' imply essences? Do degrees imply essences? Specific degrees, or any degrees? Are there an infinite variety of essences to go along with the infinite degrees of difference?
posted by lodurr at 1:38 PM on June 12 [1 favorite]


Well, look, hydrogen and helium can be said to be degrees on a continuum, but for practical purposes we assert that they are distinct from each other and all other elements. Whatever it is that makes them distinct is their essence.
posted by No Robots at 1:42 PM on June 12


Does life have an "essence" or does it just exist without an essence as well?
posted by Golden Eternity at 2:20 PM on June 12


Life is the general principle. It is esse, being itself. Things have their particular life form as their essence.
posted by No Robots at 2:30 PM on June 12


> Well, look, hydrogen and helium can be said to be degrees on a continuum, but for practical purposes we assert that they are distinct from each other and all other elements. Whatever it is that makes them distinct is their essence.

Isn't that the argument? That these are individual elements, but collectively calling them "atoms" or "matter" is not accurate enough? They are collectively "the universe," and although atoms are the most basic unit of matter, they're not the most fundamental particles in the universe. The elements have a commonality of electron configurations; people have a commonality of displaying consciousness. A distinct essence may not imply a distinct consciousness, rather a configuration or small representation of the whole of consciousness, instead.

The major problem seems to be agreeing on what the scope of consciousness should be, and I'm not even going to try to address that because we've got people and Portuguese man o' wars and all sorts of weird things on this planet.
posted by Johann Georg Faust at 2:56 PM on June 12


A distinct essence may not imply a distinct consciousness, rather a configuration or small representation of the whole of consciousness, instead

Each generic form of consciousness is indeed a particular expression of consciousness as a whole. Human consciousness is one expression of the whole of reality, horse consciousness is another.
posted by No Robots at 3:09 PM on June 12


Not to say that Tegmark is Einstein, of course, but "haw haw, this sounds silly!" isn't much of an engagement with an unfamiliar theory.

"haw haw, silly!" may not be much of an engagement, but it is for sure a true description of the theory as explained by the medium thing, and independent of the theory's being completely bonkers. That is, even if the theory is true, "computronium" and "perceptronium" are objectively silly ontological categories.

But I think some red flags about the theory that are evident just at a glance, are that it doesn't seem very aware of the philosophical literature on consciousness, or even clearly state the "hard problem." For starters, the theory might say something like: I am not emergentism/dual-aspect-theory/panpsychism/xyz because...or: I am one of those things, but it's all ok because... or even: orthodox ways of formulating this problem and the categories embedded therein are all completely forlorn because... but totally ignoring the standard scheme and skipping straight to the computronium is a total red flag. Maybe Tegmark doesn't actually do that, but it sure looks like it here...

Furthermore, I tend to think the "hard problem" formulation is responsible for a huge amount of confusing and confused literature on consciousness, within, to be sure, but coming from outside of the philosophy racket especially, and this just totally looks and quacks like that stuff. The thing is, the "hard problem" used to be called the "mind-body" problem back before computation and cognitive science and all that stuff got going and nobody could imagine how physical stuff could be that kind of stuff. When it started to make sense that it could, the worry became about how it could be like anything to be that kind of stuff even if it did compute and whatnot. That problem, i.e. the "hard problem," is the same problem as the mind-body problem, just less of it. Properly, the "hard problem" is "what's left of the mind-body problem," and I think actually calling it by that phrase would do a better job of (a) keeping track of the actual problem and it's history, and (b) giving dazzlingly heterodox outsider theorists a better idea of the kind of thing their explanations needed to explain.
posted by batfish at 4:43 PM on June 12 [1 favorite]


Does life have an "essence" or does it just exist without an essence as well?

Sure, why not?

Or, put another way: What is this "essence" of which you speak?
posted by lodurr at 5:45 PM on June 12


For example, the most accessible conclusion he comes up with is that any object that produces consciousness needs to be more ordered than a liquid but cannot be as ordered as a solid, which sounds fine to me.
(Ned G)
…which seems plausible since most of us are more ordered than a liquid and less ordered than a solid. (in fact: Metafilter: more ordered than a liquid but [not] as ordered as a solid)

it seems that from this you might be tempted to draw the conclusion: Therefore, a computer made of solid matter can never adequately simulate consciousness (AKA meat is special). But to draw that conclusion we need an additional premise: a computer made of solid matter can never adequately simulate liquid (or any other state of matter but its own). where by "adequately" I suppose I mean "to arbitrary fidelity" or similar. But that way lies either "computers cannot adequately simulate math" or "math cannot adequately simulate states of matter", either of which is bad for the project of physics.

To me, it seems like that additional premise is probably false. We regularly simulate physical systems at one level and observe the same classes of emergent properties in the simulation as in the physical systems. When we refine the simulations, we see more of the emergent properties we see in the physical systems. This position doesn't save us from the (unsavory to some) consequences of materialism while avoiding the overt introduction of dualism.

And for what it's worth I do take seriously the argument that the US itself is probably conscious, though I haven't noticed any change in my behavior that I attribute to this belief. On the other hand, some things we regularly do we would do as a consequence of adopting this belief: for instance, if we want to model what will happen when a member of the police pulls us over for speeding, we mostly do that by reference to laws or "what policemen do", and not by reference to the properties of the individual officer. In the same way we consider the person (and not the fist) when someone punches us, this makes sense if we consider an individual police officer to be a like limb of the police department, or a fingertip of the local government. It might be for other reasons that we do this (examined or unexamined reasons) but "governments are in some sense conscious, and act out their will by their human parts" also leads us to the same thing.
posted by jepler at 7:13 PM on June 12 [1 favorite]


Not to say that Tegmark is Einstein, of course, but "haw haw, this sounds silly!" isn't much of an engagement with an unfamiliar theory.

Yeah, Tegmark is blatantly, almost willfully unaware of the philosophy debates raging around mathematics and epistemology. He's pretty much at the "assume a spherical cow" point with this stuff, even before dipping a toe into neuroscience or more abstract philosophies of being.
posted by Slap*Happy at 7:33 PM on June 12 [1 favorite]


Tononi's theory starts off with objectionable stipulations, but fortunately does make falsifiable predictions about memory. (And IIRC recent data from unrelated memory research provide falsification.)

Tegmark's stuff, I can't fathom how they plan to test it critically.
posted by owhydididoit at 11:11 PM on June 12


I've never understood why that's perceived as a hard problem. To me it's always seemed more like a non-problem. We only care about it because we think we're conscious. We have no way to know when consciousness emerges, if it even makes sense to say that it does...

I would never argue that consciousness doesn't exist. But I would argue that there's no evidence it has an essence.

What is this "essence" of which you speak?


Wait. You are the one that brought up essence. I'm not sure it is a helpful concept myself, and am not well read on its use in philosophy. Perhaps you can explain more about how it pertains to the discussion.

Okay, so consciousness exists. We only care about it because we think we are conscious. Are we wrong to? What does any of this have to do with the hard problem being a non-problem? It seems that what would suffice to eliminate the hard problem is a solution for it. And why do you think we have "no way" to know when consciousness emerges? We have plenty of ways to in psychology, psychiatry, and neurology. Of course these sciences are not perfect and often rely on the reports of a large sample of subjects, or on brain lesion studies, or on animal studies of anesthetics, etc.

When it started to make sense that it could, the worry became about how it could be like anything to be that kind of stuff even if it did compute and whatnot.

I'm shocked so many people think they have imagined how computers are conscious. What computation is the conscious experience of the color red? The smell of sulfer? A musical note? Pain? etc. With all the computing power we have, if consciousness is computational, it seems to me we should have some sort of answers for the computations that suffice for simple experiences like those by now.

But I thought that Tegmark's ideas were laid out in a lucid enough fashion in the original paper that I was curious to see how the MeFi hivemind's bullshit detector responded

I'm probably the least qualified here to answer this but I'll volunteer my Dunning-Kruger thoughts on it:
Does the quantum wavefunction undergo a non-unitary collapse when an observation is made, or are there Everettian parallel universes? Does the non-observability of spacetime regions beyond horizons imply that they in some sense do not exist independently of the regions that we can observe? If we understood consciousness as a physical phenomenon, we could in principle answer all of these questions by studying the equations of physics: we could identify all conscious entities in any physical sys- tem, and calculate what they would perceive. However, this approach is typically not pursued by physicists, with the argument that we do not understand consciousness well enough.
No, it's not because they "do not understand consciousness well enough;" it's not pursued by most physicists because they believe consciousness has nothing to do with any of these things, even if they support the Copenhagen or many worlds interpretations of QM. And, there are other interpretations that have some respect as well I believe, such as variations on the de Broglie-Bohm interpretation which is purportedly able to explain all non-relativistic QM phenomena using classical statistical mechanics, but is massively non-local:

Would Bohr be born if Bohm were born before Born?
Quantum mechanics: Myths and facts

And thus the solution to the 'quantum factorization problem:'
why do conscious observers like us perceive the particular Hilbert space factorization corresponding to classical space (rather than Fourier space, say), and more generally, why do we perceive the world around us as a dynamic hierarchy of objects that are strongly integrated and relatively independent? This fundamental problem has received almost no attention in the literature [9].
does not require a theory of consciousness at all, but perhaps just a different interpretation of QM.

According to Tegmark and Tonini there are five basic principles that distinguish conscious matter from other physical systems such as solids, liquids and gases: information, integration, independence, dynamics and utility.

I think Searle makes a good argument that 'information' can't be used as a precondition for consciousness because information presupposes consciousness: Can Information Theory Explain Consciousness?

Chalmers argues in crayz's link that consciousness doesn't seem to require 'integration' because conscious experiences like sound and vision can be separate, and perhaps even more separate when one is on acid as flabdablet alluded to. I think consciousness essentially does require integration - actually not integration but "unification." Vision and sound may be separate but even the simplest experience, a single color or an acute pain let's say, I would think must require the mutual operation of multiple neurons not possible if separated by time and space as in classical physics, and it seems to me this implies non-locality is a requirement for the neural correlates consciousness. For that reason the Integration and Mutual Information and Maximizing Information sections seemed intriguing to me.

Scott scoffs at Tonini's idea that a large enough Reed-Solomon decoder would be conscious, and this doesn't make a ton of sense to me either. I do like the idea that the brain may use error coding, but I don't get the connection between classical error coding and Tonini and Tegmark's 'integration' at all actually. Though, for different reasons I'm also dissatisfied with Hopfield neural network theory.

As pointed out earlier, the United States and large corporations like Google or other large organizations like the Pentagon would seem to possess the five characteristics of conscious matter which seems absurd.

This, to me, appears to be the coolest part of Tegmark's theory:
These two problems (consciousness as a state of matter, and the physics-from-scratch problem) go hand in hand, because a generic Hamiltonian cannot be decomposed using tensor products, which would correspond to a decomposition of the cosmos into non-interacting parts, so there is some optimal factorization of our universe into integrated and relatively independent parts. Based on Tononi's work, we might expect that this factorization, or some generalization thereof, is what conscious observers perceive, because an integrated and relatively autonomous information complex is fundamentally what a conscious observer is.

...an interesting mathematical approach to the physics-from-scratch problem of analyzing the total Hamiltonian H for our physical world:

1. Find the Hilbert space factorization giving the "cruelest cut", decomposing H into parts with the smallest interaction Hamiltonian H3 possible.

2. Keep repeating this subdivision procedure for each part until only relatively integrated parts remain that cannot be further decomposed with a small interaction Hamiltonian.

The hope would be that applying this procedure to the Hamiltonian of our standard model would reproduce the full observed object hierarchy from Figure 1, with the factorization corresponding to the objects, and the various non-separable terms H3 describing the interactions between these objects.
...
"A state selected at random from the Hilbert space of a many-body system is overwhelmingly likely to exhibit highly non-classical correlations. For these typical states, half of the environment must be measured by an observer to determine the state of a given subsystem. The objectivity of classical reality - the fact that multiple observers can agree on the state of a subsystem after measuring just a small fraction of its environment - implies that the correlations found in nature between macroscopic systems and their environments are very exceptional."

This gives a hint that the particular Hilbert space factorization we observe might be very special and unique, so that using the utility principle to insist on the existence of a consensus reality may have large constraining power among the factorizations - perhaps even helping nail down the one we actually observe.
I think what he is trying to do is figure out how you get the world we experience out of the world described by quantum mechanics by doing a 'Hilbert space factorization' on the universe, or multiverse, and then posing that the act or process of this factorization essentially is consciousness (?) and is performed by neural networks (?) in our brains.
posted by Golden Eternity at 12:24 AM on June 13 [1 favorite]


You are the one that brought up essence.

By saying, as you quote: "But I would argue that there's no evidence it has an essence."

To special-plead that 'I started it' based on that is pretty much equivalent to claiming that I admit the possibility of purple cows if I say "there's no evidence for purple cows."
posted by lodurr at 2:37 AM on June 13


I'm inherently suspicious of any theoretical approach that seeks to find ways to explain why the particular way we perceive reality should be the one that some generalized mathematical description of it collapses into, because the desire to have such an explanation seems to me to be utterly cart before horse.

I may be completely barking up the wrong tree here, but my initial impression from getting started on Tegmark's paper is that he views a QM Hilbert space description of Things Generally as somehow "more real" than a detailed cataloguing of every phenomenon encountered by any observer anywhere would be.

I don't think it is. Sure, mathematical descriptions of things have generally better predictive power than simple cataloguing would do, but that's because mathematics is all about relationships between things and oftentimes those relationships are non-obvious and require some fairly heavy-duty symbol processing to reveal. But I maintain that it doesn't follow from that that the mathematics is in some way fundamental. The world is what it is; it doesn't need help from mathematics to be so. We need help from mathematics - a discipline we devised - to help us model and predict what the world's going to do next.

It feels really, really good to exercise predictive tools and have them come up with correct predictions, which I suspect might be why it so often seems that the people with the best skills in using the best available tools are also those most susceptible to mistaking their maps for the territory.
posted by flabdablet at 4:57 AM on June 13 [6 favorites]


... a discipline we devised ...

Well, that's the question, isn't it? All these arguments seem to be just the lineal descendants of the Platonic form: a true reality to be more or less uncovered, that coincidentally happens to validate our (self-perceived) special place in the universe.
posted by lodurr at 5:01 AM on June 13


All these arguments seem to be just the lineal descendants of the Platonic form: a true reality to be more or less uncovered, that coincidentally happens to validate our (self-perceived) special place in the universe.

According to Plato's doctrine, humanity is just one form among an infinite number of forms, all of them thinking, each in its own way.
posted by No Robots at 5:25 AM on June 13


Yet more evidence that Plato has been massively over-rated.

But I'll rephrase, for the sake of fidelity to a dead greek fascist: All these arguments seem to be just the lineal descendants of the Platonic form -- they assume the existence of a true reality to be more or less uncovered, that coincidentally happens to validate our (self-perceived) special place in the universe.
posted by lodurr at 5:30 AM on June 13 [2 favorites]


According to Plato, each form constitutes a particular expression of reality as a whole. This is the basis of science.
posted by No Robots at 5:37 AM on June 13


It's also a highly selective presentation of Plato's view.
posted by lodurr at 5:43 AM on June 13


It is the standard view of Plato. The same principle is found also in the work of Spinoza and Hegel. As Feuerbach put it, "Die Wissenschaft ist das Bewußtsein der Gattungen" (Science is consciousness of the genera).
posted by No Robots at 5:52 AM on June 13


The point that I (and, I think, lodurr) are objecting to is the view that the genera are in some way more real than the phenomena they're abstracted from, as opposed to being a collection of convenient mental shortcuts for grouping those phenomena in bulk and finding ways to think about them that remain useful even when much of the detail that distinguishes one phenomenon from another is ignored.

Time spent contemplating data compression algorithms is valuable, I think, when considering the Platonic notion of forms. People who take the forms seriously are happy as long as they get the gist of things; those of us who refuse to do so prefer our reality in FLAC rather than MP3.
posted by flabdablet at 6:16 AM on June 13 [2 favorites]


Saying "It is the standard view of Plato" is a bit like saying "time is relative" is the standard view of Einstein.
posted by lodurr at 7:41 AM on June 13


The point that I (and, I think, lodurr) are objecting to is the view that the genera are in some way more real than the phenomena they're abstracted from

The phenomena are only intelligible as genera. No Robots, for example, is intelligible only as a particular exemplar of the generic human. What is more, any particular genus is only intelligible in relation to other genera. The human genus, for example, is only intelligible as it stands in relation to other genera.

Saying "It is the standard view of Plato" is a bit like saying "time is relative" is the standard view of Einstein.

The view of philosophy as a whole is that science is based on the consciousness of the genera. This is self-evidently the main thrust of Plato's project.
posted by No Robots at 8:03 AM on June 13


One of the very first things I had drilled into me when I started to study philosophy is that nothing is "self-evident."

You're genericizing the discussion to an unwarranted degree. E.g., you leave out the part where Plato thought the forms had a literal reality, instead of being something we invented to organize the world.
posted by lodurr at 8:07 AM on June 13


you leave out the part where Plato thought the forms had a literal reality, instead of being something we invented to organize the world.

I'm more generous toward Plato then you, perhaps. It is certainly true that only very recently has the relativity of the genera become obvious to all:
Things in the world are classified into orders, genera and species. Science tells us that the orders, genera and species are not realities, they are only mental concepts; they are aggregates of individuals. The individuals themselves are only forms of energy. What, then, distinguishes one order from another order, one genus from another genus, or one species from another species? And it is so in all cases. Science reduces all things to forms of energy. If we reduce all things to one universal energy, then we thereby destroy the world of realities. And so, instead of creating a world of realities, science destroyed the world of realities.--Harry Waton / A true monistic philosophy v. 1, p. 82
This process of relativization has been completely realized, and now we can rebuild the genera as relative absolutes. And, in my view, in that way we finally catch up to Plato.
posted by No Robots at 8:16 AM on June 13


To put things less snarkily, I'm by no means qualified to dismiss him entirely, but it seems very strange to try to jump directly from QM to a description of consciousness. Penrose-style speculation aside, between the realms where QM applies and the realms where consciousness operates, most physicists would agree there is an approximately classical real world, full of rocks that hang around for billions of years with no one around to see them. And consciousness as we know it seems to me to require classical information that can be cloned.

So asking why consciousness perceives the world as classical seems like an obscure way of asking why there is this approximately classical real world at all, whether or not consciousness has arisen in it. Unless you are prepared to deny the existence of the real world (or make it entirely dependent on consicousness like some of the more mystical interpretations of QM), while still affirming QM. But then, where do you get QM, if not from the scientific process operating on the real world?
posted by mubba at 8:33 AM on June 13


Metafilter: This process of relativization has been completely realized, and now we can rebuild the genera as relative absolutes.
posted by rebent at 8:43 AM on June 13 [1 favorite]


if time is money, space is a bank...
posted by judson at 11:01 AM on June 13


Speaking as a person with decent practical skills in arithmetic, but zip-squat skills in the advanced maths, I believe I have a dog in this fight even if I don't know what it might look like.

Here's what I read in the article: Tegmark is exploring the idea that consciousness may be an entity similar to our other sensory organs. It may have a physical nature (somewhere between smoke and diamond--or maybe even some physical state we've not yet defined). Or one of the enzymes produced by our endocrine system. Or like one of our fingernails. It may not necessarily reside in our physical body, but, though they as yet don't know where it may "be," it may be somewhere other than an energy field generated around & by our meaty little brains. Tegmark wonders if it may not be an emergent property of our biology, a product of our biological heritage with its own constituent parts.

Something along those lines. I accept it on faith that the quantum world is a sort of sub-universe that doesn't necessarily have a direct or linear affect our macro-existence. My awareness (they say) may be in the positioning of all those colorful little quarks and quacks and spinning bosuns, so it (our awareness) may be both material and energy at the same time--if time can be said to apply here--and limitations of human perceptions may not allow any but mathematical explanations. Fine.

So, does that mean that, when I take the transporter down to Gogon V to meet Kirk, the me who arrives isn't merely a copy of the guy who got disassembled? The original me, like RAM, ceases to exist. I guess my subjective self wouldn't know the difference, but if I'm being copied like a memory stick, then I might be reluctant to go down to Gorgon V just to have lunch with the captain. Or, may I infer that consciousness is merely a type of pre-language (generated by the body) that lets us invent such notions as emotion, speculation, language (math).

I would like to know how this works before I beam down to have lunch with Kirk, or make that appointment with the guys in Cryo-Freeze to get my body transfer lined up.

Or perhaps Tegmark is implying that consciousness is something that is scaled along a continuum which might range from knowing to turn into the light (same as a petunia) to being able to grok the path of a starship through the deadly flux of a worm-hole. Okay, I wax hyperbolic, but it's hard not spin my wingnuts when thinking about this stuff.

