[Okay, so this has just an organic extension of thinking through a variety of problems via a thought experiment posted by Eric Thomson over at the Brains blog. The dialogue takes place between an alien, Al, who has come to earth bearing news of Consciousness (or the lack of it), and a materialist philosopher, Mat, who, although playing the obligatory, Socratic role of the passive dupe, is having quite some difficulty swallowing what Al has to say. It's rough, but I do like the picture it paints, if only because it really does seem to offer a truly radical way to rethink consciousness, why we find it so difficult, as well as the very nature of philosophical thought. I haven't come up with a name for Al's position, yet, so if anyone thinks of something striking (or satirical) do let me know!]
Al: “Yes, yes, we went through this ‘conscious experience’ phase, ourselves. Nasty business. Brutish! You see, you’re still tangled in the distinction between system-intrinsic base information and the system-extrinsic composite information it makes possible. Since your primary cognitive systems have evolved to troubleshoot the latter, you lack both the information and the capacity to cognize the former. It’s yet another garden variety example of informatic parochialism combined with a classic heuristic mismatch. Had you not evolved linguistic communication, your cognitive systems would never need bump against these constraints, but alas, availability for linguistic coding means availability for cognitive troubleshooting, so you found yourself stranded with an ocean of information you could never quite explain–what you call ‘consciousness’ or ‘subjective experience.’”
Mat: “So you don’t have conscious experience?”
Al: “Good Heavens, no, my dear fellow!”
Mat: “So you don’t see that red apple, there?”
Al: “Of course I see it, but I have no conscious experience of it whatsoever.”
Mat: “But that’s impossible!”
Al: “Of course it is, for a backward brain such as your own. It’s quite quaint, actually, all this talk of things ‘out there’ and things ‘in here.’ It’s all so deliciously low res. But you’ll begin tinkering with the machinery soon enough. The heuristics that underwrite your environmental cognition are robust, there’s no doubt about that, but they are far too crude and task-specific for you to conceive your so-called ‘conscious experience’ for what it is. Someday soon you’ll see that asking what redness is makes no more sense than asking what the letter m means!”
Mat: “But redness has to be something!”
Al: “To be taken up as a troubleshooting target of your environmental cognitive systems, yes, indeed. That, my human friend, is precisely the problem. The heuristic you confuse for redness was originally designed to be utterly neglected. But as I said, rendering it available for linguistic coding made it available to your cognitive systems as well, and we find this is where the trouble typically begins. It certainly was the case with our species!”
Mat: “But it exists here and now for me! I’m bloody-well looking at it!”
Al: “I know this is difficult. Our species never resolved these problems until our philosophers began diagnosing these issues the way neurologists diagnose their patients, when they abandoned all their granular semantic commitments, all the tedious conceptual arguments, and began asking the simple question of what information was missing and why. Looking back, it all seems quite extraordinary. How many times do you need to be baffled before realizing that something is wrong with you? Leave it to philosophers to blame the symptom!
“You are still at the point were you primarily conceive of your brain as a semantic (as opposed to informatic) engine, as something that extracts ‘relevant’ information from its noisy environments, which it then processes into models of the universe, causally constructed ‘beliefs’ or ‘representations’ that take the ‘real’ as their ‘content.’ So the question of red becomes the question of servicing this cognitive mode and model, but it stubbornly refuses to cooperate with either, despite their independent intuitive ease. You have yet to appreciate the way the brain extracts and neglects information, the way, at every turn, it trades in heuristics, specialized information adapted for uptake via specialized processors adapted for specific cognitive tasks. Semantic cognition, despite the religious pretension of your logicians, is a cognitive short-cut, no different than social cognition. Rather than information as such, it deals with environmental being, with questions of what is what and what causes what, much as linguistic cognition deals with communicative meaning, with questions of what means what and what implies what.