The part about when and whether computers may develop awareness may not be moot much longer. But we are still able to keep moving the goal posts regarding the definition, so I guess we'll have to await the revolution to find out how that part works. I hope they treat us kindly.
posted by mule98J at 11:38 AM on June 13 [1 favorite]


I thought this was an interesting comment from Less Wrong:

Mitchell_Porter:
I regard the paper as a very preliminary contribution to a new approach to quantum ontology. In effect he's telling us how the wavefunction divides into things, if we assume that the division is made according to this balance between minimal mutual information and some dynamics in the parts. Then he can ask whether the resulting things look like objects as we know them (reasonably so) and whether they look like integrated information processors (less success there, in my opinion, even though that was the aim).
posted by Golden Eternity at 1:32 PM on June 13


My guess on the first Anglo-American philosophy type objection to this theory is, he's committing right at the beginning to these really Cartesian assumptions about what consciousness is like, such as, it is unitary. But one of the only things that those philosophers seem to agree on is that the Cartesian cogito, with it's clear and distinct ideas, is not much like our actual consciousness.

I think that psychology supports that debunking view, too, although that might be because a guy told me that once and he seemed legit; I am pretty ignorant of academic psychology.
posted by thelonius at 6:49 PM on June 13


But one of the only things that those philosophers seem to agree on is that the Cartesian cogito, with it's clear and distinct ideas, is not much like our actual consciousness.

Oh, gawwwwd. Cogito Ergo Sum is not a bold declaration of the existence of Mind. It's the last refuge of an army in full retreat, and one that Hume destroys with his wrecking ball. Then Gettier gets involved, and now it's a party!
posted by Slap*Happy at 8:29 PM on June 13


Hume's criticism of Descartes' skepticism is not real impressive in light of modern developments, however. Sure, the "evil demon" probably isn't so evil if it's allowing us to maintain a reasonably consistent and useful set of reactions to things that happen to us in the world. But just because it isn't evil doesn't mean it's not lying its ass off to us in major ways.

Consider a world which is not made of particles at all, a world in which none of what physicists think they know is true, and which only presents us meaningful results to astrophysical measurements and particle experiments because some conscious entity has decided those are the stories the universe is meant to tell us. Such a world might allow for all sorts of woo phenomena as long as those things don't intrude too provably on the main story. Would the creator of such a world be evil? Does such a world seem likely for any reason?

Before 1980 the answer to the latter question would surely have been that such a conjecture does violence to Occam's Razor. But that was before we started to build such worlds ourselves, to the greatest fidelity possible at every turn. Are the creators of Battlezone, Doom, Halo, WoW or Second Life evil because they tried their best to make a fake world convincing to our natural senses?

Any argument that such human-created virtual worlds aren't totally convincing is not very convincing, because the state of the art has been advancing steadily for the last 30 years and this is without having millions of years for development and a computer the size of the entire Universe and direct access to the mind-world interface for data injection.

One thing I think we can say with some confidence, and it's really just a more modern and precise version of cogito ergo sum, is that both the mind and the world contain and process information. However this is a more useful formulation, because it opens up interesting questions like how much information, how persistent that information is, and exactly what the data structures look like. Those are questions that have potential answers with deep consequences, and before Turing and Shannon it wasn't really possible to ask them at all.
posted by localroger at 5:21 AM on June 14 [1 favorite]


and this is without ... direct access to the mind-world interface for data injection.

Our eyes and ears are pretty good interfaces.
posted by Golden Eternity at 6:11 AM on June 14


The phenomena are only intelligible as genera.

Not even remotely true. It's perfectly possible to encounter and examine a phenomenon and compare its attributes to that of other phenomena without labelling it as anything other than itself.

As soon as you decide that every phenomenon must be an instance of some category, rather than treating categories as detail-ignoring mental shortcuts that allow for bulk intellectual processing of multiple phenomena with similarities in certain attributes, you've privileged the idea of categories to a greater extent than it deserves. Or so it seems to me.
posted by flabdablet at 9:09 AM on June 14 [1 favorite]


One thing I think we can say with some confidence, and it's really just a more modern and precise version of cogito ergo sum, is that both the mind and the world contain and process information.

I'm not confident of that at all. I'm confident that my mind shares many attributes with the engineered information processing systems I've worked with for most of my life, but I'm not at all sure that the rest of the world does. "Information" seems to me to be one of those loose categories that we made up. It's actually a really slippery idea to pin down: if you can come up with an accurate definition of "information" that a primary school student would grasp, I'd love to see it.

The only thing that distinguishes information from worthless noise is that information is about something. That idea - aboutness - is, I think, so fundamental to the way minds work that it's easy for those of us who work frequently with information as information to fall into the same trap as do many folks who ascribe universal consciousness to the world, and ascribe universal information to it instead. But really, the world just does whatever it does regardless of what that may or may not be "about".

What we do is make sense of the world. We do that by processing information about the world. And yeah, there's plenty, plenty of room for investigation of - as you put it - how much information, how persistent that information is, and exactly what the data structures look like.

That last point is particularly interesting to me. Are there to be found, in brains, anything even vaguely analogous to something an IT worker would recognize as a data structure - i.e. some collection of stored information that can be copied? Personally I suspect not: the kind of separability between information and information processor that you'd find in an engineered IT machine is not something I'd expect evolution to settle on. I would expect evolution to find some use for every detail of a brain's internal organization, making the consciousness implemented by any given human body in-principle inseparable from that body.

If manufacturable consciousnesses do ever become a thing within my lifetime, I will be very surprised to find that it's feasible to implement on them something loosely equivalent to a JTAG port; about as surprised, in fact, as I would be to find out that somebody has worked out a way to look at Grumpy Cat videos on a web browser running under Linux on a machine constructed entirely from clockwork. It's an energy density thing.

If I understand Tegmark's position correctly, he's not suggesting that his proposed conscious states of matter are something overlaid on brains or to be found somewhere within them and in-principle isolable from them; he's proposing to consider brains themselves as an example of matter existing in a state with the particular properties that generate consciousness: properties that might potentially be susceptible to mathematical/statistical/quantum analysis and/or generalizable.

Spherical cow of uniform density thinking? Perhaps. But the history of physics is pretty much exactly a process of starting with spherical cows and seeing just how much more detail you actually need to add before you work out how to milk them.
posted by flabdablet at 10:02 AM on June 14 [1 favorite]


And to be clear, I don't think there's anything the least bit woo about the idea of a state of matter that generates consciousness. It's completely clear to me that brains are matter and that they do that, so really the only jarring part of the idea for me is the use of the word "state".

But when you think about it, a state of matter is no more than a description of matter's organizational form: the way its parts relate to one another. And sure, it takes scarcely any physical disruption to the way a brain is organized to stop one generating consciousness, compared to - say - the amount of physical disruption you need to apply to stop a liquid being recognizable as such. Still, that shouldn't stop an organizational form being analyzed as a form. It's an interesting line of thought.
posted by flabdablet at 10:16 AM on June 14 [1 favorite]


Oh God this is wrong in so many fundamental ways it's hard to figure out where to begin.

"Information" seems to me to be one of those loose categories that we made up. It's actually a really slippery idea to pin down: if you can come up with an accurate definition of "information" that a primary school student would grasp, I'd love to see it.

The fundamental unit of information is the bit, which is the answer to an unambiguous yes/no question. Information storage exists when you can save such an answer and retrieve it later; in modern computers, it's usually a question like "is the switch at memory location 10,320 on or off?"

There are such things as fractional bits; for example, if your question has a list of possible answers only one of which can be true, it's customary to take the log base 2 of the number of positions and say that's the number of bits it represents, so that for example the position of a 10-position switch represents 3.32 bits of information.

The only thing that distinguishes information from worthless noise is that information is about something.

NO, NO, NO, NO, a thousand times NO this is so wrong it's impossible to overstress how wrong it is. When we are quantifying it information is NEVER "about" ANYTHING. It's a commodity; it either exists or it doesn't, and if it exists it can be used for anything, just as a liter cup can be used to measure sugar or sulfuric acid. The bits don't care what they really mean, or whether they "mean" anything at all.

The realization that this has applications was one of the most important developments in the invention of computers, the von Neumann architecture which explicitly doesn't care whether a particular bit represents data or instructions for processing.

Now, proving that information storage exists does require that we have information storage to compare it against; that's how we can tell whether data written to the supposed information store are preserved. If we make changes and our reference store matches our prospective alternate store, the alternate store represents information. If there is no such match then the alternate store is either unreliable or isn't storage at all.

Life is so full of situations where we do this that we have a name for the special situation where it fails, "cognitive dissonance."

To give a very small and therefore manageable example, I have a car, and it has a color -- specifically, light grey. The car is in the driveway out of sight right now but I'm very confident that it's still light grey; if I were to walk out the front door and find an orange car in the driveway, I'd definitely experience cognitive dissonance, a mismatch between what my internal information map says to expect and what the external information map of the world serves up. There could be an explanation for that; it could have been painted or replaced without my knowledge, for example, no woo required. But it would be a significant event requiring investigation on my part.

While science tells us humans can perceive tens of thousands of colors I'm not super observant and I'd probably not notice if some godlike joker made my car a smidge lighter or darker in tone. I'd guess my memory of the car's color represents around 10 bits of information, the log base 2 of all the colors I'd be likely to notice are distinct from one another should I see them on a car.

Now, as it happens, this information is mirrored in other places where it can be cross-checked. Other people have seen my car and know its color, I have photographs of it, and the DMV has a record which is printed on the registration certificate. That all of those sources agree my car is light grey strongly suggests that those bits are really stored and served up consistently, from multiple sources, all agreeing with one another when cross-checked. That is exactly how you verify that information storage exists and is reliable, and for most of us it happens for billions of bits per day as we go about our lives without being too surprised at the fundamental state of the world.

Tegmark's conjecture would be no less meaningful -- but much more ridiculous sounding, and rightly so -- if you replaced every instance of the word "consciousness" with "computerness." Don't forget that computers, like the theories that made them possible, are a very new thing, totally unthinkable as anything other than magic for most of human history.

But computers are not, in fact, magic, and most likely neither is consciousness.
posted by localroger at 12:43 PM on June 14 [2 favorites]


Oh God this is wrong in so many fundamental ways it's hard to figure out where to begin ... The fundamental unit of information is the bit ... NO, NO, NO, NO, a thousand times NO this is so wrong it's impossible to overstress how wrong it is.

I understand your vehemence, I really do.

But the simple fact is that pointing to some feature of the world and saying "that's a bit" can't be done. You might as well point to some feature of the world and say "that's a seven".

Which is not to say that you can't point to a floating gate in a flash PROM and say "that's a structure that can store a bit". Of course you can do that. But it's physically inaccurate; the floating gate doesn't store a bit, it stores electrons. And the electrons are not the bit. The bit is in the fact that this floating gate stores appreciably more electrons than that floating gate does.

The bits don't care what they really mean, or whether they "mean" anything at all.

I agree! Only we can care about that. But the point I think you may have missed is that it's also only us who actually gets to define whether bits are present in any physical structure in the first place.
posted by flabdablet at 11:21 PM on June 14 [2 favorites]


But the point I think you may have missed is that it's also only us who actually gets to define whether bits are present in any physical structure in the first place.

This is, quite simply and absolutely, wrong. Any recognizable pattern represents information. It represents information whether any consciousness is there to recognize it or not. The pattern of craters on the Moon is information. The arrangement of planets in the Solar System is information. We can say with some confidence that this information was encoded in these arrangements of matter long before humans were around to observe them because that is the simplest interpretation of how they come to be observable by us now in their current state.

Today we can pick out one of those craters and we can read its bits to tell how large the impactor was, how fast it was going, which direction it came from, if we visit personally and collect samples what it was made of. All of that is information, and it's been stored on the surface of the Moon for billions of years. Today we're able to read those bits, but those craters would be on the moon, matter encoding bits recording impacts that happened billions of years ago, whether humans ever came along to look at them or not.

Any arrangement of matter other than a uniform gas implicitly represents information. Any broadcast of energy other than a uniform glow represents information. (In fact, even the spectrum of a uniform glow represents a little information; it has a temperature.) That's the feature of the word I can point to and say it's a bit -- the fact that any part of the world is observably different from any other part is bits. The fact that you can look at it twice and recognize a feature demonstrates that those bits are stored somehow for later retrieval as opposed to just being random noise generated from moment to moment.

We assume those bits represent the positions of atoms and molecules, but the usefulness of calling what our senses gather information is that we can talk more generically about other ways that information could be stored, processed, and conveyed to our senses which would look the same. Video games and movies with CGI special effects look an awful lot like the world but the information sources that create the images are very different.

Traditionally philosophers have assumed that only some kind of evil trickster entity would go to the trouble to represent the natural world as something other than what it "really is." But even with an ability to manipulate information that is very primitive compared to the scale of the universe, we've been doing it avidly for decades, and it is rather strange to deny the possibility of the world doing something that we do ourselves.

As for that gate in the flash PROM -- in fact, the fact that it has electrons stored in it is a bit. The answer to the question "Does this gate have electron stored in it?" is either yes or no -- that is, by design, the very definition of a bit. The special arrangment of the PROM isn't what makes it encode bits though; the fact that it's nonuniform solid matter does that. What's special is that it lets us conveniently change and read back some of those bits.

In another complex nonuniform bit of matter, the brain, matter encodes changeable and readable bits in a different way, mostly by growing dendrites and synapses. But the end result is still a pattern that can be recalled; the color of my car is light grey in my own synaptic firings, in the patterns of magnetism on a disk drive in the DMV's computer, in the symbols printed on the printed registration, in the silver halides of some old photographs, and in the photons that are being reflected from it right now. All of those things encode information, and more interestingly they encode the same little chunk of information in ways which are different but all of which align when they are compared. That is both why information is a thing that exists and why it is sometimes more useful to discuss than things like "arrangements of matter." All arrangements of matter are information, but not all information represents arrangements of matter.
posted by localroger at 6:48 AM on June 15 [1 favorite]


Any recognizable pattern represents information. It represents information whether any consciousness is there to recognize it or not.

Do you really, honestly not see the flat contradiction between those two sentences? Because I think that's really the only thing we're disagreeing about here.

Without pattern recognizers, there are no patterns; there is only everything.

It's a subtle point to grasp. Doing so requires full engagement with the consequences of the idea of one's own mind not existing, which seems to me to be a difficult task unless one first lets go of the assumption that the self has object permanence to begin with.

Any arrangement of matter other than a uniform gas implicitly represents information.

I contend that there is plenty of information one could encode about a uniform gas. You might not care about the precise position of any given molecule within that gas, but it is certainly in principle encodable - as is the history of every virtual particle surrounding every real particle within that molecule.

The fact that you can look at it twice and recognize a feature demonstrates that those bits are stored somehow for later retrieval as opposed to just being random noise generated from moment to moment.

Sure, and the bits concerned are the ones stored in our brains. It also seems to me that the fact that features are repeatably identifiable and correlatable says much more about the operation of the feature-extraction engine than it does about what the engine is presented with.

it is rather strange to deny the possibility of the world doing something that we do ourselves.

I don't think so. We're pretty specialized representers.

All arrangements of matter are information, but not all information represents arrangements of matter.

My fundamental objection to the idea that arrangements of matter are information comes from having noticed over the years that arbitrary amounts of information can be extracted by prolonged observation of anything. It seems to me that if the information were inherent in matter itself there would be some kind of limit to that process, but there doesn't seem to be; the more one understands about any physical system - that is, the more information one has already encoded about it - the more questions it makes sense to ask, opening up possibilities for generating further information.
posted by flabdablet at 7:52 AM on June 15


We will have to agree to disagree. I think our differences are more linguistic than anything else. We simply have different definitions of what "information" is. However, I am confident that my definition is more in line with the engineers and mathematicians who created information theory.

Without pattern recognizers, there are no patterns; there is only everything.

I flatly reject this as a ridiculous notion. Were those craters on the Moon when nobody was there to recognize them?

My fundamental objection to the idea that arrangements of matter are information comes from having noticed over the years that arbitrary amounts of information can be extracted by prolonged observation of anything.

I suspect this is the core of your misunderstanding. If the universe is coded the way physicists think it is, then it represents a truly vast (but not infinite) mass of information. For example, the position-velocity vector of a single electron, if we have to locate it within a sphere 13 billion light years in radius, is around 220 bits. That's what we would need to record its position and velocity to the precision allowed by the Heisenberg Uncertainty Principle.

So most arrangements of matter actually contain a lot more information than we notice; as you point out, even a uniform gas has molecules each representing a position-velocity vector. So yes, it is possible to look more closely at any given physical system and see more and more information, because there is in fact a hell of a lot of information there, and most of it isn't ultimately very important to us. This does not, however, mean that it doesn't exist or isn't in fact more important than we realize.

I suspect this is why you have hung your coat on this "recognition" thing. Now it is my own conjecture that we might not recognize the difference between a universe made up of subatomic particles and one made out of a really well done video game engine. I think you are saying that this means the underlying engine that can't be seen somehow isn't there at all, or "isn't information."

But there has to be some engine storing and processing the information that ultimately ends up at our senses, because that information (and surely you will admit that sensory data transmitted through the nervous system is information?) displays an apparent consistency which requires both internal and external storage and processing. If either our mind or the outside world did not store information, then we would not have the illusion of our memories lining up so regularly with what we observe in the world. That is a basic and necessary thing we can say about both our mind and the world we live in without needing to definitely say much else about how either really works.
posted by localroger at 8:18 AM on June 15 [1 favorite]


Were those craters on the Moon when nobody was there to recognize them?

I fully agree that easily the most plausible hypothetical is that if there had been an observer with a structure similar to ours, doing an activity similar to what we would consider to be looking at the structure we now refer to as "the Moon", at a time before human observers existed, then that observer would have noticed the kind of structures we now refer to as "craters".

Such an act of noticing would have caused some information to be encoded in that observer's brain, and then retained there.

And if that observer were to repeat the observation, then the repeat observation would also cause some information to be encoded in its brain, and the similarities between that information and that encoded at first observation would likely have been quite apparent.

If the universe is coded the way physicists think it is, then it represents a truly vast (but not infinite) mass of information. For example, the position-velocity vector of a single electron, if we have to locate it within a sphere 13 billion light years in radius, is around 220 bits. That's what we would need to record its position and velocity to the precision allowed by the Heisenberg Uncertainty Principle.

And if that cataloging were to be done - if every fundamental particle that exists or has ever existed - real and virtual - were to have its momentum, velocity, charge, spin and so forth faithfully and precisely listed in some Matrix or Library of Babel, then that would indeed constitute a stupendous, almost unimaginably large (but not infinite) quantity of information.

But it would still not be enough to predict, with certainty, whether a given atomic nucleus would decay at a given time and in a given place, ending the life of some particles and creating a bunch of new ones about which yet more information would need to be stored. Because our best available model of the universe - the one you're talking about, the one that features entities like electrons and quarks and bosons - is at its heart probabilistic.

It tells us the kinds of things that are likely to happen, and lets us make really solid bets about how frequently we will see those things actually happen under specified conditions, but no amount of information about the entities that comprise the model is enough to characterize the physical universe that the model represents.

Because we made up those particles, just as we made up the Moon as Moon and the craters as craters. We didn't just snap our fingers and poof there they were: that would be a ridiculous assertion, worthy of exactly the kind of visceral rejection you responded to my earlier comments with.

What we did do, though, and do do every waking moment, is to break up our inherently unpredictable and fully connected universe into things we can identify and tell ourselves stories about those things.

The story that physicists tell is a really good story. It represents reality to a truly astonishing level of detail. But it and all the information contained in it is a map, not the territory, and I think that identifying the one with the other is a fundamental mistake.

surely you will admit that sensory data transmitted through the nervous system is information?

I think it's a mixture of information and noise.
posted by flabdablet at 9:47 AM on June 15 [1 favorite]


If either our mind or the outside world did not store information, then we would not have the illusion of our memories lining up so regularly with what we observe in the world.

I can see why you say that, and would encourage you to give more careful scrutiny to the assumption that information exists anywhere except inside structures capable of encoding, replicating and processing it.
posted by flabdablet at 9:50 AM on June 15


We also made up temporal and spatial separation, for what that's worth; really useful map features, those.
posted by flabdablet at 9:58 AM on June 15


Traditionally philosophers have assumed that only some kind of evil trickster entity would go to the trouble to represent the natural world as something other than what it "really is."

Or an artist. Egyptian, Greek, etc., etc., art does not represent the world very realistically, and this certainly looks to be intentional.

The pattern of craters on the Moon is information.

By this definition anything that can be differentiated from some other thing is information, thus basically everything is information, and it seems to me this definition of information is kind of a useless concept for explaining anything.

Electrical engineering information theory (Shannon, etc.), concerns bits of information that can be encoded in a medium and then compressed or transmitted and received. (Having to launch asteroids at the moon to create craters is not a very convenient way to encode bits of information.) Bits themselves don't have any meaning in EE information theory, they are just bits.

Our "minds" have a very special feature (language) in that we can use things to represent other things. We can use scratches on paper or sounds to represent anything else we experience in life. This information is observer dependent in that it only makes sense to minds that have a previous understanding of what is being encoded and that it is being encoded. If we decided to use the pattern of craters on our moon to represent gods or something, this information would not be available to an alien observer.

If it makes sense to talk about craters at all without an observer or mind, it could only contain the information: "crater." Perhaps one could argue that causal relationships allow one physical thing to represent another physical thing in some way without a mind involved - a thermometer for example. One can't arbitrarily encode information on a thermometer, but a more sophisticated physical system like a computer could perhaps arbitrarily change what causal functions particular registers are performing. But I am not satisfied that this sort of causal connection between physical things is the same as or sufficient for the meaning and representation (intentionality or 'aboutness') a conscious observer or mind is capable of, and I see no reason to believe such a causal connection produces (or is) conscious experience some how.

But it would still not be enough to predict, with certainty, whether a given atomic nucleus would decay at a given time and in a given place, ending the life of some particles and creating a bunch of new ones about which more information would need to be stored, because our best available model of the universe - the one you're talking about, the one that features entities like electrons and quarks and bosons - is at its heart probabilistic.

As I understand it, this is not necessarily true. Variations of de Broglie-Bohm interpretations of QM are deterministic, but they are massively non-local (essentially position and momentum of one particle depends non-locally on the position, momentum, etc. of every other particle in the universe) and are very incompatible with quantum relativity (the other interpretations also have problems with quantum relativity).

I found this to be a pretty good discussion of Tegmark's article by Sabine Hossenfelder - Consciousness and Physics from Scratch:
Physics from Scratch

Tegmark interprets the “physics-from-scratch problem” as the question how to identify subsystems of the whole Hilbert space that can be separated as well as possible. These subsystems I believe are eventually supposed to give rise to the neatly separated (and almost classical) objects we experience, not to mention our own brains. He thus sets out to find a basis in which the interaction Hamiltonian between subspaces is minimized.
[...]
Before reading Tegmark’s paper, I would have envisioned the physics-from-scratch procedure as follows. First you need to identify space and time from your Hamiltonian. Space and time are roughly the degrees of freedom that make the rest look as local as possible. Once you have that, you should be able to write down the Hamiltonian in a series of local, or almost local, operators of various dimensions. You need to define a vacuum state, then you can start building your Fock space. The rest is basically effective field theory. That, needless to say, is all “in principle”, not that anybody could do this in practice.

Just why the world we observe contains large things that are almost classical is probably not a question we can answer by looking at the properties of Hilbert-space decompositions in general, but it depends on the specific Hamiltonian. If we didn’t have confinement and if we didn’t have gravity our universe might just be a quantum soup.
This "physics from scratch" problem seems to be pretty essential understanding to know what the heck Tegmark is talking about. If anyone has any pointers to explanations of why it is so hard to explain classical objects using standard QM I would be very appreciative. Maybe this will be helpful: Max Tegmark, "Consciousness as a State of Matter"
posted by Golden Eternity at 10:57 AM on June 15


By this definition anything that can be differentiated from some other thing is information, thus basically everything is information...

For what it's worth, that does seem to be the position of a lot of people who are able to fashion what they view as pretty rich and coherent systems of thought around the idea. E.g., I'm pretty sure that's what Rudy Rucker thinks.

I think that approach is essentially religion, and fundamentally mistaken, but that's just me. It still makes for a nice fantasy world.
posted by lodurr at 12:22 PM on June 15


flabdablet: Your concept of what information is seems self-consistent and might be useful. It is not, however, what most people mean by the word "information."

But it would still not be enough to predict, with certainty, whether a given atomic nucleus would decay at a given time

So what? This has nothing to do with anything. You seem to think that if we don't have all the information to perfectly recreate a system then we don't have "information." That's not how it works. You are really poking here around the edges of chaos theory, which in turn depended on information theory and information processing machinery for its development. Unfortunately, both words "information" and "chaos" come pre-loaded for some people with associations that cloud their meaning as terms of art.

Claude Shannon's paper would not be considered revolutionary by most people today. What made it revolutionary was that it forged a bridge between "analog" phenomena in the real world and the kind of systems being described by Turing and von Neumann. It was news to a lot of people, and staunchly disbelieved by a lot of old timers, that an analog transmission medium like the radio spectrum might have a finite bandwidth. Shannon introduced "bits" to describe those limits, but the paper was revolutionary because people quickly realized that one could use that idea of bits to describe everything else in much more generic terms than anyone had previously suspected.

GE: By this definition anything that can be differentiated from some other thing is information, thus basically everything is information, and it seems to me this definition of information is kind of a useless concept for explaining anything.

Before Shannon described it as information nobody understood that the radio spectrum had a finite bandwidth as a transmission medium. Everything is information, and the usefulness of that is we can ask how much information it is and how that information is coded.

lodurr: For what it's worth, that does seem to be the position of a lot of people who are able to fashion what they view as pretty rich and coherent systems of thought around the idea. E.g., I'm pretty sure that's what Rudy Rucker thinks.

Yes, this would be what is called the "mainstream definition" within the science and math community. If it's a religion it has turned out to be a phenomenally useful one in practical terms.
posted by localroger at 1:09 PM on June 15 [1 favorite]


Claude Shannon's paper would not be considered revolutionary by most people today. What made it revolutionary was that it forged a bridge between "analog" phenomena in the real world and the kind of systems being described by Turing and von Neumann. It was news to a lot of people, and staunchly disbelieved by a lot of old timers, that an analog transmission medium like the radio spectrum might have a finite bandwidth. Shannon introduced "bits" to describe those limits, but the paper was revolutionary because people quickly realized that one could use that idea of bits to describe everything else in much more generic terms than anyone had previously suspected.

Before Shannon described it as information nobody understood that the radio spectrum had a finite bandwidth as a transmission medium. Everything is information, and the usefulness of that is we can ask how much information it is and how that information is coded.


I don't think this is an accurate description of Shannon's formula, C = B*log2(1 + S/N), if that's what you're talking about. The maximum pulse rate for a given bandwidth comes from Nyquist. Hartley's Law set out the maximum pulse rate and number of pulse levels given signal power, signal bandwidth, and receiver resolution. Shannon capacity gives the maximum possible bit rate given channel bandwidth, signal power, and noise (AWGN) - allowing for any pulse shape and arbitrarily low bit error rate depending on error correction coding. Which is incredibly impressive for the time, don't get me wrong. But it doesn't say anything about radio spectrum having a finite bandwidth or the medium at all actually. Signal bandwidth is a given in the equation. It is just a math problem involving pulse waveforms and AWGN noise. As far as these guys were concerned, they could probably get the noise pretty close to zero if they could go to zero Kelvin and use an ideal receiver (the primary source of noise in point to point rf communications is thermal noise in the receiver). Note the capacity blows up to infinity as noise goes to zero.
posted by Golden Eternity at 9:02 PM on June 15


If it's a religion it has turned out to be a phenomenally useful one in practical terms.

It is and it has.
posted by haricotvert at 9:33 PM on June 15


So what? This has nothing to do with anything.

I was specifically responding to your point that the amount of information that would be required for a complete description of the universe is finite.

You seem to think that if we don't have all the information to perfectly recreate a system then we don't have "information." That's not how it works.

Just because information is capable of approximate representation via lossy encodings, it does not follow that everything capable of lossy representation is itself information.

Everything is information ... this would be what is called the "mainstream definition" within the science and math community.

If everything is information, then information as a concept - that is, information as distinct from any other kind of thing - loses any useful meaning, and we might as well just drop the word and use "everything" instead.

The kind of information I'm talking about is the kind that's capable of being encoded without loss into some restricted set of symbols, can be duplicated without loss, and can be fed into one end of a lossy communication channel and extracted the other end with a probability of corruption that depends on the channel's lossiness and symbol transfer rate. That's the kind of information that Shannon's writings pertain to, as far as I know.

You appear to me to be working with some kind of generalisation of that idea. If you're not simply mistaking the map for the territory, could you explain how your generalisation differs from what I've outlined here?

The part of your thinking I'm having the most trouble reconciling with my own understanding of the nature of information is that information, as I understand it, is replicable in ways that are independent of its encoding. You can encrypt it or store it in a PROM or keep it inside the dynamical state of an acoustic mercury column or scribble it on a bit of paper - it really doesn't matter how you encode it; provided only that what you're using is capable of representing enough of the symbols that comprise your information that you don't need to lose any, it's the same information regardless.

But I can't see how you'd go about demonstrating that the universe as a whole has this property, which seems to me to be fundamental to anything that can reasonably be labelled "information". In fact it's perfectly clear to me that it does not have this property, on the very simple grounds that the universe as a whole is by definition everything that is and/or was and/or will be and there isn't anywhere else to replicate it to.

Simply positing such Universal Information and claiming that it is in-principle replicable, requiring only another universe to replicate it into, strikes me as incoherent, self-contradictory nonsense.

Can you help me resolve this difficulty?
posted by flabdablet at 2:21 AM on June 16 [1 favorite]


I am not satisfied that this sort of causal connection between physical things is the same as or sufficient for the meaning and representation (intentionality or 'aboutness') a conscious observer or mind is capable of, and I see no reason to believe such a causal connection produces (or is) conscious experience some how.

By way of contrast, I'm perfectly comfortable with the idea of my own bodymind being exactly one such representation and self-representation arrangement, and feel no need to posit any mysterious extra ingredients beyond those already identified by mainstream physics to account for the resulting subjective experience. Emergent phenomena can be pretty phenomenal and it's fun for this recognition engine to label itself as one.
posted by flabdablet at 2:39 AM on June 16


Yes, this would be what is called the "mainstream definition" within the science and math community.

No, I really don't think it is. It's a mainstream view, but I sincerely doubt that scientists and mathematicians either understand the view that clearly, or would explicitly agree with it if they did. (That's true of most religions: Explain the theology of their religion to most people as though it were some other religion, and they'll disagree with important parts of it.)

Casual scrutiny reveals it to be a ridiculous definition, though. You're basically saying that you can have information before there's a consumer for information. It's like having language without having something to speak the language -- it's a meaningless concept.

The reality of the underlying stuff is not in question. Literally not in question, because there would be nothing to ask the question.

I'll put this another way: The situation you're describing is directly analogous to the situation of an omnipotent and omnipresent god. Those two facts concurrent in the same being make it incapable of acting -- it essentially nullifies itself as an actor. Whenever any 'scientific' framework is directly analogous to a mainstream religious view, I generally get pretty suspicious.
posted by lodurr at 3:22 AM on June 16


Everything is information ... this would be what is called the "mainstream definition" within the science and math community.

Kinda-sorta but not really. In physics, physical information is simply what can be used to distinguish one thing from another. Under that working parameter, yes, everything is information. If you try to narrow down that definition even a smidge, especially in relation to how people interact with information, you wind up describing knowledge rather than information, and and any tautology based on "everything is information" falls apart.
posted by Slap*Happy at 4:08 AM on June 16


Knowledge and information have nothing whatsoever to do with one another. The very idea that they might be related in any way shows a complete misunderstanding of one or the other.

I suspect that this is the main difficulty here. The problem obviously has more to do with words than concepts. Unfortunately since the words I'm using obviously don't mean the same thing to some other people that they do to me, I cannot think of any other words to use that might take their place.
posted by localroger at 5:13 AM on June 16


Disagreement about the terms is very important here, yes, but I think it goes deeper than you seem to be assuming. You seem to be taking what other people call 'information' and calling that 'knowledge,' and taking what other people call 'reality' and calling that 'information.'

Also, even at that, to say they have 'nothing to do with one another' is rather odd.
posted by lodurr at 5:31 AM on June 16 [1 favorite]


You seem to be taking what other people call 'information' and calling that 'knowledge,' and taking what other people call 'reality' and calling that 'information.'

What?

Knowledge and reality are philosophical concepts. Information as I've been using it here is an engineering term of art. This is in fact different from how lay persons informally use the word "information," but I think I've been very clear on that. If that's not clear after I name-dropped Shannon, Turing, and von Neumann, I really don't know what else to say.

Should you acquire knowledge of reality, it would be in the form of information, but that doesn't mean the concept of knowledge is related to the concept of information any more than Thomas Harris' book Red Dragon is somehow related to Webster's Dictionary because the former is made of English language words.
posted by localroger at 9:45 AM on June 16


I actually do get that you're using these terms of art. But that you can say "Information as I've been using it here is an engineering term of art" and still think that information exists in the absence of engineering is ... kinda problematic.
posted by lodurr at 11:34 AM on June 16 [1 favorite]


But that you can say "Information as I've been using it here is an engineering term of art" and still think that information exists in the absence of engineering is ... kinda problematic.

Um, no it isn't. We use other engineering terms of art in the absence of engineering all the time; for example, rocks have tensile and compressive strength, planets have angular momentum, radiation sources have a luminal intensity, and so on. These properties exist regardless of whether humans come around to measure them or to use them in pet projects or not.
posted by localroger at 12:02 PM on June 16


This is the root of the problem. I don't think anyone here disputes that if an actor existed who was capable of measuring properties that conformed to tensile and compressive strength, angular momentum, etc., that the bodies would have those properties.

But what you're saying basically implies that we're describing reality. We aren't. We're describing an approximation of reality. Hundreds of years ago we might have said something like "phlogiston exists whether humans come around to measure it or not," and by our current scientific understanding, we'd have been wrong. In a hundred years, people will look back at what we measured and deem us wrong about what we thought we were measuring.

IOW, we are inherently wrong about reality, but to paraphrase Mr. Box: Sometimes errors are useful.
posted by lodurr at 12:10 PM on June 16 [2 favorites]


I don't think anyone here disputes that if an actor existed who was capable of measuring properties that conformed to tensile and compressive strength, angular momentum, etc., that the bodies would have those properties.

No actor required. The properties exist whether they are ever measured or not.

The reason we can say this with some confidence is that we regularly tease out such properties in situations where the original object is long destroyed or unavailable by measuring some far removed effect that it had, and whenever we do this the results are always consistent with some external store of information keeping everything lined up. Whether that store is encoded in terms of particles bound together or rays of light or bits in a computer doesn't matter, but we can say that it must exist and have a certain capacity in order for the world to look and act the way it does.

Now in this sense I am talking about reality, in the sense that I am setting a bare minimum for what it can be, and I do that in terms of information because that is the engineering term of art for the minimal operation and capacity measurement of a system which can save and recall patterns. However, we have been over this ground and numerous people here have stated that they do not believe the word "information" means what it very clearly does in this context, so I am done arguing about it.
posted by localroger at 3:15 PM on June 16


The properties exist whether they are ever measured or not.

Could you speak to lodurr's point about the assured existence of phlogiston? Because that's absolutely central to the disagreement here. While you're at it, have a crack at justifying the underlying physical basis for epicycles.

It doesn't really help to point out that you can infer properties for pieces of reality not spatially or temporally proximate to the observer and demonstrate that those inferred properties are consistent with the rest of your information model. If the model's any good, that's exactly how you'd hope it would behave.

The contention on this side is that the model itself is something only brought into existence by modelers, from which it follows that the specific properties that the model is built from don't exist until measured and encoded into information (that's "information" in the precise term-of-art sense in which you're also using it). The universe that one can measure in order to evaluate the properties exists, but the properties themselves are information about the universe rather than being the universe.

Refusing to make this distinction does, in my mind, amount to mistaking the map for the territory on a pretty epic scale. Maps get better; reality just is what it is.
posted by flabdablet at 6:53 PM on June 16


flabdablet, here's the kicker: Both the map and the territory are information. The territory contains more information than the map, and the map might not be a completely accurate representation of the territory, but inasmuch as anyone agrees it is a map at all then the map and the territory represent to some extent the same information. The fact that you can compare them and find similarity is the most fundamental evidence that both are information.

I wrote and deleted another paragraph but I realized I was just revisiting crap that didn't work 100 comments up. We're done here.
posted by localroger at 7:49 PM on June 16


How is that view testable, even in principle? The only way I can think of to do that is to compare what I know about reality with what I know about information, and on at least one fundamental test, that comparison comes out unequal.

If reality is in some sense made of information, then any arbitrary part of it ought to be transformable into some other encoding with perfect fidelity. We have good observational evidence to suggest that such is not the case: every measurement is approximate, and every prediction is to some degree probabilistic.

The fact that you can compare them and find similarity is the most fundamental evidence that both are information.

I'm unaware of any way to compare a measurement of reality with reality itself, except by making some other measurement and then comparing the two measurements. It's perfectly clear that both measurements are information, but I still fail to see how that fact, or any fact about the content of that information, can justify a belief that the underlying reality is so as well.

Believe it or not, I'm not being deliberately obtuse here. I'm genuinely interested in finding a way through these objections, if possible, because the idea that reality is made of information is an interesting one that I have already spent a great deal of time contemplating. My current opinion is that it's an incorrect idea, but if you have arguments in favor of it that I haven't already considered and found unsatisfactory, I'd enjoy contemplating those as well.
posted by flabdablet at 9:02 PM on June 16


No actor required. The properties exist whether they are ever measured or not.

Thank you for not acknowledging that the 'properties' are merely best current approximations of the actual properties, and thus -- not real, in an absolute sense.
posted by lodurr at 5:22 AM on June 17


If reality is in some sense made of information, then any arbitrary part of it ought to be transformable into some other encoding with perfect fidelity.

What? I guess if you think a silly thing like this then yeah, you're very hung up. The fact that a system contains information does not mean that information is perfectly accessible. Quick, what's the byte stored at memory location $1023 in your computer? Wait, you don't know? I can assure you there's such a location and there's a byte stored there, and if it was different your computer might very well crash. The fact that you don't have a tool to retrieve and examine it at will doesn't mean it doesn't exist.

As many epistemological arguments demonstrate we don't know much, if anything, about reality. We also don't know much, if anything, about our own consciousness. Quite a pickle, eh? But both reality and consciousness serve up patterns. If those patterns are repeatable, they represent information. It doesn't matter how those patterns are stored or how they are transmitted or how they arrive at the comparison; the fact that they are consistent across comparisons is the very definition of information. If you don't understand that then I'm not going to argue it further because that's the definition of a word, and I'm tired of arguing about the definitions of words.

Now one can question whether either the universe or our consciousness really contains information, i.e. those patterns repeat consistently, or if that apparent consistency is some kind of illusion. The thing is, if either the universe or our consciousness was not storing these patterns, then they would not line up. The fact that I am never surprised by the color of my car when I leave the house in the morning suggests that the color of my car is stored and recalled reliably both by the universe and by my own mind. If either the universe or my mind simply made up a random color each time I encountered my car, then I would expect to constantly be surprised by it.

Now, there are many possible explanations for both the outer and inner consistency of this comparison, and we could argue all day long about exactly how my car's color is encoded both in the universe and in my mind. But what we can pin down pretty well is that my car's color is in fact stored and recalled with some reliability both by the universe and by my mind. The ability to store and recall a pattern is one way to state the definition of information. This suggests a functionality which unmistakably exists.

And the difference between bits and a number or static measurement is that bits are a measurement of functionality. Three is not a thing that exists; a centimeter is not a thing that exists. But a bit is a switch that can store a value for later retrieval. It either works or it doesn't, and if it doesn't it's not a bit.

Now in practice all of the bits we encounter appear to be implemented by physical arrangements of matter and energy. But we can (and do) argue about whether those arrangements of matter and energy are real, and how they work. We can't say for sure whether the world is made of subatomic particles or is a great big video game engine. We don't know nearly as much as we should about how our own minds work. But what we do know is that both sources serve up patterns, and those patterns regularly match, and that means both the world and our minds must contain information, because if either didn't then those matches would not line up.

Now we can go further, and ask exactly how much information these patterns represent and what kind of processing would be necessary to create phenomena like consciousness and the passage of time. The usefulness of putting it in terms of information is that it doesn't bind us to any assumptions about the physical systems that might be encoding the information, but it does establish a baseline which must be met by any candidate system that might be responsible for those phenomena.
posted by localroger at 5:41 AM on June 17


You're basically making definitive statements about stuff that you happily acknowledge you can't know.
posted by lodurr at 5:56 AM on June 17


The fact that a system contains information does not mean that information is perfectly accessible.

Sure. But the fact that a thing can't be accessed opens up at least two possibilities: it may be that the means to perform that access are not currently available, or it may be that whatever you're trying to access doesn't actually exist. Absence of evidence isn't evidence of presence.

Quick, what's the byte stored at memory location $1023 in your computer? Wait, you don't know? I can assure you there's such a location and there's a byte stored there

Sure. But the only reason you can assure me of that, and that I happily accept that it is so, is that both of us hold relevant information about computer engineering. I remain unconvinced that similarly relevant information is available about reality.

But both reality and consciousness serve up patterns. If those patterns are repeatable, they represent information.

That's one way to look at it, and indeed I did look at it that way for many years. My present view is that consciousness or any other representation engine generates and stores responses to stimuli, but unless and until it can be clearly demonstrated that some process of symbolic encoding is involved, it is better to remain agnostic about the nature of both responses and stimuli.

The example you gave earlier of the response to light of silver halide grains in a photographic film to projected images is apposite. I will happily agree that the pattern of the oxidation states of the silver atoms in the film constitutes a symbolic encoding of the scene projected onto the film at exposure time: it is in principle feasible to enumerate those atoms and represent each one's oxidation state as representation-independent bits. However, I remain utterly unconvinced that the silver atoms themselves, or the halide atoms proximate to them, or the electrons that rearrange themselves to form those oxidation states are information in and of themselves. Storage media yes; stored information no.

I'm drawing a distinction here between the symbols stored in a structure capable of symbol storage (clearly information in the technical sense) and the storage structure itself (not at all clearly information in the technical sense) and I think that's a completely legitimate distinction.

The fact that I am never surprised by the color of my car when I leave the house in the morning suggests that the color of my car is stored and recalled reliably both by the universe and by my own mind.