“Now as I said, red no more possesses being than ‘m’ possesses meaning. Soon you will come to see that what you call ‘qualia’ are better categorized as ‘phenomemes,’ the combinatorial repertoire that your environmental cognitive systems uses to make determinations of being. They are ‘subexistential’ the way phonemes are ‘subsemantic.’ They seem to slip into cognitive vapour at every turn, threatening what you think are the hard won metaphysical gains of another semantic myth of yours, materialism. You find yourself confronted with a strange dilemma: either you make a fetish of their constitutive, combinatorial function and make them everything, or you stress their existential intractability say they are something radically different. But you are thinking like a philosopher when you need to think like a neuropsychiatrist.
“The question, ‘What am I bloody well looking at?’ exhausts the limits of semantic cognition for you. Within those limits, the question makes as much sense as any question could. But it is the product of a heuristic system, cognitive mechanisms whose (circumstance specific) effectiveness turn on the systematic neglect of information. So long as you take semantic cognition at its word, so long as you allow it to dictate the terms of your thinking, you will persist in confusing the informatic phenomena of smonsciousness with the semantic illusion of consciousness.”
Mat: “But semantic cognition is not heuristic!”
Al: “That’s what all heuristics say–they tend to take their neglect quite seriously, as do you, my human friend! But the matter is easily settled: tell me, in this so-called ‘conscious experience’ of yours, can you access any information regarding its neural provenance?”
Mat: “No.”
Al: “Let me guess: You just ‘see things,’ transparently as it were. Like that red apple.”
Mat: “Yes.”
Al: “Sounds like your cognitive systems are exceedingly selective to me!”
Mat: “They have to be. It would computationall–”
Al: “Intractable! I know! And evolution is a cheap, cheap date. So then, coarse-grain heuristics are quite inevitable, at least for evolved information systems such as ourselves.”
Mat: “Okay. So?”
Al: “So, heuristics are problem specific, are they not? Tell me, what should we expect from misapplications of our heuristic systems, hmm? What kind of symptoms?”
Mat: “Confusion, I suppose. Protracted controversy.”
Al: “Yes! So you recognize the bare possibility that I’m right?”
Mat: “I suppose.”
Al: “And given the miasma that characterizes the field otherwise, does this not place a premium on alternative possibilities?”
Mat: “But it’s just too much! You’re saying you’re not a subject!”
Al: “Precisely. No different than you.”
Mat: “That you experience, but you don’t have experience!”
Al: “Indeed! Indeed!”
Mat: “You don’t think you sound crazy?”
Al: “So the mad are prone to call their doctors. Look. I understand how this must all sound. If qualia don’t exist because they are ‘subexistential,’ how can they contribute to existence?
“Think of it this way. At a given moment, t1, qualia contributes, and you find yourself (quite in spite of your intellectual scruples) a naive realist, seeing things in the world. You see ‘through’ your experience. The red belongs to the apple, not you, and certainly not your brain! Subsequently, at t2, you focus your attention on the redness of the red, and suddenly you are looking ‘at’ your experience instead of through. (In a sense, instead of speaking words, you find yourself spelling them).
“The thing to remember is that this intentional ‘directing at’ that seems so obvious when attending to your attending is itself another heuristic–at best. You might even say it’s the ‘Master Heuristic.’ Nevertheless, it could, for all you know, be an abject distortion, a kind of water-stain Mary Magdelene imposed by deliberative cognition on cognition. Either way, by deliberating the existence of red, you just dropped a rock into the chipper, old boy. ‘But what is red!’ you say. ‘It has to be something!’ You say this because you have to, given that deliberative cognition possesses no information regarding its own limits. As far as its concerned, all phenomenal rocks are made of natural wood.
“So this is the dilemma my story poses for you. Semantic cognition assumes universality, so the notion that something that it says exists–surely, at the bare minimum!–does not exist sounds nonsensical. So when I say to you, information is all that matters, your Master Heuristic, utterly blind to the limits of its applicability, whirs and clicks and you say, “But surely that information must exist! Surely what makes that information informative is whether or not it is true!” And so on and so forth. And it all seems as obvious as can be (so long as you don’t ask too many questions).