To my way of thinking, using the word "stored" when speaking of the color of the car itself (as opposed to the information you have about the car) is begging the question; storage, in this context, only makes sense as an operation applied to information. I prefer to think of the color of the car as a measurement of the world, and the resulting information in my mind as information about the world.

Three is not a thing that exists; a centimeter is not a thing that exists. But a bit is a switch that can store a value for later retrieval. It either works or it doesn't, and if it doesn't it's not a bit.

Again, this statement appears to me to conflate the bit itself with the structure where it might be stored. Given the representation-independent nature of information, that appears to me to be an error.

There are two ways to generate bits, as far as I know: (1) combine other bits or (2) measure something. I understand that you're comfortable considering both of those processes to be (variants of?) the same thing. I'm not, and I hope by now that I've offered enough reasons why that's so to take at least some of the heat out of your initial No! No! No! No! response to what I wrote.

We can't say for sure whether the world is made of subatomic particles or is a great big video game engine.

Agreed. I personally quite enjoy trying to work out what kinds of thing I can't know, and I do my best to avoid making firm assertions about those things. Many of them are fun ideas to play with in a pass-the-bong kind of way, but I think taking them seriously is a mistake.

We don't know nearly as much as we should about how our own minds work.

Agreed. I have a strong suspicion, but cannot prove, that there's nothing going on in them that couldn't plausibly emerge from the operation of physical structures; I don't think consciousness needs special sauce, just appropriate physical organization. Reasonable people like Golden Eternity disagree about that. Pass the bong.

But what we do know is that both sources serve up patterns

I know that my mind generates patterns in response to stimuli, which I view as an instance of case (2) above; I agree that it's reasonable to discuss those patterns as patterns, and that having been recognized as such they can reasonably be construed to be information, but I cannot honestly assert that whatever it is I'm measuring in order to recognize those patterns must itself be made of information, or even inherently patterned.

and those patterns regularly match

We wouldn't notice them unless they did. And in fact we have no idea at all and can have no idea at all of how much or what kind of non-patterned inconsistent activity is going on within and around us as well. Best we can do is generate statistical information about some of it when we notice patterns in its side effects.

and that means both the world and our minds must contain information

I'm comfortable with the idea that my mind contains information (which, given that my mind is part of the world, trivially means that the world does too), and I'm comfortable with the idea that information can be derived by measuring the world.

However, as you point out yourself in your sub-argument about video games, we have no way to know whether the information we hold about the world actually says much about its fundamental nature. Which is why I'm still unwilling to make, or even take seriously, firm assertions such as "everything is information".
posted by flabdablet at 8:20 AM on June 17 [1 favorite]


Of course, if what you're actually saying is that the only thing we can ever work with is information, then that's a much weaker and much more plausible claim - and one with which, given my current opinion about the nature of consciousness, I'm inclined to agree.
posted by flabdablet at 8:35 AM on June 17


I'm comfortable with the idea that information can be derived by measuring the world.

Great! Where do you think this information comes you derive by measuring the world comes from?

You are obviously hung up on some idea that the existence of information has some relationship to how it is perceived or used. It doesn't, that's a basic definition, and there's really no point arguing any more in this circle.
posted by localroger at 9:27 AM on June 17


Where do you think this information comes you derive by measuring the world comes from?

The measurement process.
posted by flabdablet at 9:41 AM on June 17


You are obviously hung up on some idea that the existence of information has some relationship to how it is perceived or used.

No, I think I'm just a little more rigorous than you appear to be about distinguishing the idea of information from other, related ideas.

It seems to me that in order for something to qualify as information in the relevant technical sense, it needs to be encoded in symbols, replicable without loss, and independent of its storage medium. With which, if any, of those attributes do you disagree?
posted by flabdablet at 9:45 AM on June 17


It seems to me that in order for something to qualify as information in the relevant technical sense, it needs to be encoded in symbols, replicable without loss, and independent of its storage medium. With which, if any, of those attributes do you disagree?

All of them. You are thinking of an entirely different concept than I am, and I'm out of ways to try to explain the difference because you keep circling back to just not understanding what I'm talking about because it's not your concept.
posted by localroger at 11:00 AM on June 17


Okay then, so if those are not the fundamental attributes of information (in the technical sense referred to in the Shannon paper you raised earlier), then how do you distinguish information from non-information?

If you have no way to do that (which I currently suspect you don't, given your insistence that everything is information) then either the idea of information you're using is redundant, or you're working within a conceptual framework that truly does mistake the map for the territory.

In particular, if "information" is merely a synonym for "everything", then none of your previous arguments re. using the word in its technical sense hold water.

But perhaps you do indeed have some useful referent for the word "information" that is as yet undisclosed. As I noted much earlier, it's a slippery idea to pin down.
posted by flabdablet at 11:12 AM on June 17 [1 favorite]


I'm out of ways to try to explain the difference because you keep circling back to just not understanding what I'm talking about because it's not your concept.

That's actually a pretty good account of my own internal experience when I first started making a serious attempt to understand the idea of information in anything other than the kind of vague and nebulous way you get from working with information technology every day.

Having started from a position very much like your current one, it came as a bit of a shock to find that not only was I personally incapable of ascribing a definite objective meaning to the word "information" but that everybody else's attempts at definitions also ended up bouncing fruitlessly from ill-defined abstraction to metaphor to ill-defined abstraction before sliding off into one form or other of philosophical waffle. The whole thing was just a big ugly ball of self-referential mud.

Seriously, have a look at Wikipedia's page on it, which is pretty typical of the genre. The very first sentence: "Information is that which informs". Gee, thanks a bunch; big help! But in fact that's about as good as it gets, once you strip away all the side trips and obfuscation.

After following all the other definitional threads around for a while it became apparent that the word as commonly used is too slippery and mutable to be useful for genuine understanding of anything, and that some rather more restrictive technical meaning would have to be stuck to if any progress was to happen at all. And that's when I started to take seriously the idea of finding attributes of information that would let me distinguish it from non-information.

I've been pretty happy with the attributes I've been working with so far (symbolic encoding, perfect replicability, representation independence) because they're a good fit with all the information-related math I've encountered. But if you can find some non-handwavey case where they don't work, I'm really keen to know about it.
posted by flabdablet at 11:59 AM on June 17


In particular, I'd love to see a tight technical definition of "information" in terms that don't imply or presuppose some form of consciousness, that doesn't need the symbolic encoding and perfect replicability criteria.
posted by flabdablet at 12:08 PM on June 17


In particular, I'd love to see a tight technical definition of "information" in terms that don't imply or presuppose some form of consciousness, that doesn't need the symbolic encoding and perfect replicability criteria.

I've always been partial to Gregory Bateson's "a difference that makes a difference" definition, but I'm not sure it gets us out of all of the difficulties you describe. It does, at least, free us from the need for a consciousness that actively "understands" the information it receives.
posted by yoink at 12:29 PM on June 17


Information is the functionality that makes it possible provide the same answer to a particular question repeatably. Full stop.

Now I hear you thinking up weird edge cases so I'll address a few of them.

Questions are posed through transmission channels. They may consist of sensory observations, measurements, or actual exchanges where a "question" signal is sent to request a response. All such channels are mathematically identical.

All information transmission channels are potentially subject to noise, so the answers we get through them may be tainted. (This was the exact focus of Claude Shannon's paper.) If we ask a question to which we know the answer should be "yes" and we get an occasional "no," we can use that to gauge the amount of noise in the channel. By asking questions repeatedly we can to a great extent (but never perfectly) cancel out the presence of noise. This is why your requirement for perfect fidelity, at best, has nothing to do with anything useful.

As for what is and isn't information, the too-simple answer is that if you get the same answer every time it means there's information (AT THE OTHER END, it is by no sane assessment associated with the transmission channel except that transmission might fuck it up). In the presence of noise you look for bias. If you're lucky you can ask questions whose answers you know to gauge the noise level so you know how much repetition you need. There's math for that.

Now there are a couple of situations where the question "does that answer represent information" gets a little muddy. Information can change; the answer to the bit "Is New Orleans in the northern hemisphere?" is always yes, but the answer to "is it daylight there?" changes every 12 hours. In 99.999% of cases it is very obvious whether such changes are random or reflect processing being performed in a systematic way on the information you're querying, but in some edge cases, such as cryptography and image enhancement, you have to haul out the math.

It's also muddy if you get the same answer to a broad range of questions (such as "what value is stored in that memory location" when "that memory location" is actually outside the range physically implemented on a computer, and the answer is always zero). A lot of times this has to do with how the question is posed ("how the information is encoded"). For example in the fractal dust space we call the Universe for nearly every possible location the answer to "is there a particle here?" is NO. It would be very rare to get YES because you hit a proton. But if you rephrase the question "List all the particles and their positions," the same information looks much richer. Again there's a lot of math addressing this, most of it finite math taught in computer engineering courses. The same information can be represented for transmission in many different ways, giving different strengths and weaknesses in the transmission capacity requirements and ability to cancel noise.

Now what you deride as "you're just saying information is everything" is what I actually phrase as "everything is information," and there's a reason for that, because the statement "everything is information" actually says something useful. Even if you know nothing else about the system providing those answers information itself can be quantified and manipulated (which again is what Shannon's paper is about).

So if we know nothing about a system except that it contains information, and information we can only imperfectly access at that, we can still speculate as to how the information is coded, what the best way to phrase questions is (I suppose you might call that "symbolic coding" but that's a very narrow term for something that's really very broad in practice), and most importantly how many questions can the system answer -- what is it s capacity, or size.

The reason this is important is that even if you can't tell whether the universe is a big mass of subatomic particles or a video game engine, you can say that it represents a certain amount of storage and processing (which is what causes information to not always be the same, see above). You can even say that certain implementations would imply greater or lesser requirements.

And despite my use of the word "question" here, this in no way implies the involvement of a consciousness; machines and physical processes create the mathematical equivalent of these "questions" all the time. In many cases we can infer the existence of information we can't query directly because we know something about how the Universe's information is being processed, and we can infer that certain patterns must have existed in the distant past in order to produce the distantly related answers we get today -- such as that asteroid that hit the Moon from a certain angle and speed 4 billion years ago, which we read today from physical measurements of the crater.

Anyway, I realize there isn't anything in this post I haven't said four times already. I believe the biggest problem is that you are trying to use words to describe something which is really mostly math, and which is much more obvious from that perspective.
posted by localroger at 3:32 PM on June 17


Localroger, I don't see how, in your definition, you make any rigorous distinction between information and noise. And, of course, Shannon himself said that white noise is "information rich." The problem with the kind of reified conception of 'information' that you're trying to lay out here seems to me precisely the problem that Bateson's "difference that makes a difference" points to. That is, you want simple "difference" to count as "information"--the more "differences" there are, the more information there is. But noise is full of "differences." Maximally entropic states are precisely the most "information rich" states in that sense--they are uncompressible precisely by virtue of that density of information. But they contain no "signal"--they answer no "question" (to use your terms). They are differences that do not make a difference.

And, of course, even more problematically, information is "noise" when it is not information that matters to you. And no, I'm not smuggling consciousness back into the equation here. If I open a YouTube video and then another video autoplays over the top of it as I browse around listening to the first one, that stream of "information" is "noise" to me because it is preventing my capacity to seek the "answer" I'm looking for in listening to the first video. But if someone else deliberately opens both of those videos together in YouTube Doubler then exactly the same stream of differences is purely "information" for them because they "question" they brought to the event was different than the "question" I did. And the same thing is true whether those "questions" are conscious "questions" or unconscious "questions."
posted by yoink at 9:29 AM on June 18 [1 favorite]


Yoink, the very first line of my post says that answers which do not repeat are not information. That immediately disqualifies white noise.

When Shannon calls white noise information rich he is using the word in a completely different sense, which is happening a lot in this discussion. With that comment Shannon is focusing on the transmission channel and white noise takes up bandwidth. In that sense it's common to speak of the noise as if it is information because it is taking up a resource through which we expect to receive information, but the noise does not indicate the presence of a thing with persistent existence, which is what I'm speaking of when I say "X contains information" or "X is an information structure."

It's really kind of backward to ask what the properties of information are because as we work with it in the real world information is actually a property of other things. However there are situations where information is the only property of a thing we can access, and in that case it makes sense to consider the information itself and the sort of storage and processing which might be necessary (however those inaccessible things are implemented) to support it.

Information which does not matter to you is not noise, it is interference, which is a different thing but of course also a problem of transmission channels and also has nothing to do with whether the channel has connected you with a source of information.
posted by localroger at 9:55 AM on June 18


the very first line of my post says that answers which do not repeat are not information

But that's wrong. Pi doesn't repeat, but it's information. Of course, for the purposes of most "questions" it's essentially noise--but if your question is "what is the ration of a circle's diameter to its circumference" then that non-repeating, apparently random sequence of digits is, most definitely, information.

If your answer to that is "but the answer itself gets repeated every time you ask the question" then the same thing can be said of white noise. That is, it's possible to ask questions to which white noise is a meaningful answer and will always be the sought-for answer.

So, again, your definitions do not work.
posted by yoink at 10:19 AM on June 18


Pi doesn't repeat, but it's information.

This is so off the wall I'm not even sure how to respond to it.

Pi is the same every time you calculate it. Yes, that's what I mean by "repeats." The fact that the values which are the same every time you read them themselves contain no apparent order does not matter.

(Actually the infinitely non-repeating irrational number PI is a thing that doesn't exist within the universe, so by "Pi" I assume you mean the calculated approximation which is the result of processing being performed according to an algorithm. All of which is information.)

White noise does not repeat every time you get it, that is the definition of noise. Saying "white noise is the answer" conflates the map with the territory in a serious way. The concept of "white noise" can be an answer which is always the same, but white noise itself contains no information.

It is also possible to record white noise and the recording will be information, because if you play it you will get the same random-looking numbers every time, which is my definition of information. To do this you will of course need to store the noise. You will find the device which does that is able to do so because it represents information storage, and yes, it's literally a machine that in that case turns noise into information.
posted by localroger at 12:37 PM on June 18 [1 favorite]


White noise does not repeat every time you get it, that is the definition of noise. Saying "white noise is the answer" conflates the map with the territory in a serious way. The concept of "white noise" can be an answer which is always the same, but white noise itself contains no information.

But, again, this is exactly my point. You are saying "there is information just there in the world, and it can be quantified and analyzed without ever asking the question 'information for whom' or 'information under what circumstances.'" But you've just admitted that you can have a burst of noise that under one set of circumstances gets defined as "informationless white noise" but under another set of circumstances is "repeatable signal which contains some specific information." In other words, there is no analysis you can perform on any actual real world burst of "white noise" that determines that that is either "information empty" or "information rich."

Do you finally see the problem with your position now? Because I really don't think I could make it any more clear.
posted by yoink at 12:49 PM on June 18 [1 favorite]


But you've just admitted that you can have a burst of noise that under one set of circumstances gets defined as "informationless white noise" but under another set of circumstances is "repeatable signal which contains some specific information."

That is not what I said at all. A description of a signal is not the signal. The word "elephant" does not have a trunk. White noise does not contain information, although the fact that a channel is occupied with white noise is itself a bit of information that is not part of the white noise itself. (In actuality, white noise is not a very good example for this because it may contain a little information, such as a spectrum or volume, but these are usually more a product of the transmission channel than the information-free source. A better example would be a string of random numbers generated by some process like radioactive decay. In the real world all signals are a mix of information and noise, and there is no such practical thing as a signal that contains none of either.)

In other words, there is no analysis you can perform on any actual real world burst of "white noise" that determines that that is either "information empty" or "information rich."

If you can read it twice and it's the same, it's information. If you can't, it's noise. Whether the information means something is completely 100% irrelevant to anything. A description of the signal is not the signal; that is new information created by processing the signal after it is received.

Ten seconds of white noise spewing from a 100 kbps channel is a megabyte but it is not information because you can't get it back. A recording of those ten seconds which is the same every time you play it is a megabyte of information. The label you put on the recording is something else entirely, and has nothing to do with the recording itself.

A more difficult critique would be the transmission channel that outputs nothing, because nothing could be blank potential information storage or a lack of information. Only observation over time can straighten that out if you have no other knowledge of what's at the other end of the channel. Lots of real world systems that do contain information spend most of their time outputting zero.
posted by localroger at 1:11 PM on June 18


On the staircase -- the objection that the presence of white noise represents information is ontological in nature and says more about the limitations of language than my concept of what information is.

To clarify, go back to my definition of a bit; it is the repeatable answer to an unambiguous question. Now white noise is not a positive answer to the question ARE YOU A WHITE NOISE GENERATOR; that would be a bit, and would be information. White noise is a random string of responses to some other question, like WHAT IS YOUR TEMPERATURE. There is nothing in the white noise that says HI I AM A WHITE NOISE GENERATOR. The fact that the data is white noise instead of the information we want is something we have to determine by analysis -- processing, which can create new information.

As soon as you change the questin to IS THERE CHANGING BUT POSSIBLY MEANINGLESS DATA ON THE LINE you have essentially created a different information channel for what we generally call metadata. And the answer is not as clear as you might think; there is a substantial difference between a source which rebroadcasts the same ten megabytes of crud over and over and one which sources purely random numbers all day long, and it's a difference you will not be able to detect with a five megabyte sample.

So the fact that a white noise generator is hooked up might be information and it might be information we can infer from an analysis of the white noise, but it is not information which is transmitted by the white noise itself.

This may seem pedantic, but it is no more so than the fact that "exists" and "big" may both be adjectives but they do not imply that God exists via the ontological argument's use of them.
posted by localroger at 2:56 PM on June 18


It's really kind of backward to ask what the properties of information are because as we work with it in the real world information is actually a property of other things.

In other words, information is about something.
flabdablet: The only thing that distinguishes information from worthless noise is that information is about something.

localroger: NO, NO, NO, NO, a thousand times NO this is so wrong it's impossible to overstress how wrong it is.
If you can read it twice and it's the same, it's information. If you can't, it's noise.

Not a fan of jam bands, I take it :-)

Ten seconds of white noise spewing from a 100 kbps channel is a megabyte but it is not information because you can't get it back.

White noise is demonstrably existent, and yet it is not information? Also, not information because not replicable? Works for me.

Also, a small clarification on exactly what I have always meant by "perfect replicability": that would be replicability where the probability of corruption can be made arbitrarily small.
posted by flabdablet at 4:28 AM on June 19 [1 favorite]


It's really kind of backward to ask what the properties of information are because as we work with it in the real world information is actually a property of other things.

In other words, information is about something.


No. That's another misuse of words as imperfect tools.

Words make it seem as if there is a distinction between "X contains information" and "information is a property of X." There is, in fact, no distinction. The difference is words.

LIGHT GREY is the information-containing answer to the question WHAT COLOR IS LOCALROGER'S CAR. Another way of putting that is that light grey is a property of my car. Still another way of putting that is OWNS A CAR and CAR COLOR IS LIGHT GREY are properties of localroger. These statements are all equivalent and they all represent exactly the same 10 bits or so of information.

But the information that my car is light grey, whether it's a real car made of countless subatomic particles or a simulation in Grand Theft Auto, is not "about" anything except that my car is light grey. It's not my gang sign, it's not my political party's color, it's not even a color I like because frankly I bought the car because it was on the lot and cheap. It's just a few bits I need to be able to pattern-match when I need to find my car in a large parking lot.

The bottom line for me is that we have entire industries devoted to making devices that "store information," and any definition of information which cannot include what those devices return to us -- whether they are new and full of zeroes or having been used to record white noise -- is WRONG. I am not as ignorant as most people with engineering backgrounds and I am aware of the skeptical arguments. And for me, this is bedrock; I may be made of particles or simulated. The world may exist or it may be an illusion. But my experience represents an interaction between systems exchanging information. That information may be incomplete, it may be wrong, it may be recorded noise, it may be the zeroes of an as-yet unused storage module, but that's what it is. It is a real thing, and in fact the only real thing I can reasonably argue actually exists.

The question of how can we know some particular thing presented to us is information is a completely separate thing from the question of what information is. Language is a particularly stupid thing when dealing with edge case arguments about transmission channels, because we use the same word for six different things. Yeah, the random crap coming through from a white noise generator can be called information because it's the answer to the question "what is that status of the comm line this microsecond?" But that is qualitatively different from information that is mostly fixed, such as the properties of an asteroid which has been floating in space for a billion years without any interaction with other matter, and that's different still from the content of a storage device which can remember and regurgitate random answers on command.

I am confident it's possible to build an argument using words that would justify my concept of information. But the problem is with the words, not the concept, and the argument isn't really about what information is; it's about whether "exists" and "big" are just both adjectives or whether one property is fundamentally different from the other.
posted by localroger at 4:56 PM on June 19


LIGHT GREY is the information-containing answer to the question WHAT COLOR IS LOCALROGER'S CAR. Another way of putting that is that light grey is a property of my car. Still another way of putting that is OWNS A CAR and CAR COLOR IS LIGHT GREY are properties of localroger. These statements are all equivalent and they all represent exactly the same 10 bits or so of information.

And yet another exactly equivalent way of putting that is "LIGHT GREY is information about localroger's car" and "OWNS A CAR and CAR COLOR IS LIGHT GREY are pieces of information about localroger".

Note particularly that this formulation both explicitly specifies that the property concerned is of type "information" and that it does not require the use of words such as "question" and "answer" that imply the involvement of an agent with an intentional stance.
posted by flabdablet at 11:43 PM on June 19


They're only exactly equivalent if the speaker is held constant. "LIGHT GREY" is a subjective value, after all; and for a native speaker of American English, it also includes the information that the speaker is very likely from a commonwealth nation. From the perspective of the commonwealth speaker, that information is not present, because it's the result of an analysis by the hearer. The information ("very likely from a commonwealth nation") is not objective information any more than the value of "LIGHT GREY" is. Even the means by which a common color value for LIGHT GREY could be stipulated (say, "#ddd") is likely to be based on characteristics of the human perceptive apparatus. The result of analysis, after perceptual manipulation by the brain.

What exists prior to that wasn't 'information' in anything like the same sense that it was after. So at the very least, the term 'information', applied across this range, is being applied to qualitatively different things.
posted by lodurr at 2:40 AM on June 20


that information is not present, because it's the result of an analysis by the hearer.

Personally I have "analysis" filed under the same rough category as "measurement", and am comfortable with the idea that what either of these processes generates is information.
posted by flabdablet at 6:36 AM on June 20


I'm assuming 'measurement' includes transformation?
posted by lodurr at 6:50 AM on June 20


OWNS A CAR and CAR COLOR IS LIGHT GREY are pieces of information about localroger".

..that imply the involvement of an agent with an intentional stance.


Ownership certainly requires an intentional stance and it seems to me a phenomenal stance. There is no physical property of paper that makes it money or a car that makes it Roger's. They are mental properties. Even in purely behaviorist definitions, which I don't think that would be adequate, the term would apply. After all, Dennett is a materialist and he coined the term, I believe.

"Question" and "answer" are good terms to use because they show how information theory can be applied to both physically measured and mentally (consciously) observed things.
posted by Golden Eternity at 7:08 AM on June 20


I'm assuming 'measurement' includes transformation?

The way I have it all filed in my head, reversible transformations are just re-encodings of the same information, but non-reversible transformations and measurements both generate new information that is about whatever is being non-reversibly transformed or measured. I draw the distinction between reversible transformation of existing information and measurement because, unlike localroger, I don't see the world as a finite information store.

Ownership certainly requires an intentional stance

Sure, but the only ownership in that example is the one included in the sample piece of information. My point was that casting the statement in the form "X is a piece of information about A" avoids the loaded terms "question" and "answer" in localroger's formulation "X is the information-containing answer to the question 'what (property P) is A'", and makes explicitly clear that X is information in a way that his other formulation "X is a property of A" does not.

Personally I think "question" and "answer" are both quite poor terms to use when attempting to come up with a way to think about information outside the context of choices made by entities capable of intent.

Of course it's still possible to undermine my own formulation by making an issue of "about, according to whom?" but I still like it more because "about" directly labels the relationship between X and A in a way that speaking of questions and answers does only indirectly.

I understand that you don't consider "information" to be a property of things except in relation to conscious entities; personally I'm closer to localroger on that, in that I consider information at least as useful and general-purpose an idea as energy or momentum. Where localroger and I mainly differ, I think, is that localroger's view seems to be that the world has inherent properties, information content among them, while I don't think it has any at all: in my view, we ascribe properties to the world in order to understand it.

The way I see it, all properties (both their values and their characterizations) are information generated by observers via measurements, and objectivity is the extent to which observers doing their very best to perform their measurements honestly and well are likely to agree.

Mathematics I see as the general science of patterns: a huge information-manipulation framework we've built over time. Physics is the application of mathematics to measurements, which we do both in order to increase our objectivity and to compress the information we have to the greatest extent we can by finding as many patterns inside it as possible.

Given this conceptual structure, I see the old questions of "who or what breathes fire into the equations to create the actual universe" and "why is there something rather than nothing" as coming from a viewpoint that's pretty much exactly backwards; they cause me no difficulty whatsoever, because I'm perfectly happy to take the universe's existence as a (in fact the only) given.
posted by flabdablet at 9:06 AM on June 20


casting the statement in the form "X is a piece of information about A" avoids the loaded terms "question" and "answer" in localroger's formulation "X is the information-containing answer to the question 'what (property P) is A'"

I would defend my use of the words "question" and "answer" in that there is really no such thing as passive information gathering or measurement; in a sense all potential information we receive is in the form of answers to questions, most often "what happens to X type particles that bounce off this thing." The fact that those particle interactions are in fact queries which inevitably affect the thing being measured is the basis of the Heisenberg Uncertainty Principle.
posted by localroger at 2:48 PM on June 20


there is really no such thing as passive information gathering or measurement; in a sense all potential information we receive is in the form of answers to questions

We appear to have switched positions :-)

There's an awful lot of information transfer that doesn't involve answers to questions. Even much of the information we receive and many of the measurements we make are completely gratuitous.
posted by flabdablet at 12:23 AM on June 21


We appear to have switched positions :-)

Words are coarse things with which to make fine distinctions. Even the most solid seeming definitions become situational if you parse them too closely.

Even much of the information we receive and many of the measurements we make are completely gratuitous.

Actually about the only thing we can observe which isn't the answer to a question in the form of a particle bouncing off of something is the emission of a primary radiation source. Just because we didn't deliberately ask the question doesn't mean it isn't one (and this is one reason I find the idea that consciousness has anything to do with the definition of information so silly).
posted by localroger at 6:47 AM on June 21


Words are coarse things with which to make fine distinctions. Even the most solid seeming definitions become situational if you parse them too closely.

I think a point often missed by those of a more mathematical persuasion is that mathematics itself, and to a much greater extent physics, suffers from exactly the same thing. It's a language, and as such it has words too.

Mathematics is all about patterns and is to a great extent self-referential, so it's possible to nail definitions down extremely tightly in most cases. There are still places where it "leaks" - ideas such as "set" that would have been considered unambiguous when first devised that now have lots of subtly different variants. But physics, being a pile of information about the real world, relies much more on the definitions of things that are not themselves parts of physics.

Because physics (which, to my way of thinking, includes ordinary common-sense knowledge about the way the world behaves as well as the more precise and accurate mathematical branches) is so, so good at producing good predictions given well-formulated questions, there's a really strong tendency to forget that it's only a model.

I think the idea that consciousness is fundamental to the definition of information makes perfect sense, but only as a special case of the idea that consciousness is fundamental to the act of definition itself; information is not special in this regard. And this is why I take the view that in a hypothetical universe exactly equivalent to this one but for an absence of consciousness, there would not only be no craters but no moon - there would be only everything.

Note particularly that I am not defending and do not subscribe to the position that there were no craters or moons in this real universe before the advent of human consciousness; I take relativity seriously enough that my model of this universe is pretty much forced to include its entire history and future (and that's future, not futures; totally not a fan of many-worlds).
posted by flabdablet at 7:40 AM on June 21 [1 favorite]


The fact that those particle interactions are in fact queries which inevitably affect the thing being measured is the basis of the Heisenberg Uncertainty Principle.

I don't think that's really correct. It may be what Heisenberg was thinking, but I don't think it is how physicists typically look at the problem today. I love the below description from Edward Witten of how QM and the Heisenberg relation originate from the singularities (real singularities not the mind uploading kind) in Newton's law of gravity and Coulomb's law:

Newton's Law of Gravity: f = -g*m1*m2/r^2
Coloumb's Law: f = k*q1*q2/r^2

Witten (PDF):
Obviously, the inverse square law means that the force becomes infinite for r -> 0. This singularity did not cause great difficulties for Newton, since (for instance) the Moon was always at a safe distance from the Earth, far from r=0. However, once the electron and atomic nucleus were discovered almost a century ago, the 1/r^2 singularity did become a severe problem. A simple calculation based on nineteenth-century physics showed that, because of the strong force at small r, the electron should spiral into the nucleus in about 10^−9 seconds. This was obviously not the case. To cure this problem, quantum mechanics was invented. In quantum mechanics the position x and momentum p of a particle do not commute, but obey Heisenberg’s relation [p;x] = −ih; with h being Planck’s constant. This relation gives a sort of “fuzziness” to the electron and other particles. Because of this fuzziness, one never really gets to r=0, and the problem is averted.
Also Wikipedia:
The original heuristic argument that such a limit should exist was given by Heisenberg, after whom it is sometimes named the Heisenberg principle. This ascribes the uncertainty in the measurable quantities to the jolt-like disturbance triggered by the act of observation. Though widely repeated in textbooks, this physical argument is now known to be fundamentally misleading.[4][5] While the act of measurement does lead to uncertainty, the loss of precision is less than that predicted by Heisenberg's argument; the formal mathematical result remains valid, however.

Historically, the uncertainty principle has been confused[6][7] with a somewhat similar effect in physics, called the observer effect, which notes that measurements of certain systems cannot be made without affecting the systems. Heisenberg offered such an observer effect at the quantum level (see below) as a physical "explanation" of quantum uncertainty.[8] It has since become clear, however, that the uncertainty principle is inherent in the properties of all wave-like systems,[4] and that it arises in quantum mechanics simply due to the matter wave nature of all quantum objects. Thus, the uncertainty principle actually states a fundamental property of quantum systems, and is not a statement about the observational success of current technology.[9] It must be emphasized that measurement does not mean only a process in which a physicist-observer takes part, but rather any interaction between classical and quantum objects regardless of any observer.[10]
The de Broglie-Bohm interpretation I mentioned before apparently attributes Heisenberg uncertainty strictly to our lack of knowledge of initial conditions as in classical mechanics. I have no idea how de Broglie-Bohm gets around the singularity problems, then.

I understand that you don't consider "information" to be a property of things except in relation to conscious entities

I'm not sure. I'm still trying to figure out what the concept of "information as a property of things" really offers, but I think academia should encourage looking at problems in different ways, so I'm all for it. Shannon Entropy and physical information theory or whatever seem interesting. I'm seeing it as a study of how to preserve and transfer deterministic patterns in random processes. It seems to me the math involved could be useful in answering many interesting question that do relate to consciousness:

* How much data appears to be stored and processed by the mind?
* How much, and in what way is data lost by the mind?
* How much data can it be said the mind is processing when performing certain tasks?
* What does this say about the data storage and processing capacity of the brain?
* How much data can be communicated from one part of the brain to another? (from the retina's to the visual cortex through the optic nerve?)
* How does this compare to the data storage capacity of a given a physical medium like CMOS? Or the data bandwidth of digital communication technologies like COFDM, etc?
* What does this tell us about the necessary physical properties of the brain?
* etc.

However, I think answers to these questions still leave us very much in the dark as to an explanation of consciousness. Nothing in any physical theory shows why physical stuff should start to have conscious experiences, but we can correlate the properties of our experiences with physical properties, and I suppose this could include "information" as a property. I think the best definition of "information" should just refer to empty patterns, or bits, or symbols devoid of meaning. I'm tempted to say only a mind can give them meaning, but that doesn't sit all that well with me. You could derive materialist, behaviorist conceptions of meaning based on causal relationships between different patterns. Memory locations in the frame store of a graphics controller determine the "color" of particular pixels, and could be said to represent colors. But this is very much lacking because equating causality with representation is tenuous, and it is still missing conscious experience; and meaning to us is primarily experiential it seems to me.

Obviously, all science is done by making measurements and observations, but I think we should resist inserting the act of observation in physical theories except as a purely physical interaction until it becomes absolutely necessary. Perhaps Tegmark sees that a consciousness-based theory of observation is required to explain how the classical world we observe comes out of QM, but I don't understand it.

Given this conceptual structure, I see the old questions of "who or what breathes fire into the equations to create the actual universe" and "why is there something rather than nothing" as coming from a viewpoint that's pretty much exactly backwards; they cause me no difficulty whatsoever, because I'm perfectly happy to take the universe's existence as a (in fact the only) given.

I'm not sure what you are saying here. Does anyone doubt the universe's existence? I don't see how taking the universe as a given addresses the ultimate existential question. One could imagine the church saying to Galileo, "we take 'why does the sun rotate around the earth' as coming from a viewpoint that's exactly backwards because we take it as a given." Maybe you are saying "existence precedes essence," like Sartre or whatever.
posted by Golden Eternity at 9:05 AM on June 21


I don't see how taking the universe as a given addresses the ultimate existential question.

What it does, if you take it seriously, is flip the existential question back on itself. From the review you linked:
His first interlocutor, the philosopher Adolf Grünbaum, throws down a challenge: Why be astonished at being at all? To marvel at existence is to assume that nothingness is somehow more natural, more restful. But why? The ancients started with matter, not the void; perhaps nothingness is stranger than being.
Having accepted that "nothing" labels an idea, this is a viewpoint that follows very naturally, putting the "ultimate existential question" firmly in pass-the-bong territory and disqualifying it as anything worth actually taking seriously.

As for the business about being uncomfortable with what came "before" the Big Bang: again, if you accept the existence of the universe as a given, you need to start looking sideways at the idea of "before". If our understanding of the way things work says that no present-day feature's history (including the spatial separation between things) can be traced back further than fourteen billion years, but we know for certain that the universe exists, and there seems to be some conflict between that certainty and some of the consequences of our understanding of how things work (i.e. a paradox) then we need to be looking harder at our understanding, not at the universe. There's nothing wrong with the universe, but there are plenty of ways to misunderstand it.

For example, the idea that there must have been something before the Big Bang rests on a number of generally unstated assumptions: (1) the Big Bang was a specific event (2) every event has at least one cause (3) events have an unambiguous ordering in time (4) time is something other than a way to specify a kind of separation between events (and no doubt many others). None of these four assumptions holds up under close scrutiny, and the falsehood of any of them is enough to render the concept "before the Big Bang" meaningless.

And yet there's a nagging feeling attached to both of these questions that says that attempts to undermine them in ways like this are somehow not legitimate because of course they're big important questions and you can't just pick them to pieces like that because... because... well, I don't know, you just can't, that's all. I want answers to these Big Questions and if all I'm going to get from you is this kind of fucking smartarse nitpicking I'll seek those answers elsewhere because you're not even a proper philosopher so what would you fucking know.

One of the things that having experienced psychosis has taught me is that it is completely feasible to be absolutely certain, to the core of your very being, of things that are just flat wrong. So I've learned not to trust that nagging feeling. If a proposition doesn't pass careful reality checking, it needs to get tossed.

For what it's worth, if some non-human and possibly engineered device one day claims to me to be conscious, and gives me a convincing account of itself, I'd be inclined to give it the benefit of the doubt. Emergent properties really are often quite remarkable.
posted by flabdablet at 9:51 AM on June 21 [1 favorite]


By the way, existence is something I do my very best to take as the only given, as I have yet to conceive of a more fundamental observation. Existence doesn't rely on any kind of distinction. It just is.

That's the best counter I can offer to your point about Galileo and the Sun.
posted by flabdablet at 9:59 AM on June 21


A proper philosopher
posted by flabdablet at 10:13 AM on June 21


I have no idea how de Broglie-Bohm gets around the singularity problems, then.
A simple calculation based on nineteenth-century physics showed that, because of the strong force at small r, the electron should spiral into the nucleus in about 10^−9 seconds.
Simple calculations based on nineteenth century physics also showed that ideal black bodies at thermal equilibrium should radiate infinite energy.

The singularity problem is a problem caused by using the best available formulae of the nineteenth century to model things sufficiently unlike those that were observed in order to derive those formulae. I'd be astonished to learn that de Broglie-Bohm has a similar problem, but if it does, all that means is that it - like every physical theory - has a limited domain of applicability, occasionally fails to make useful predictions, and could be improved by either patching or refactoring.
posted by flabdablet at 11:03 AM on June 21


totally not a fan of many-worlds

Amusingly relevant to this discussion, Rudy Rucker (whose concept of what information is does seem very similar to my own) once argued that the Many Worlds universe does not contain any information because it never makes a decision. I believe he intended that as a reduction to absurdity :-)
posted by localroger at 1:37 PM on June 21


Well yeah, except that you could apply exactly the same reductio to the unitary universe by noting that in such a universe the fact that things happen means that they were always going to happen, which means that there have never been any decisions to make in the first place... pass the bong.
posted by flabdablet at 11:43 PM on June 21


This sounds like good reading
posted by flabdablet at 4:13 AM on June 22


And yet there's a nagging feeling attached to both of these questions that says that attempts to undermine them in ways like this are somehow not legitimate because of course they're big important questions and you can't just pick them to pieces like that because... because... well, I don't know, you just can't, that's all. I want answers to these Big Questions and if all I'm going to get from you is this kind of fucking smartarse nitpicking I'll seek those answers elsewhere because you're not even a proper philosopher so what would you fucking know.

Grunbaum and others, I find, don't like the question of why there is something rather than nothing because it is inconvenient for their staunchly naturalistic views. They have an overwhelming psychological need to fit everything into their Procrustean bed of a meaningless universe filled with dead matter -- and damn anything wondrous that would upset the apple cart.

I mean, hey, why assume that the universe is explicable at all? Why wonder why things fly or float? Why assume that there is any rational correspondence between what we think and what is?

There's as much or as little warrant for taking unshakable ignorance about all questions whatsoever as a given as there is for taking the existence of the universe as a given.
posted by shivohum at 7:08 AM on June 22


Grunbaum and others, I find, don't like the question of why there is something rather than nothing because it is inconvenient for their staunchly naturalistic views. They have an overwhelming psychological need to fit everything into their Procrustean bed of a meaningless universe filled with dead matter -- and damn anything wondrous that would upset the apple cart.

That's a rather more polite paraphrase of "what would you fucking know" but it amounts to the same thing. Neither is an actual argument.

I can't speak for Grunbaum obviously, but I can assure you that in my own case there is no aversion to wonder whatsoever. I am frequently filled with wonder, and delight in understanding things. There are so many wonderful real things to wonder at that it seems a pity to be distracted from them by trumped-up non-issues. There will always be more things in heaven and earth than are dreamt of in my philosophy, so I want to make as much room in here for as many of them as will fit. If that strikes you as Procrustean, you're welcome to your view; it's no skin off my nose.

Also, I have to take issue with your implication that matter is inherently dead or that people who refuse to waste time taking nonsense seriously think it is: I am made of the stuff after all, and I don't want to go on the cart.

As for meaning: meaning is where you find it. Personally I don't find any at all in the operation of systems far removed from the human scale. And I love that lack of meaning. Nature is huge and wild, and I think if it all had to mean something, especially if that were something tiny enough for people to grasp, the world would be a far duller place than it is. Again, your mileage may vary and that's fine by me.

By the way, if you don't take existence as a given, what do you take as your starting point for inquiry, and why? As I mentioned upthread, the reason I take "this is" as the sole given is because it's the only observation I have ever made that requires no conceptualization at all: no distinctions, no categories, no understanding, no descriptions, no labels - just pure raw observation. I'd be interested in your thoughts about how I'd even begin to go about doubting it.

There's as much or as little warrant for taking unshakable ignorance about all questions whatsoever as a given as there is for taking the existence of the universe as a given.

I've tried, but I completely fail to see the point you're making here. Unshakable ignorance about all questions whatsoever is an attitude; raw undifferentiated existence is an observable. I see no equivalence between them. Would you care to expand on this?
posted by flabdablet at 11:13 AM on June 22


I'd be astonished to learn that de Broglie-Bohm has a similar problem, but if it does, all that means is that it - like every physical theory - has a limited domain of applicability, occasionally fails to make useful predictions, and could be improved by either patching or refactoring.

Yeah, that was probably a pretty dumb comment by me. I think de Broglie-Bohm would use the ground state of an electron in an atom as defined by Schrodigner's equation, and something similar for gravitational singularities.

I am frequently filled with wonder, and delight in understanding things.


I think the "experience of wondering at the existence of the world" (or the existence of existence) that Wittgenstein talked about may be a bit different:
"The truth is that the scientific way of looking at a fact is not the way to look at it as a miracle. For imagine whatever fact you may, it is not in itself miraculous in the absolute sense of that term. For we see now that we have been using describe the experience of wondering at the existence of the world by saying: it is the experience of seeing the world as a miracle."
--Wittgenstein

"The most beautiful emotion we can experience is the mystical. It is the power of all true art and science. He to whom this emotion is a stranger, who can no longer wonder and stand rapt in awe, is as good as dead. To know that what is impenetrable to us really exists, manifesting itself as the highest wisdom and the most radiant beauty, which our dull faculties can comprehend only in their most primitive forms- this knowledge, this feeling, is at the center of true religiousness. In this sense, and in this sense only, I belong to the rank of devoutly religious men."
-- Albert Einstein
But whatever, for me the admittedly nonsensical, incoherent thought "why does existence exist?" produces a positive psychological experience that I find beneficial especially when life is not going well. It is because the question can't possibly have a satisfactory answer that it does so. Obviously, others see it differently - for the most part as just nonsense I imagine, or to the extent that it makes sense a scientific answer can be thought about. It is dumb to argue about it, and contrary to the Einstein quote the point is not to diminish other's experiences and perspectives. Previously: Much Ado About Nothing

For what it's worth, if some non-human and possibly engineered device one day claims to me to be conscious, and gives me a convincing account of itself, I'd be inclined to give it the benefit of the doubt. Emergent properties really are often quite remarkable.