“Information is systematicity. You need to see yourself philosophically the way your sciences are beginning to see you empirically: as a subsystem. You rise from your environments and pass back into them, not simply with birth and death, but with every instant of your life. There is no ‘inside,’ no ‘outside,’ just availability and applicability. Information blows through, and you are little more than a spangled waystation, a kind of informatic well, filled with coarse-grained intricacies, information severed and bled and bent to the point where you natively confuse yourselves with something other than the environments that made you, something above and apart.
“Information is the solvent that allows cognition to move beyond its low-resolution fixations. It’s not a matter of what’s ‘true’ in the old semantic sense, but rather ‘true’ in the heuristic sense, where the term is employed as a cog in the most effective cognitive machine possible. The same goes for ‘existence’ or for ‘meaning.’ These are devices. So we make our claims, use these tools according to design as much as possible, and dispose of them when they cease being effective. We help them remember their limits, chastise them when they overreach. We resign ourselves to ignorance regarding innumerable things for want of information. But we remember that the cosmos is a bottomless well of information, both in its sum and in its merest part.
“And you see, my dear, materialist friend, that you and all your philosophical comrades–all you ‘thinkers’–are actually tinkers, and the most inventive among you, engineers.
Mat: “You have some sense of humour for an alien!”
Al: “Alienation comes with the territory, I’m afraid.”
Mat: “So there’s no room for materialism in your account?”
Al: “No more than idealism. There is just no such thing as the ‘mind-body dichotomy.’ Which is to say, the mind-body heuristic possesses limited applicability.”
Mat: “Only information, huh?”
Al: “Are you not a kind of biomechanical information processing system, one with limited computational capacity and informatic access to its environments? Is this not a cornerstone tenet of your so-called ‘materialism’?”
Mat: “Yes… Of course.”
Al: “So is not the concept ‘materialism’ a kind of component device?”
Mat: “Yes, of course, bu–”
Al: “But it’s a representational device, one that takes a fundamental fact of existence as its ‘content.’”
Mat: “Exactly!”
Al: “And so the Master Heuristic, the system called semantic cognition, has its say! So let me get this straight: You are a kind of biomechanical information processing system, one with limited computational capacity and informatic access to its environments, and yet still capable, thanks to some mysterious conspiracy of causal relations, of maintaining logical relations with its environments…”
Mat: “This is what you keep evading: you go on and on as if everything is empirical, when in point of fact, scientific knowledge would be impossible without a priori knowledge derived from logic and mathematics. Incorrigible semantic knowledge.”
Al: [his four eyes fluttering] “I’m accessing the relevant information now. It would seem that this is a matter of some controversy among you humans… It seems that certain, celebrated tinkers taught that the distinction between a priori and a posteriori knowledge was artificial.”
Mat: “Yes… But, there’s always naysayers, always people bent on denying the obvious!”
Al: “Yes. Indeed. Like Galileo and Einste–”
Mat: “What are you saying?”
Al: “But of course. You must excuse me, my dear, dear human friend. I forgot how stunted your level of development is, how deeply you find yourself in the thrall of the processing and availability constraints suffered by your primate brain. You must understand that there are no such thing as logical relationships, at least not the way you conceive of them!”
Mat: “Now I know you are mad.”
Al: “You look more anxious than knowledgeable, I fear. No information system conjures or possesses logical relationships with its environments. What you call formal semantics are not ‘a priori’–oh my, your species has a pronounced weakness for narcissistic claims. Logic. Mathematics. These are natural phenomena, my friend. Only your blinkered mode of access fools you otherwise.”
Mat: “What are you talking about? Empirical knowledge is synthetic, environmental, something that can only be delivered through the senses. A priori knowledge is analytic, the product of thought alone.”