It does seem like it would be pretty difficult for me to falsely believe a robot is conscious which would be spooky, but I'm less convinced after seeing stuff like this:

Computer allegedly passes Turing Test for first time by convincing judges it is a 13-year-old boy
Creepy robot can mimic facial expression
MIT's Nexi MDS Robot: First Test of Expression

The technology is getting closer. As you said it is very possible for us to be absolutely certain of things that are just flat wrong. So relying solely on our subjective judgement to determine if something is conscious seems dubious. It would be better, imo, to try - as impossible as it may be - to understand more about how consciousness comes about for us before being too willing to attribute it to AI.

The Turing test seems a little backwards to me. Doing psychological testing to determine under what circumstances we are capable of attributing minds to other things tells us a lot about our minds' facility for doing this, but alone isn't completely convincing evidence that the other things necessarily have minds.

Also, I have to take issue with your implication that matter is inherently dead or that people who refuse to waste time taking nonsense seriously think it is: I am made of the stuff after all, and I don't want to go on the cart.

"Matter" is just a theory. As physics has advanced matter has become less and less 'material.' Perhaps it has a ways to go before it can start to explain consciousness and better explain life.

Chomsky: The machine, the ghost and the limits of understanding, Paraphrasing:
Mind-body is meaningless. If there's no body, there's no mind-body problem. ... The term physical is just sort of like an honorific word. It's kind of like the word 'real' when we say 'the real truth.' It doesn't add anything. To say something is physical today just says "you've gotta take it serious," but there's no further concept of physical, or material, or body, so there can't be a mind-body problem.
posted by Golden Eternity at 3:18 PM on June 22


That's a rather more polite paraphrase of "what would you fucking know" but it amounts to the same thing. Neither is an actual argument.

I never claimed it was an argument; it's an analysis of a cultural malady. Imagine someone who lacked any interest in morality -- not who was immoral, but who simply had no interest in or care about morality. "What's the big deal? Who cares what's right and wrong? I see nothing in those terms." this person might say.

Suppose this attitude was prevalent in a culture, because people had been taught that was normal. The minority who still cared about morality could ultimately only shake their heads at the attitude, I think. In the end there could be no knockdown "argument" about such a thing, only pointers to the psychological incentives for the culture to support such a strange view.

Also, I have to take issue with your implication that matter is inherently dead or that people who refuse to waste time taking nonsense seriously think it is: I am made of the stuff after all, and I don't want to go on the cart.

Any philosophy which holds that matter is not primarily dead -- i.e. insentient -- is panpsychist, which is not a materialist position at all; it's dualism.

I know there are materialists trying to have their cake and eat it too. That's what emergent and mysterian theories of consciousness are. They don't want to destroy everything that's human.

But in the end the whole appeal of materialism is a rigid simplicity: everything that exists is publicly knowable through the sense organs aided by instruments. That, after all, is what matter is. Get rid of that and you break the beauty of the system, and you might as well embrace any number of other far more "mystical" philosophies.

Dennett and others realize this, which is why he and the most clear-eyed materialist philosophers know they have to be eliminativists about consciousness -- and thus about meaning, free will and really pretty much everything else valuable in life (as all these are in consciousness).

By the way, if you don't take existence as a given, what do you take as your starting point for inquiry, and why?

Taking something as a given doesn't make it inappropriate to question it. "Something exists" is a fair axiom, but it is far from the case that one may not ask why it could not be otherwise. By questioning existence we're not asking why we should assume something exists, but why in fact something does exist, when there's nothing logically contradictory about a world where it didn't.

Would you care to expand on this?

It's what I said above about givens; taking something as a given or self-evident proposition does not make asking questions about it inappropriate. That would be to draw an arbitrary line that bounds inquiry ("the universe just does exist, and there's nothing meaningful more to be said than that"), and if we're going to bound inquiry arbitrarily, we could just as justifiably bind it completely, on every topic whatsoever.
posted by shivohum at 6:35 PM on June 22 [1 favorite]


Imagine someone who lacked any interest in morality -- not who was immoral, but who simply had no interest in or care about morality...

As far as I can tell, this is a complete red herring. I see no analogy at all between a lack of interest in morality and a deliberate policy of trying not to take self-contradictory nonsense seriously.

Any philosophy which holds that matter is not primarily dead -- i.e. insentient -- is panpsychist, which is not a materialist position at all; it's dualism.

I don't see that at all. Life and sentience are two (non-equivalent!) properties observable in matter that's organized in certain characteristic forms. You might as well claim that any philosophy which holds that matter is not primarily pink is pangreenist.

But in the end the whole appeal of materialism is a rigid simplicity: everything that exists is [in principle] publicly knowable through the sense organs aided by instruments. (insertion mine).

The really interesting questions, though, are how knowable, and to what ends? A lifetime's experience has taught me that every advance in human understanding prompts a flood of new questions, and I have no reason to think that's going to change.

"Something exists" is a fair axiom, but it is far from the case that one may not ask why it could not be otherwise.

One may ask that. One may also ask "What is the difference between a duck?" But one shouldn't be surprised when one gets answers every bit as silly as "One of its legs is both the same".

That would be to draw an arbitrary line that bounds inquiry ("the universe just does exist, and there's nothing meaningful more to be said than that"), and if we're going to bound inquiry arbitrarily, we could just as justifiably bind it completely, on every topic whatsoever.

This is not reasonable. It's toddler logic: a piece has crumbled off the side of my cookie so I'm just going to hurl the whole thing to the floor.

Axioms are starting points for reason, not limits on it, and "this is" is truly the only genuinely self-evident truth. It's not a postulate, it's not arbitrary; it's a matter of direct experience. And "describe this" offers plenty of scope for inquiry.
posted by flabdablet at 11:18 PM on June 22


But whatever, for me the admittedly nonsensical, incoherent thought "why does existence exist?" produces a positive psychological experience that I find beneficial especially when life is not going well. It is because the question can't possibly have a satisfactory answer that it does so.

Nifty toys are indeed fun to play with, fun is really important, and paradoxical questions are nifty toys. Answers to them? Not so much.

The Turing test seems a little backwards to me. Doing psychological testing to determine under what circumstances we are capable of attributing minds to other things tells us a lot about our minds' facility for doing this, but alone isn't completely convincing evidence that the other things necessarily have minds.

I don't think we'll ever have "completely convincing" evidence of that; I think it's always going to be more of a convenient working assumption that makes sense of a whole pile of otherwise non-understandable behavior than anything else.

You yourself are passing my own Turing test quite plausibly, for what it's worth :-)

Relevant material here.
posted by flabdablet at 11:29 PM on June 22


As far as I can tell, this is a complete red herring. I see no analogy at all between a lack of interest in morality and a deliberate policy of trying not to take self-contradictory nonsense seriously.

But the amoralist would say the same thing about morality. He'd call it self-contradictory nonsense. He'd be wrong, too, of course, but he'd have the same grounds for making the argument (i.e. his sense that the question of morality was just a linguistic confusion).

There's nothing self-contradictory about asking why everything is not, basically, something analogous to whatever it is we (do not) experience in deep, dreamless sleep.

"Why is there something rather than nothing?" is essentially the same as asking "what the hell IS all this?" What IS this world, experience, life, movie, video game we are in? What's behind the curtain? What's the higher purpose?

You're welcome to say "there is nothing behind the curtain and there is no higher purpose." But that's still an answer, however impoverished and unsatisfactory, to a valid question.

For it IS a great mystery. Obviously it cannot be answered in terms of science. It could only be answered in terms of something higher than science. Even philosophy can only speculate, of course.

Calling the mystery "self-contradictory nonsense" is just the kind of cultural disease I was talking about: an unwillingness or really, a taught inability, to see the obvious where it is inconvenient to the materialist religion; to dismiss the most important with lazy linguistic sleight-of-hand; and to hold up a lack of imagination as a proud shield against engaging with the profound.

Most of the educated class are infected, unfortunately, and there's no cure in sight.
posted by shivohum at 5:35 AM on June 23


"Why is there something rather than nothing?" is essentially the same as asking "what the hell IS all this?"

I can't immediately think of a statement with which I could disagree more strongly.

For what it's worth, my own approach to dealing with "what the hell IS all this?" is to break it into an answer and a challenge: It is what it is. Now describe it as accurately as you can.

What's behind the curtain?

Essentially a good question, but framed in a way that presupposes that there does exist a man we're not supposed to be paying attention to; I object to the framing, not to the question, because it seems to me that such presuppositions are not aids to clear thinking.

What's the higher purpose?

Leading the witness again, m'lud.

You're welcome to say "there is nothing behind the curtain and there is no higher purpose." But that's still an answer, however impoverished and unsatisfactory, to a valid question.

Am I also welcome to say "I don't like the way you've framed those questions; before I deal with those, I would prefer to examine the idea of purpose carefully, and find out whether it is reasonable to apply it universally"?

For it IS a great mystery. Obviously it cannot be answered in terms of science. It could only be answered in terms of something higher than science.

I have no problem with the idea that self-contradictory nonsense would present itself as being "higher" than science.

an unwillingness or really, a taught inability, to see the obvious where it is inconvenient to the materialist religion

Given that the question in question completely hinges on an unwillingness or perhaps a taught inability to acknowledge the single most obvious thing there is - existence itself - I find this claim ironic. Also, for somebody clearly speaking from a religious perspective to use "religion" as a pejorative is a cheap rhetorical trick unworthy of a serious thinker, and disrespectful to the speaker's own religion besides.
posted by flabdablet at 6:52 AM on June 23


Given that the question in question completely hinges on an unwillingness or perhaps a taught inability to acknowledge the single most obvious thing there is - existence itself - I find this claim ironic

That's a conceptual confusion. A "given" is not unquestionable. Again, a questioning of why something couldn't be different has nothing to do with acknowledging the existence of the thing.

I acknowledge that the capital of the United States is Washington, D.C. -- there is no "inability to acknowledge that," but that has nothing to do with whether or not it couldn't be different.

For what it's worth, my own approach to dealing with "what the hell IS all this?" is to break it into an answer and a challenge: It is what it is. Now describe it as accurately as you can.

That shows a misunderstanding of the question. When we ask the question, we're not looking for a detailed list of everything in creation. We're looking for the basis of creation itself, which obviously must lie in something OTHER than the created. That's the psychological need, that's the philosophical question -- to which the materialist response is "that's irrelevant; that doesn't exist; that isn't a valid question; *malfunction* *error code 169*" -- essentially an endless series of evasions, because materialism doesn't HAVE an answer. Of course I've no doubt they're sincere evasions; they're tragically self-deceptive, that's all.

but framed in a way that presupposes that there does exist a man we're not supposed to be
paying attention to


Who says it's a man? It would by definition have to be something inconceivable: Kant's noumenon. Nor is it something, then, that we're not supposed to be paying attention to, but something that we CANNOT pay attention to, that is by definition beyond our ability to attend to it, at least with our normal tools of knowledge acquisition.

Am I also welcome to say "I don't like the way you've framed those questions; before I deal with those, I would prefer to examine the idea of purpose carefully, and find out whether it is reasonable to apply it universally"?

Of course you can. But you're simply going to come to your pre-determined culturally mandated conclusion. No?

I mean, suppose we stood next to a giant blazing magical volcano filled with unicorns and mermaids who swam in lava. They couldn't be touched or talked to, prodded or manipulated, nor could they leave the volcano, but there they were: in total violation of every rule of science.

And suppose people came in wonder and talked about the volcano, and asked "what IS this?"

And suppose a staunch scientific materialist walked by and said, "I don't even find that question intelligible. That's a meaningless question."

What could one say to such a person? Would it not be clear that they were denying the wonder of the volcano because it was inconvenient for an ideology?
posted by shivohum at 7:18 AM on June 23


We're looking for the basis of creation itself, which obviously must lie in something OTHER than the created.

Why do you think that's obvious? It's not at all obvious to me. Could you explain the reasoning that renders it obvious to you?
posted by flabdablet at 10:17 AM on June 23


Who says it's a man?

The noted materialist philosopher L. Frank Baum.
posted by flabdablet at 10:23 AM on June 23


Why do you think that's obvious? It's not at all obvious to me. Could you explain the reasoning that renders it obvious to you?

Because to explain something is to subsume it into a larger, ordered whole. A thing can never, therefore, be explained by itself, or by some subset of itself. That would be tautological, or incoherent -- it would add no pertinent information.

If I ask "why did the ball fall?", no adequate explanation could be simply that there is a ball, or that something fell, or even both facts together -- that the ball fell. Any adequate explanation would have to invoke more: that someone dropped the ball, or because of the law of gravity, or because that was Fate. Whatever. Facts and order outside the facts to be explained must be added, in other words, to arrive at explanation.

But what happens when we ask about the explanation for all visible facts and orders together as a whole? If we crave an explanation for that, then from what we have just said, it could not possibly lie within any particular fact or ordering of particular facts, but must lie in something superior to both.
posted by shivohum at 3:19 PM on June 23


So if I'm reading your reasoning correctly, we have a choice between

(a) allowing an exception to the general rule that a thing can only be understood by invoking other things outside itself, in the case where the thing to be understood is "the totality of all things"

and

(b) accepting a self-contradictory definition of "the totality of all things".

Personally I favour (a). I do so on the grounds that (1) every rule has edge cases and grey areas, while existence itself (in other words, the totality of all things) is simply not deniable and (2) I am unwilling to base what I take to be genuine understanding on a deliberate contradiction, because from a contradiction anything follows and I wish to keep my self-contradictory nonsense firmly in its place in the toybox.

Is there anything wrong with that reasoning, from your point of view?
posted by flabdablet at 11:22 PM on June 23


It's not a self-contradictory definition of the totality of things. There are possibilities outside the totality of things that could explain it: namely, non-things. Things that are not limited in the way we understand limitation. Entities beyond human conceivability and comprehension.

Of course we can only point, refer, and speculate about those things. They are in Kant's noumenon; we cannot know them directly, but only make guesses.
posted by shivohum at 5:59 AM on June 24


There are possibilities outside the totality of things that could explain it: namely, non-things.

Do these posited non-things exist or not?

If they comprise part or all of existence, they exist and are therefore things, regardless of whether or not anybody can conceive of or comprehend them. If they're not part of existence, they're made-up nonsense and I see no justification for taking them seriously.

And if they're things, regardless of how incomprehensible they might be, then calling them non-things is a flat contradiction.
posted by flabdablet at 7:34 AM on June 24


No, it's a false dichotomy. The noumenal is, by definition, beyond the categories of "existence" and "non-existence" and "thing" and "not-thing"; that is, those terms do not apply to them, because those terms are defined in the parlance of human understanding.

Indeed, that would have to be the case if the noumenal were to be the explanatory groundwork for the existence of all things.

That's the meaning of "incomprehensible" as I have been using it. Incomprehensible doesn't mean "something just like we know, except that we do not know it."

It is something that defies all our categories entirely; that's exactly what enables it to be what explains them.
posted by shivohum at 7:56 AM on June 24


It is something that defies all our categories entirely; that's exactly what enables it to be what explains them.

If we can't know anything about it, we can't know how it "explains" anything. If we know that it "explains" out existence, we know rather a lot about it. You really, really can't have that both ways.
posted by yoink at 9:16 AM on June 24


If we can't know anything about it, we can't know how it "explains" anything. If we know that it "explains" out existence, we know rather a lot about it. You really, really can't have that both ways.

We don't know how it explains things. In fact, technically we don't even know for certain that it explains things; certainly any explaining it did would be in a mode that we cannot imagine. We merely posit a concept of it, which we acknowledge immediately as inadequate, and suggest its having explanatory ability as the most appealing supposition and ground for an overall worldview.

At the same time, within that concept is nestled some negative knowledge of it: it is not bounded by our concepts. Nothing certain can be said of it except that nothing certain can be said of it.

We are clearly engaged in speculation, not knowing, about it, trying to use words we acknowledge as utterly impoverished to ponder the ultimate. And when we do that, the meaning of those words necessarily distorts, as light does near a black hole.
posted by shivohum at 9:25 AM on June 24


It is something that defies all our categories entirely

...except insofar as you've just categorized it as "the noumenal".

If I know about it but can't sense it even in principle then it's an idea, which is part of my mind, which is part of me, which is part of everything.

The noumenal is where lieth all those things with a sort of raffia-work base, that has an attachment; it is the means by which all the missing left socks transmute into empty wire coat hangers in long-abandoned rooms; it is where people go after successfully travelling back in time and shooting their own grandfathers. It is a silly place.
posted by flabdablet at 9:28 AM on June 24 [1 favorite]


except insofar as you've just categorized it as "the noumenal".

Right; that's just a limitation of language. It means "not anything we can understand."

If I know about it but can't sense it even in principle then it's an idea, which is part of my mind, which is part of me, which is part of everything.

Well we may be sensing it every moment. It may run though everything; everything may in fact be an infinitesimal manifestation of it. The question is whether we can understand it, which is a different issue.
posted by shivohum at 9:31 AM on June 24


Well we may be sensing it every moment

In which case it exists, and is therefore part of that which we seek to understand, and is therefore not outside that which we seek to understand, and therefore (by your logic) cannot explain that which we seek to understand.
posted by flabdablet at 9:49 AM on June 24


In which case it exists

Not necessarily. It may, for example, be capable of manifesting in existent things, but that does not mean that it is adequately describable by the term "exists" with respect to whatever it is in its essence.
posted by shivohum at 9:53 AM on June 24


So far the only sense I can make of this idea is as some kind of ill-defined placeholder for the cause of uncaused events, invented for the purpose of forcing all events into a conceptual web of causality.
posted by flabdablet at 10:02 AM on June 24


There are plenty of benefits it gives us:

-It gives us a sense that there might be a hidden meaning behind life, death, suffering, and existence, even if we cannot fully know it.
-It serves as a heuristic, pointing out that there might be non-discursive ways of knowing, ways beyond science and observation
-It gives us a sense that there might be a fundamental intelligence behind things, even if we cannot comprehend it
-It gives us a sense that we might be part of something mysterious and truly incredible, and not our limited little selves
-It also makes an excellent object of speculation, assuming we keep in mind that the speculation will never catch up and fully understand. Viz much of philosophy and theology.
posted by shivohum at 10:14 AM on June 24 [1 favorite]


Yes, it's a nice toy. But it's no help at all in resolving the question of why there is something rather than nothing, because all it is is a semantic end-run around the idea of "something", and not even a particularly elegant one.
posted by flabdablet at 10:22 AM on June 24


I never said it "resolved" the question of existence. It obviously doesn't, given that I've said it's a big Mystery. What it does is show how parochial and ideologically motivated is the idea that the question itself is meaningless. Which is what we were originally discussing.
posted by shivohum at 10:37 AM on June 24


No, it doesn't show that at all. What it does do is show that all that needs to be done in order to make meaningless questions seem profound is to keep doubling down on meaningless abstraction until reason cries Uncle.
posted by flabdablet at 10:45 AM on June 24


It's not meaningless abstraction, it's pointing out the necessarily arational basis for reason.

Anyhow, this discussion has gone as I predicted, no? The materialist ideology stifles the ability to see issues that it cannot handle - basically everything fundamental in human life.

And it's to the point that after all this you cannot even admit that the question of existence is a cognizable question, even if you disagree with my answer.
posted by shivohum at 11:01 AM on June 24


Never denied it's worth pondering. Spent a lot of time pondering it myself; in fact it's exactly by pondering that question, but refusing to accept nonsense, that I worked out that the closest thing it has to a correct answer is as follows:

The reason there is something rather than nothing is that "nothing" only has meaning as part of something else. Nothingness is always relative: to say that there is "nothing there" is to identify some restricted region of everything that includes no things, or at most no things worth considering.

But you can't get rid of all the things. This answer exists, which makes it something, and that alone is sufficient to demonstrate that nothingness cannot possibly be universal.

It is certainly possible to construct mathematical models of empty universes, but they do not model this one; the existence of this answer is proof enough of that. It is also possible to model this universe in ways that imply that it has a finite age - in fact, our best available model is one such - and it then becomes quite natural to wonder what, if anything, was going on before this all began. However, that still doesn't remove the directly observable fact of this universe's existence; questions about what (if anything) might exist outside the spatial and temporal ranges covered by such models are not identical to the question of why there is something rather than nothing, and conflating them is an error.

Since there is simply no way around the fact that something does exist, the simplest and most direct answer to the question "Why is there something rather than nothing?" is "because nothing is simply not an available option". Existence is inescapable. It's a brute fact and if we're honest with ourselves we are simply forced to accept it.

Of course, we don't like being forced to lump things, especially big obvious things beyond our control. We're a bolshie lot, and we'd rather be happy than right, so the natural tendency is to try to argue with reality and rules-lawyer our way around it. We invent imaginary classes of thing that we declare not-things by decree, and then ascribe all kinds of mysterious properties to them including ill-defined but somehow compelling powers of "explanation". Or we cling doggedly to the assertion that our "why" demands a better answer by acting as if causality rather than raw existence is the brute fact and starting point that must be taken as given. Or we simply hurl our cookie to the floor and declare that "it is a mystery"; or if we're really desperate to save face we might even capitalize that and declare that "it is a Mystery".

So yes, the question is worth pondering, but mainly for the reason that doing so reveals more about ourselves than it does about the putative subject.
posted by flabdablet at 12:13 PM on June 24


But you can't get rid of all the things. This answer exists, which makes it something, and that alone is sufficient to demonstrate that nothingness cannot possibly be universal.

You've made this identical mistake over and over. The answer exists, but one can imagine a world where it didn't exist. It's called a counterfactual.

Anyway, I'd suggest reading in philosophy a lot more, as these kinds of basic conceptual mistakes cripple your ability to make a cogent counterargument. William James addresses the whole topic of something and nothing well in one of his essays.

A blind man arguing against the existence of color would make many of the same arguments that you do. "What is this red? A made-up word!" "What, the sun looks like an orange tastes? What gibberish!"

Blindness - not knowing - become so painful that psychologically there was a need to deny the very existence of color, to invent long self-justifying arguments to deny the most basic human intuitions so long as it contributed to that result, and so on - this is precisely what we see happening in the culture today with the materialists.

Against this psychological compulsion no ordinary argument can prevail.
posted by shivohum at 2:41 PM on June 24


...one can imagine a world where it didn't exist. It's called a counterfactual.

You don't get to label an imagined thing a counterfactual simply because it isn't factual; you must specify some set of conditions under which it could possibly be factual. And in the particular case of the existence-free universe, no such conditions exist or indeed could possibly exist; the concept is analytically self-contradictory.

The argument you're trying to support is that nothing could be a feasible alternative state of affairs to something, implying that the existence of everything-as-a-whole demands explanation. This is simply not analogous to an argument that $UNFAMILIAR_ATTRIBUTE is a meaningful property of certain things.

Blindness - not knowing - become so painful that psychologically there was a need to deny the very existence of color, to invent long self-justifying arguments to deny the most basic human intuitions so long as it contributed to that result, and so on - this is precisely what we see happening in the culture today with the materialists.

Ignorance - not knowing - is so painful for some people that psychologically they have a need to invent causes for things that demand none, to invent long and convoluted arguments for ignoring the most obvious truths in order to fool themselves into believing that doing so conceals their ignorance, and so on - this is precisely what we see happening in the culture today with the fundamentalists.

See what I did there? Sneering is not an argument, and extensive reading in philosophy is not the only way to acquire a sensitive bullshit detector.
posted by flabdablet at 4:59 PM on June 24


Ignorance - not knowing - is so painful for some people that psychologically they have a need to invent causes for things that demand none, to invent long and convoluted arguments for ignoring the most obvious truths in order to fool themselves into believing that doing so conceals their ignorance, and so on - this is precisely what we see happening in the culture today with the fundamentalists.

Honestly, I think this kind of applies to materialist explanations of consciousness. The obvious truth is we can't get much of an explanation of consciousness out of molecules flying around and interacting with each other, no matter how we invoke information theory, emergentism, 'forrest through the trees,' etc.
posted by Golden Eternity at 5:23 PM on June 24 [1 favorite]


The obvious truth is we can't get much of an explanation of consciousness out of molecules flying around and interacting with each other -- Yet.

FTFY.

Seriously, consciousness is one of the few remaining of a long, long line of mysteries that Very Serious People thought Couldn't Possibly Be Explained by Mere Material Stuff flying around. The planets were once thought as mysterious as thought; today we drive radio controlled cars around on those with suitable surfaces and drop probes into the atmosphere of Jupiter and fly others through the rings of Saturn. The stars were once considered as mysterious as thought; now we know that they're made of matter, we know their life cycles, and we even know some of them are circled by planets.

The thing is, we already know enough about how the brain works to know that it encodes a truly vast amount of information in the form of synaptic connections, and that can do a vast amount of parallel processing in the form of synaptic firing. But as vast as those numbers are they don't look nearly as vast today as they did in, say, 1980, and the idea that there is something unattainable there is just as reasonable as the ideas that we will never be able to know what the planets and stars are made of.

The brain is a marvelous but very much material electrochemical machine and consciousness is an emergent property of its chaotic operation. (I use the term chaos here as math, not philosophy.) To suppose otherwise is to put yourself in the camp that thought the planets, stars, and phenomena light light and energy were mysteries from the gods instead of things that could be described and understood.
posted by localroger at 7:15 PM on June 24 [1 favorite]


we can't get much of an explanation of consciousness out of molecules flying around and interacting with each other

That's the same kind of dismissive language a creationist would use: an eye could never have arisen by random chance ergo evolution is bunk.

The interesting questions are all about how those molecules interact with one another. I strongly suspect that most of them will need to be doing something quite unlike "flying around".
posted by flabdablet at 7:50 PM on June 24


You don't get to label an imagined thing a counterfactual simply because it isn't factual; you must specify some set of conditions under which it could possibly be factual. And in the particular case of the existence-free universe, no such conditions exist or indeed could possibly exist; the concept is analytically self-contradictory.

For which your argument is that "this answer", meaning your response above, exists, and that that is "something," thus making the idea of a non-universe an impossibility.

Except that in the empty universe, your response would NOT exist. Nor would any fact of the matter about the empty universe - no, not even its emptiness. For there would be nothing to make the statement of.

You've confused this universe with the counterfactual one. That's why I'm suggesting you don't understand the concept of a counterfactual.

It's like I said "suppose there was a universe in which Hitler died as a child." In response you tell me that's not a real counterfactual and in fact hitler-child-death universe is analytically impossible because hitler did live to adulthood in this world.

See what fallacious nonsense that is?
posted by shivohum at 9:44 PM on June 24


Honestly, I think this kind of applies to materialist explanations of consciousness. The obvious truth is we can't get much of an explanation of consciousness out of molecules flying around and interacting with each other, no matter how we invoke information theory, emergentism, 'forrest through the trees,' etc.

Absolutely, but good luck trying to convince them of that. They're as fixed in their Science Faith as Ayn Rand robots are in the power of the free market to solve everything.
posted by shivohum at 9:50 PM on June 24


They're as fixed in their Science Faith as Ayn Rand robots are in the power of the free market to solve everything.

If Ayn Rand had cured polio or landed a rover on Mars I'd be a lot more impressed with whatever beliefs enabled her to do that. As far as I know all her beliefs ever let her accomplish, other than collecting groupies, was giving herself lung cancer.
posted by localroger at 5:15 AM on June 25 [1 favorite]


Speaking of Ayn Rand:
Whatever the degree of your knowledge, these two—existence and consciousness—are axioms you cannot escape, these two are the irreducible primaries implied in any action you undertake, in any part of your knowledge and in its sum, from the first ray of light you perceive at the start of your life to the widest erudition you might acquire at its end. Whether you know the shape of a pebble or the structure of a solar system, the axioms remain the same: that it exists and that you know it . . . Existence is Identity, Consciousness is Identification.
Shit! If I seem to be agreeing with Rand to some extent, I must be wrong. Hopefully I'm taking her out of context.
posted by Golden Eternity at 7:05 AM on June 25 [1 favorite]


...an eye could never have arisen by random chance ergo evolution is bunk.

The interesting questions are all about how those molecules interact with one another. I strongly suspect that most of them will need to be doing something quite unlike "flying around".


With evolution we have mathematics to show how very complicated structures can grow out of simpler pieces. If we stick with a materialist, mechanistic definition of life, I don't think there is the same problem.

The problem with a conscious experience like color or pain is they can't be put in physical terms to begin with. You could say color equals wavelength of light, or pain equals a behavior. This is clearly wrong. Wavelength is measured in meters, not redness. Pain is an actual feeling, not just a behavior. We can have a lot of math that groups molecules together and shows causal interactions between them, but it won't be able to produce actual "red" or "pain," since these things can't be reduced to position, momentum, volume, kinetic energy, or other physical quantities.
posted by Golden Eternity at 7:31 AM on June 25


We can have a lot of math that groups molecules together and shows causal interactions between them, but it won't be able to produce actual "red" or "pain," since these things can't be reduced to position, momentum, volume, kinetic energy, or other physical quantities.

They reduce to the firing of specific neurons, most likely in the Thalamus. I like Erich Harth's work on that particular topic.

I keep bringing up astronomy because it's very striking how you don't have to go very far in the past to see a lot of very smart people putting a lot of effort into understanding what's out there and being hopelessly and embarrassingly parochially wrong about everything. They were wrong about the heavens and the aether. They were wrong about the humours and elements and the transitions undergone by materials in the course of chemical reactions. They were wrong about the laws of motion and momentum and inertia. They were wrong about the nature of gases and the atmosphere. They were wrong about what causes disease. They were wrong about the difference between living and dead matter.

We have only had fast MRI, roughly the equivalent of Galileo's first telescope for observing the living brain in action, for thirty years. In that time we have proven that a hell of a lot of what the brain does is expressed physically by neural firing. We have proven that specific areas of the brain are active during specific types of mental activity. Close to the motor and sensory humonculi we've even been able to observe individual neurons performing specific functions, such as staging to perform a physical movement.

We know from people who sustain injuries that this physical expression is necessary for many specific types of mental function, including some that would seem indistinguishable from any function of "consciousness." If a brain injury can cause you, as the title of the book said, to mistake your wife for a hat, what the hell is this numinous duality stuff doing? Because if it had anything to do with me as I conceive of myself then it would be immune to such physical manipulations causing me to mistake my loved ones for inanimate objects. But that happens, and stuff like that happens with great regularity and repeatability.

So we don't know exactly how the brain works, how it teases concepts like "red" out of the sensory detection of long-wavelength light. But we know it's a hell of a complicated machine, though it gets less intimidating by the decade as our own technology improves.

If I seem convinced of something unknowable, it's only because the lesson of history suggests it isn't unknowable, just unknown, and that it won't stay unknown forever. And just as the planets aren't mysterious cosmic forces but just other worlds where we can drive radio controlled cars around, the mind will turn out to be patterns of information we can understand, duplicate, emulate, and enhance. I can't necessarily know that -- not yet at least -- but I know where I'm putting my money when the wagers are taken.
posted by localroger at 9:42 AM on June 25 [1 favorite]


Looking back at my last couple of responses I think I was a little unnecessarily hostile. My apologies. Anyway, it's been interesting.
posted by shivohum at 1:35 PM on June 25




You've confused this universe with the counterfactual one. That's why I'm suggesting you don't understand the concept of a counterfactual.

It's like I said "suppose there was a universe in which Hitler died as a child." In response you tell me that's not a real counterfactual and in fact hitler-child-death universe is analytically impossible because hitler did live to adulthood in this world.

See what fallacious nonsense that is?


You've suggested to me that I should perhaps read more widely, so I shall return the favour by suggesting to you that you should perhaps make some attempt to understand what you read, as opposed to merely skimming it for talking points.

They're as fixed in their Science Faith as Ayn Rand robots are in the power of the free market to solve everything.

Fair play to Him, your God.
posted by flabdablet at 8:59 PM on June 25


They reduce to the firing of specific neurons, most likely in the Thalamus. I like Erich Harth's work on that particular topic.

First of all, you can't say conscious experience reduces to a physical process in this way because they belong to different ontological categories. Neurons firing are not red they are grey; a red color doesn't have any neurons in it no matter how closely you look. But even still a correspondence theory between neuron firing and color, pain, or other conscious experience doesn't work until you've put together a coherent theory that shows precisely how neurons firing corresponds to conscious experience. As I understand it, the brain is still basically understood as a Hopfield network. I think you could come pretty close to describing the entire network completely with two vectors, and a learning formula. A firing vector that indicates which neurons (or inputs) are firing at any instance in time, and another vector that shows the weights between all neurons and inputs and the firing thresholds:

O1 = sum(C1,2*I2 + C1,5*I5 + C1,8*I8 .....) > thresh1
O2 = sum(C2,1*I2 + C2,7*I7 + C2,12*I12 .....) > thresh2
.
.
.

The learning formula would adjust connection weights based on previous firing states. Obviously brain theory is already way more complicated than this, and synapses can't be described by simple weights, but are extremely complicated, nonlinear processes that are a function of neurotransmitter content and whatever.

Anyway, how do you get the color red or the pain of a sunburn out of this formula? Red itself is not hugely complicated, it is just what we experience when we see certain wavelengths of light. I don't see why someone couldn't take a shot at writing down the neuron firing vector for red and show how it is distinguishable from the vectors for pain or sound. To me it doesn't make sense, they are both just numbers representing inherently exactly the same physical process - neurons firing. Nothing has explained how vastly different experiences, let alone any experiences, could come out of this. I don't think the Hopfield network is capable of showing a correspondence between neuron firing and conscious experience that makes any sense. I know absolutely nothing about it, but I think the Hopfield netwok may just show how neuron firing is coordinated ( how the brain gets the appropriate neurons to fire "together") and there has to be something else besides just "firing" going on that corresponds properly with conscious experience. Maybe Penrose and Hameroff were on to something. Maybe Tegmark is seeing a bigger problem that needs to be solved. Who knows.

I keep bringing up astronomy because it's very striking how you don't have to go very far in the past to see a lot of very smart people putting a lot of effort into understanding what's out there and being hopelessly and embarrassingly parochially wrong about everything. They were wrong about the heavens and the aether. They were wrong about the humours and elements and the transitions undergone by materials in the course of chemical reactions. ...


And that's what makes it even more astonishing to me that they could now be so wrong in saying that neuron firing reduces consciousness. The better scientific position, if you ask me, is not to pretend to say anything that serious about consciousness at all yet, but to willingly engage in wild speculation and crazy theories and encourage others to do the same.

We have only had fast MRI, roughly the equivalent of Galileo's first telescope for observing the living brain in action, for thirty years. In that time we have proven that a hell of a lot of what the brain does is expressed physically by neural firing. We have proven that specific areas of the brain are active during specific types of mental activity.


In a way this is like saying the church had proven the sun orbited the earth by watching the sun orbit the earth. It's a terrible example, but imagine we left a television with a primitive tribe somewhere in the Amazon that knew nothing about electronics or physics. They opened it up started trying to understand how it worked, and they noticed that when they turned up the volume the audio power amplifiers got hot, and when they picture came up, the video processor got warm. So some tribesman declared they fully understood how the television produced video and audio. Audio produced by the television is expressed as heat in one section of the board, and video produced by the television is expressed as heat in a different section. Thus, video and audio reduces to heat. But some of the others are like, "wait, what? I think you've had too much Peyote my friend; maybe we actually don't understand much of what's going on here." MRI and brain lesion studies are great, but I don't think it gets us that close to an actual "theory" of consciousness.
posted by Golden Eternity at 9:24 AM on June 26


To me it doesn't make sense, they are ... just numbers representing inherently exactly the same physical process - neurons firing. Nothing has explained how vastly different experiences, let alone any experiences, could come out of this.

You and I have gone over a fair bit of this ground before, so I won't rehash that. But localroger alludes to a point we haven't covered much, which is that explaining how something happens is related, but not identical to, showing that it happens.

If it can be shown reliably and repeatedly that interference with certain neural structures either causes or prevents their owner having certain experiences, it becomes reasonable to assert that those experiences are the activity that is or would otherwise be going on in those structures. Note well: the claim to be tested is not that the external observer's measurements of that activity are the experimental subject's experiences; it's that the activity itself is the experimental subject's experience as it appears from the outside.

The objection that the neurons whose current firing pattern is characteristic of their owner's recall or recognition of a red thing are not themselves red strikes me as a herring of that very colour; a kind of category-error-driven version of the old saw about the blind men arguing furiously over the true nature of the elephant, if you will.

You'd probably be interested in a piece I ran across in the New Scientist the other day, about a research group that's doing interesting work on artificially re-injecting pre-recorded memories back into brains whose own recall processes are temporarily suppressed using drugs. The basic approach is to use very fine electrode arrays along with the kind of multiple-input multiple-output signal processing math more often seen used in multiple-antenna WiFi routers.

Once this kind of thing becomes safe enough to be done ethically with subjects the fact of whose consciousness is uncontroversial, we might move into some very interesting times.
posted by flabdablet at 10:04 AM on June 26 [1 favorite]


New Scientist, as is its way, gets a little carried away and has a bit of a froth and flutter about moving memories from one brain to another and then trying to work out who exactly the resulting composite personality is.

Personally I'd be quite surprised to find that any two human brains would be interoperable enough to make this an issue. Recording and playback are all very well, but as this very thread demonstrates, we have trouble making each other understood even when using a language whose entire purpose is information sharing. The chance of being able to share memories directly, given independently-evolved and therefore idiosyncratic experience ↔ firing pattern mappings, strikes me as probably very slim.

But if there were indeed some kind of cunningly trainable DSP that could be pressed into service for that as well? Hmmmm.
posted by flabdablet at 11:32 AM on June 26


If it can be shown reliably and repeatedly that interference with certain neural structures either causes or prevents their owner having certain experiences, it becomes reasonable to assert that those experiences are the activity that is or would otherwise be going on in those structures.

No it doesn't. As in my example, when IC's do anything they emit heat and consume power electrically. However a particular IC could be a wifi transceiver, an audio processor, a video processor, etc. Emitting heat and consuming power is not identical with wifi transceiving or audio/video processing just as neurons firing is not identical with conscious experience. If you were to look at the pattern of current draw on the power supplies, in all cases you would probably see big spikes during clock transitions. This would likewise tell us nothing about what the chip is actually doing. Current spiking is not identical with the function of the IC.

the activity itself is the experimental subject's experience as it appears from the outside.

Again, it seems to me it is not appropriate to use identity here. I think it is more correct to say, "the activity is observed from the outside during the experimental subject's experiences." The category error I am seeing is when we say A = B, and A and B are from different categories (ontological or epistemic or whatever) but I don't actually know the philosophy that well.


The objection that the neurons whose current firing pattern is characteristic of their owner's recall or recognition of a red thing are not themselves red strikes me as a herring of that very colour


This wording is a bit strange to me. When I am seeing a color I am "seeing red" not "recognition as red." Seeing something may be one way we recognize it as different then something else but recognizing != seeing.

interesting work on artificially re-injecting pre-recorded memories

I didn't find New Scientist article, but it seems to me what this is doing is getting the right neurons to fire together to trigger a memory. I don't think it is recording the actual memory, and I suspect there may be much more happening in the hippocampus and in the individual neurons and synapses when an actual memory experience is occurring, and this is just kicking off a process that is inherent in, or encoded in, this area of the brain. It hasn't actually recorded the memory and injected it, it is just triggering the process which perhaps itself contains more information than the signal transmitted. MIMO, as I understand it, is a way that an individual transmitter and receiver can communicate in the presence of many other tx-rx's using the same frequency band due to fading.

I do think brain prosthetic or synthetic brain cells are a valid way to prove a theory of consciousness, whereas the Turing test doesn't really suffice. If color were reducible to a neuron firing vector, we should be able to determine what differentiates this vector from a vector that is identical with a sound or pain, and should be able to say if more primary colors are possible or if there is something about the structure of the vector that only allows the colors we are able to see, etc. Maybe someday you could actually "wire" in the new firing pattern and cause someone to experience new things. But it seems clear to me, that a single vector space of neurons firing is not capable of doing this.
posted by Golden Eternity at 12:18 PM on June 26


When I am seeing a color I am "seeing red" not "recognition as red."

This is exactly what Dr. Harth's theory, which I mentioned upthread, addresses. Recognition is the operation of a feedback mechanism in the Thalamus which sharpens patterns found in sensory input toward stored patterns from the cortex. His candidate algorithm makes many of the same perceptual mistakes in simulation which actual humans and animals make.

In this model consciousness isn't the firing of "a" neuron or a "vector" (whatever the hell that is supposed to mean) but the tightening of a feedback loop which balances competing potential associations. "Seeing red" would be tightening of the feedback to maximize the return from the pattern detector which has been programmed to simply fire on exposure to the color red.

It's by no means a whole theory of consciousness, but it is definitely a theory of the corner you are picking on here.
posted by localroger at 12:47 PM on June 26


If it can be shown reliably and repeatedly that interference with certain neural structures either causes or prevents their owner having certain experiences, it becomes reasonable to assert that those experiences are the activity that is or would otherwise be going on in those structures.

No it doesn't. As in my example, when IC's do anything they emit heat and consume power electrically.


On the other hand, if removing an IC reliably cut the audio, particularly if it did so on a variety of models using a mix of chips of which that one was the only common element, it would be quite reasonable to conclude that that chip has something to do with audio processing.
posted by localroger at 12:49 PM on June 26 [1 favorite]


If you were to look at the pattern of current draw on the power supplies, in all cases you would probably see big spikes during clock transitions. This would likewise tell us nothing about what the chip is actually doing.

"Nothing" might be overstating the case a little.
posted by flabdablet at 8:45 PM on June 26


In this model consciousness isn't the firing of "a" neuron or a "vector" (whatever the hell that is supposed to mean) but the tightening of a feedback loop which balances competing potential associations. "Seeing red" would be tightening of the feedback to maximize the return from the pattern detector which has been programmed to simply fire on exposure to the color red.

By vector I meant a list of the states all of the neurons and sensory nerve inputs in the body (either firing or not firing). So if there are 100B neurons, it would be a list of 100B 1's or 0's. Each position in the list would correspond with an individual neuron. A 1 would indicate that neuron is firing, a 0 would indicate it is not. If neuron firing patterns are conscious experiences then this vector is conscious experience. If a second time dimension is added, the array would be an entire lifetime of conscious experience. By my calculation ten seconds of human conscious experience would then fit on a 2TB hard drive (100B neurons at 10msec resolution). I guess the hard drive would be experiencing this stuff all together simultaneously for a lifetime. If pain is just an information pattern, it may be unethical to put anything but noise in any storage medium until we know the exact information patterns of intolerable pain.

In a Hopfield network you could see feedback loops by monitoring oscillation patterns in the nodes or neurons over time, and could map out the feedback in the connections between nodes. Thus the network and all of the information it contains can be described by a mathematical structure (I guess a mathematical graph might be the correct term) that could describe the behavior of a corresponding physical structure.

I'm not quite sure what you mean by tightening of a feedback loop. What you describe sounds a little bit like Dennett's multiple drafts - which sounds good to me but doesn't explain consciousness. It seems to me the same problems exist. Why is one feedback network the conscious experience of color, and a different feedback network a different experience? What differentiates them? I would think they actually look very much the same topologically and mathematically. How could the experiences be so different if they just are these very similar networks? Why aren't feed-forward networks consciousness? I suspect such a theory would fail to show an isomorphism with conscious experience.

"Nothing" might be overstating the case a little.

Yeah that was an overstatement indeed. Even still, the information present the DPA (pdf) could not be used without already understanding a lot about RSA and the microprocessor being analyzed I would think. And similarly, I think there is a lot more to be understood about the brain before we can say too much about MRI patterns.
posted by Golden Eternity at 10:22 PM on June 26


By my calculation ten seconds of human conscious experience would then fit on a 2TB hard drive (100B neurons at 10msec resolution). I guess the hard drive would be experiencing this stuff all together simultaneously for a lifetime.

This is the same kind of persistent category error I perceive you making every time you and I discuss this stuff.

What's on your hypothetical hard drive is information about the activity of the brain that was measured to collect that information. It's not the activity itself, any more than a stored debug trace of a running executable is the operation of that same executable or a photograph is the scene that was photographed or the pits on a CD are the music.

I'm guessing that the difficulty you're having with trying to conceptualize that-which-you-think-of-as-consciousness springs from the same root as the difficulty I have with trying to conceptualize that-which-I-think-of-as-time. Your hypothetical hard-drive brain-state snapshot movie is a special case of one potentially useful way to think about reality as a whole, the block universe, which - given that Relativity both works so well and denies the universality of simultaneity - strikes me as the least erroneous way to conceive of that whole as a whole.

However, trying to reconcile the kind of extreme determinism inherent in the block universe with other important concepts that are not inherent in it - particularly, motion and change - is generally a pretty unsatisfactory experience.

If I'm right that you and I are both actually chewing on different ends of the same chunk of conceptual gristle, it seems to me that your inquiry into what it is that's actually experiencing the present is the flip side of my own desire to understand what motion means with respect to a block universe. And unfortunately I know of no better answer to that, and no better approach to answering that, than to say that it is what it is and observe that every description of it must to some extent just be unsatisfactory.

I think that's inherent in the nature of explanation, description and understanding itself, which I see as analogous to a compression function that has to be lossy: there is simply no way I can expect the entire universe to fit inside my own mind, much as that sounds like it might be fun.

I was convinced that it had done, for a while, and it was fun. Most fun I've ever had, in fact. But I was mad at the time, and that turned out not to be a sustainable way to live :-)
posted by flabdablet at 11:25 PM on June 26


By vector I meant a list of the states all of the neurons and sensory nerve inputs in the body (either firing or not firing).

This is something of an oversimplification but I'd be comfortable speculating that "what I am thinking / perceiving" at any given moment is encoded by this "vector" of yours applied over the thalamus, but not the cerebral cortex. There is a 1:1 correspondence between thalamic areas (though they are packed 3D instead of across a 2D surface) and cortical areas.

I'm not quite sure what you mean by tightening of a feedback loop.

It's kind of hard to explain without referring to Harth's theories. Unfortunately The Creative Loop is long out of print and his scholarly articles are very hard to understand without an underpinning in the math and neuroscience he uses.

Basically, Harth believed consciousness was an ongoing dynamic interaction between the thalamus and the cortex, with the cortex recognizing and encoding patterns (based both on sensory input and other recognized patterns of both higher and lower abstraction) and the thalamus using dynamic feedback to recognize those patterns as they might appear in sensory input or in more abstract expressions based on combinations of sensory input at higher abstractions. He called this dynamic interaction "the creative loop," and called his popular book about it that.

Harth's system is most definitely not backpropagation; in addition to being based explicitly on the thalamic neural wiring which Harth was investigating, it had significant mathematical advantages in avoiding "getting stuck." Thermal noise is an important component of that, a feature not a bug.

The thing I like about Harth's theory is that, in addition to demonstrably making the same kinds of mistakes biological systems do at low levels of abstraction, it suggests a mechanism by which the process of "thought" occurs as patterns are dynamically balanced against the results of sensory input and memory to explore potential patterns of outward activity, which might end up getting sharpened into signals sent to the muscles.

An interesting detail which often gets overlooked is that there is no sensory feedback to the cerebral cortex; neither pleasure nor pain nor anything other positive or negative feedback signal (except for some smells, since in the distant evolutionary past the cortex was an olfactory organ) are routed there. So as important as the cortex obviously is for encoding functionality that enables us to be ourselves, whatever actually drives us to seek some things and avoid others and to build complex models to help us do so must live elsewhere. Harth's theory does a good job of explaining that.
posted by localroger at 5:36 AM on June 27


This is the same kind of persistent category error I perceive you making every time you and I discuss this stuff.

What's on your hypothetical hard drive is information about the activity of the brain that was measured to collect that information. It's not the activity itself, any more than a stored debug trace of a running executable is the operation of that same executable or a photograph is the scene that was photographed or the pits on a CD are the music.


If we have the complete information about an activity such that we can reproduce the activity in its entirety, possibly using a different medium, then we can't expect doing the activity itself to produce something that is not already described by the information we have. As I've stated, there is nothing in the description of the Hopsfield netwok to indicate what is color experience, pain experience, or to indicate that there is any experience at all when active or not active.

It is absurd to think the hypothetical hard drive is conscious; on this we agree. The point I want to make is the same absurdity applies to the assertion that conscious experience is the activity going on in neural structures if that activity is assumed to be the functioning of something like Hopfield network based on the spiking neuron model or the activity going on in a computer, or anything that is just doing computations on or manipulating meaningless symbols, often described as 'information processing.'

It would be conceivable to reproduce the full neural activity (as a simplified Hopfield network) in the HD with the addition of a small micro-controller with a SATA interface or whatever. In addition to the 'neural firing vector' I described, the entire graph describing the connection between each node is added. This is a list of neuron pairs with a number indicating the strength of the connection between them (unconnected neurons can be listed as 0 or left off of the list). An additional vector is added which lists the firing threshold for each neuron that the sum of all inputs multiplied by their strengths must cross for the neuron to fire, and another vector is added to store the previous firing vector.

A program for updating the firing vector is then added. This is a very simple program that just copies the current firing vector to the 'previous' firing vector memory, then updates each node in the firing vector using the connection graph, threshold vector, and previous firing vector.

For psychosomatic pain or internally generated thought this should be sufficient. For other sensory experiences, we could store a recording of inputs from all of the retina cells and other sensory input cells connected to the nervous system. So now we may be talking about a few thousand of TB's of data. Still not unimaginable for today's technology.

I think the same thought experiment shows the idea that the activity of the micro-controller + HD is conscious experience to be just as absurd. 99% of all of the HD is lying just as dormant as before 99% of the time. When there is activity it is just flipping bits in the neuron firing vector. Each 'neuron' or bit in memory is basically disconnected from any other bits. It is only by virtue of the HD controller that it can be considered a bit, and only by virtue of the program that it has any connection to any other bits. The micro-controller itself only requires a few instructions and a few registers for storing addresses and incoming data and a simple ALU that only needs to do multiplication, addition, and comparison. It only computes one neuron output at a time. The other 100B 'neurons' are just lying dormant. When we looked at the HD before we could see it is just a meaningless pattern of magnetization on a disk just sitting there. Now it is still a dormant meaningless pattern on a disk, except 1% of it is being flipped around. We don't even half to use a magnetic disk, we could put all of the bits on punch cards or bar codes and spread them out all over the universe. How can this activity explain conscious experiences like color, pain etc? We can't even pick out which bits are color and which bits are pain in any meaningful way.

Actually, the same goes for modern AI, in which the CPU is only operating on a few memory locations at one time, and all of the other 'information' is just lying dormant in memory cells 99% of the time.

If you were to implement the Hopsfield network in IC's so it could operate in real time by implementing each neuron as a series of programmable gain amplifiers connected to a summation amplifier and a programmable comparator, the situation is a little different but the same thought experiment pertains, imo. It is still just a lot of flipping meaningless bits (or adjusting meaningless quantities in an analog scenario), and most of the "information" is lying dormant most of the time with no active connection to anything else. Only by looking at past history can you tell how a neuron firing is connected to other neurons that fired previously - everything is separated by space and time according to classical physics so it is really hard to see how you can even look at the network as a whole individual thing capable of individual conscious experiences like color. When something is "active" it is only a small portion of the structure, and the activity doesn't suggest consciousness in any way.
posted by Golden Eternity at 10:05 PM on June 28


When we looked at the HD before we could see it is just a meaningless pattern of magnetization on a disk just sitting there.

The thing is this is true of any general purpose computer if you look at it without understanding of the I/O and coding that the patterns in memory are based on. It's also true of actual biological brains, where most of the neurons embedded in what looks like a "meaningless" tangle of white matter spend most of their time doing nothing at all.

You also seem to be missing a point which is also missing in a lot of models, which is that dendritic growth and synapse formation are probably the method by which new patterns are encoded, and we're very sketchy on the details of how that occurs. But with more knowledge we could certainly simulate that too.

Anyway I fail to see how your thought experiment doesn't exclude the actual brain. All of your objections apply to it too. Yet, we readily call what happens when the brain is doing its thing consciousness.
posted by localroger at 5:28 AM on June 29 [1 favorite]



The thing is this is true of any general purpose computer if you look at it without understanding of the I/O and coding that the patterns in memory are based on.


If we are trying to explain how a computer itself becomes conscious on its own, how we look at it can't be part of what makes it conscious or not. There are patterns all over the place in the world. Without me looking at it with prior knowledge of what it is, the only thing that distinguishes the pattern on a hard drive from patterns on my carpet or anywhere else in the world is that a CPU may or may not at some point connect to it and read the pattern and then change the pattern or do something else through an I/O and the past history of how the pattern was formed initially. If consciousness is activity, we can't use the dormant inactive patterns on the hard drive as part of conscious experience. The only activity is in the CPU and in the few bits that are actually being read or changed on the HD and this activity has no real connection to the other bits in memory (the other bits could be put on punch cards or 2d bar-codes instead of a HD and spread all over the universe). The actual activity in the CPU at any given time is manipulation of a few bits using boolean algebra and it is hard to see how this activity is conscious experience like vision, sound, etc.

It's also true of actual biological brains, where most of the neurons embedded in what looks like a "meaningless" tangle of white matter spend most of their time doing nothing at all.
...
Anyway I fail to see how your thought experiment doesn't exclude the actual brain. All of your objections apply to it too. Yet, we readily call what happens when the brain is doing its thing consciousness.


There is a difference here. The tangled white matter is itself capable of becoming active simultaneously. Some of what it does is obviously involved in conscious experience. My argument is that whatever it is doing that directly relates to consciousness, it is not just computation on meaningless symbols or 'information processing' as in a Hopfield network, which as I understand it is the basis of a lot of brain theory. This can be seen by doing a thought experiment where the Hopfield network or any other computational model could be instantiated in a digital computer and seeing that it is absurd to think the computer is conscious in these scenarios because 99.999% of the computer is just inactive empty patterns sitting in memory not connected to anything else in any real way, and the small amount of activity that is present at any given time is just a few bit manipulations. I guess one could argue that consciousness is still computational but architecture dependent. Computations must be executed such that many bits are being changed simultaneously instead of one at a time for consciousness to occur; I still see similar problems with this. I tend to think that when a conscious experience occurs the physical activity involved must be connected together nonlocally somehow. This isn't the case with a computer.

Not that the Hopsfield network won't continue to be helpful in explaining a lot of what is happening in the brain, but I'm thinking it controls how conscious processes are triggered and sustained, but it doesn't explain consciousness itself.
posted by Golden Eternity at 1:25 PM on June 29


The tangled white matter is itself capable of becoming active simultaneously.

On what do you base this? My understanding is that it is just biochemical wires.

This can be seen by doing a thought experiment where the Hopfield network or any other computational model could be instantiated in a digital computer and seeing that it is absurd to think the computer is conscious in these scenarios because 99.999% of the computer is just inactive empty patterns sitting in memory not connected to anything else in any real way, and the small amount of activity that is present at any given time is just a few bit manipulations.

Again, how is this different from a real brain? Is it "absurd to think the brain is conscious" because 99.999% of its neurons are inactive at any given time?

Of course memory patterns "just sitting in memory" ARE connected to something if there is a something waiting to activate them, whether they are being activated at the moment or not, and whether that activation is through a dendrite or over a routed network. My computer currently has a connection to the server at metafilter.com, but 99.9% of the time neither it nor the metafilter server are actually exchanging information. Yet the result of this interaction is a conversation.

You are not describing any real differences here. You are using words to encode a biochauvanism which is not justified if you look at what neurons and dendrites actually do compared to what CPU's and memory actually do.
posted by localroger at 3:22 PM on June 29


how we look at it can't be part of what makes it conscious or not

No, that would be self-determined.

However, the way we look at a system is crucial to our opinion on whether it is conscious or not. For example, I'm personally of the opinion that the entity that signs its messages "Golden Eternity" probably is one such system.

A conscious entity whose internal organization is dissimilar enough from our own that we don't instantly empathize with it might or might not be concerned with how we perceive it, or even (given that it may well be working on a timescale far removed from our own; Turing equivalence says nothing at all about processing speed) whether we even notice it.
posted by flabdablet at 6:13 AM on June 30


Imo, this blog has a great discussion of IIT (including the comment section): Consciousness Wars: Tononi-Koch versus Searle

On what do you base this? My understanding is that it is just biochemical wires.

I'm not sure what you meant by tangled white matter. I assumed you meant dendrites, axons, synapses, neurons, neurotransmitters and the whole lot.
Again, how is this different from a real brain? Is it "absurd to think the brain is conscious" because 99.999% of its neurons are inactive at any given time?

Good point; that was very muddled. Neurons not involved in a particular experience being inactive during the experience I do not see as a problem. The claim is that consciousness is the activity of neurons firing in the brain and that the neuron firing is equivalent to computation or information processing. If this is true, then for a given experience (let's say I hear tires sceeching for one second) the activity that is this experience must be active for the one second. This is tautological based on the language we are using.

The point I'm ultimately trying to make is that 'neurons firing' is not a sufficient explanation for consciousness even in a biological brain; the Micro+HD thought experiment reveals this more clearly to me. In the micro+HD case it is really only a single, or actually a portion of a single neuron that is ever active at a single time. In the case of a biological brain some pattern of neurons (millions let's say) fire in a similar pattern for the one second. But when can we say the neurons are actually active? Perhaps only on the rising and falling edges of the firing pulse? It is doubtful they are all firing together exactly simultaneously. Because of the way neurons are separated by space and time according to classical physics they effectively are each in their own worlds totally disconnected from each other, except by a causal link. This part I need to think about more, but the cause of an individual neuron firing is not available at the time it is firing and is not part of its activity.

As another thought experiment regarding the problems with using causality to explain consciousness or meaning, suppose every neuron input were connected to a random noise source rather than other neurons. The neurons are then each spread light years apart all over the universe. After trillions and trillions of years, by chance alone the same one second firing pattern occurs in the millions of relevant neurons. Would the sound of screeching tires now be experienced? All of the neuron firing activity is the same so if consciousness just is neurons firing it must be. But there is no causal connection between neurons in this case; all of the random noise generators are independent.

You are using words to encode a biochauvanism which is not justified if you look at what neurons and dendrites actually do compared to what CPU's and memory actually do.

Biochauvanism seems like a very silly accusation to make, especially considering how weak and speculative any current "theory" of consciousness is, but I guess I am actually a physics-chauvinist not a biochauvinist. I suspect there is a lot we don't understand yet about what is happening in individual neurons let alone large networks of them, and an adequate explanation or description of how consciousness fits into the universe is going to require new physics.

However, the way we look at a system is crucial to our opinion on whether it is conscious or not.


Which is more reason to try to look at it in the most 'scientific' way possible. Hopefully through empirical psychology, neurology, biology, etc., we will eventually find more detailed and more meaningful neural correlates of consciousness.
posted by Golden Eternity at 10:47 PM on June 30


I'm not sure what you meant by tangled white matter.

White matter consists entirely of bundles of nerve fibres. Grey matter consists mostly of neurons and their glial support cells. That there is such a distinction is one of the more fundamental morphological features of brains.

If this is true, then for a given experience (let's say I hear tires sceeching for one second) the activity that is this experience must be active for the one second.

No, that is not how it works even in computers. You seem to have a real problem understanding the idea of an encoded representation.

The claim is that consciousness is the activity of neurons firing in the brain and that the neuron firing is equivalent to computation or information processing.

That is an oversimplification, but yeah kinda. Consciousness is the firing of neurons in a particular dynamic pattern which performs a particular type of ongoing information processing.

Your thought experiments do not tease this out. The pattern can be established at any rate, including one where only one neuron switches state at a time; it is the dynamic pattern of firing, however quickly or slowly it evolves, which we call consciousness. Your neurons being fed noise obviously are not participating in such a pattern. In the brain neurons in the thalamus only have "meaning" based on the pattern detectors to which they are hard-wired in the cortex; the pattern detectors in the cortex only have "meaning" based on the other nearby pattern detectors and thalamic inputs which they are detecting. If you take any part of the system away then it won't be conscious, and it will very obviously not be conscious if you are observing it at any scale because it won't be acting like a consciousness.
posted by localroger at 10:44 AM on July 1


White matter consists entirely of bundles of nerve fibres. Grey matter consists mostly of neurons and their glial support cells. That there is such a distinction is one of the more fundamental morphological features of brains.

In any event, it seems to me white matter activity is critical to consciousness, or at least the activity of the pulses travelling within the fibers is important. So we could say white matter is dormant when there are no signals travelling through them, but active when relevant signals are travelling through them.

No, that is not how it works even in computers. You seem to have a real problem understanding the idea of an encoded representation.

Hmm. I don't think it is trouble understanding "encoded representation," but I don't think "encoded representation" explains or is consciousness. Before it was said that certain activity (neuron firing) is consciousness. Are you saying this is indeed the case, but even though the conscious experience happens at time A, the activity that is the experience could happen at time B, even if they are separated by minutes or hours or millions of years? I don't think you can say consciousness is identical with an activity if the activity is not contemporaneous with the conscious experience.

If you take any part of the system away then it won't be conscious, and it will very obviously not be conscious if you are observing it at any scale because it won't be acting like a consciousness.

We could change the thought experiment to put all of the white and grey matter back in the body. However in every connection between a nerve input and the nerve we insert a device that feeds the nerve with random pulses, and just sinks the actual input pulses. You are right that this would not "act like consciousness" for 99.999999....% of the time. However, if we could repeat this experiment for trillions and trillions of years there is a chance that for a brief period of time all of the neuron firings would be identical to a brain working normally, and the body would act no differently. It would "act like consciousness." Would it be consciousness?
posted by Golden Eternity at 11:48 AM on July 1


I don't think you can say consciousness is identical with an activity if the activity is not contemporaneous with the conscious experience.

It is well known that conscious perception of sensory input lags the actual sensory input and is subject to a great deal of distortion. flabdablet and I have both told you that conciousness is a system, and you keep picking out bits and pieces and saying "this isn't it" and "that can't be it." Of course not because a piece of a system isn't the system. I've also said it is a dynamic system, and if the system is prevented from functioning by being disconnected or having one of its parts deactivated or whatever other silly thought experiment you want to throw at it, then it's not conscious at that point even though it is the mechanism through which consciousness is expressed. Such as when, for example, you are in dreamless sleep.

Your quest would be like asking exactly which part of a car makes it a car. The answer is that there is no one thing; it is a mix of features and functionality that add up to say a thing is a car and not a kiddie wagon or a moped or a pickup truck or a bicycle. You can pull things off a car all day long and say "Well this isn't what makes it a car, because pickup trucks have them! This isn't what makes it a car, because cars in India don't have them!" I've said at least four times that consciousness is a dynamic interaction between brain systems.

This actually supports one of your contentions, which most neuroscientists would consider quite bizarre, that "the white matter is active." In a sense it is, because the routing of messages (determined by where nerve fibres go) is much more important than the actual frequency of pulses on particular ones.

However, if we could repeat this experiment for trillions and trillions of years

If you got monkeys to type gibberish for a trillion years you'd eventually get the works of Shakespeare, but you would not be able to find them and the result would not be a library. And such arguments have less than nothing to do with consciousness, which is its own thing.
posted by localroger at 4:33 PM on July 1 [1 favorite]


Strictly reductionist analysis of consciousness: a metaphor
posted by flabdablet at 10:07 PM on July 1


« Older Dan Savage, the University of Chicago, free speech...  |  Milena Sidorova is a soloist i... Newer »


This thread has been archived and is closed to new comments