Al: “And is your brain not part of your environment?”
Mat: “Huh?”
Al: “Is your brain not part of your environment?”
Mat: “Of course it is.”
Al: “So you derive your knowledge of mathematics and logic from your environments as well.”
Mat: “No. Not at all!”
Al: “So where does it come from?”
Mat: “Nowhere, if that question is granted any sense at all. It is purely formal knowledge.”
Al: “So you access it… how?”
Mat: “As I said, via thought!”
Al: “So from your environment.”
Mat: “But it’s not environmental. It just… well… It just is.”
Al: “Symptoms, my good fellow. Remember what I said about symptoms. One thing you humans will shortly learn is that these kinds of murky, controversy-inspiring intuitions almost always indicate some kind of deliberative informatic access constraint. The painful fact is, my dear fellow, is that not one of your tinkers really knows what they are doing when they engage in logic and mathematics. Think of the way you need notation, sensory prosthetics, to anchor your intuitions! But since no information regarding the insufficiency of what little access you have is globally broadcast, you assume that you access everything you need. And then it strikes you as miraculous, the connection between the formal and the natural.
Mat: “Preposterous! What else could we require?”
Al: “Well, for one, information that would let you see your brain isn’t doing anything magical!”
Mat: “It’s not magical; it’s formal!”
Al: “Suit yourself. Would you care to know what it is you’re really doing?”
Mat: “Please. Enlighten me.”
Al: “That was sarcasm, there, wasn’t it? Wonderful! Have you ever wondered why logic and mathematics had to be discovered? It’s ‘a priori,’ you say. It’s all there ‘already,’ somewhere that’s nowhere, somehow. And yet, your access to it is restricted, like your access to environmental information, and the resulting knowledge is cumulative, like your empirical knowledge. ‘Scandal of deduction’ indeed! The irony, of course, is that you’re already sitting on your answer, insofar as you accept that you are a kind of biomechanical information processing system with finite computational capacity and limited informatic access to its environments. Some things that system discovers via system extrinsic interventions, and others via system intrinsic interventions. Your ‘formal semantics’ belongs to this latter. Not all interaction patterns are the same. Some you could say are hyperapplicable; like viruses they possess the capacity to manage systematic interventions in larger, far more complex interaction patterns. Your magical… er, formal semantics is simply the exploration of what we have long known are hyperapplicable interaction patterns.”
Mat: “But I’m not talking about ‘interaction patterns,’ I’m talking about inference structures.”
Al: “But they are the same thing, my primitive, hirsute-headed friend, only accessed via two very different channels, the one saturated with information thanks to the bounty of your environmentally-oriented perceptual systems, the other starved thanks to the penury of your brain’s in situ access to its own operations. The one ‘observational,’ thanks to the functional independence of your cognitive systems enjoy relative to your environments, the other performative, given that the interaction patterns at issue must be ‘auto-emulated’ to be discovered. The connection between the formal and the natural strikes you as miraculous because you cannot recognize they are one and the same. You cannot recognize they are one and the same because of the radical differences in informatic access and cognitive uptake.”
Mat: “But you’re reasoning as we speak, making inferences to make your case!”
Al: [sighs] “You are, like, so low-res, Dude. Why do you think the status of your formal semantics is so controversial? Surely this also speaks to a lack of information, no? When trouble-shooting environmental problems, your systems are primed for ‘informatic insufficiency’–and well they should be, given that environmental informatic under-determination kills. That blur, for all your ancestors knew, could be a leopard.
“The situation is quite different when it comes to trouble-shooting your own capacities. Whenever you attend to what you call ‘first-person’ information, sufficiency becomes your typical default assumption. This is why so many of your philosophers insisted for so long that ‘introspection’ was the most certain thing, and disparaged perception. The very thing that persuaded tinkers to doubt the reliability of the latter was its capacity to flag its own limitations, its capacity to revise its estimations as new perceptual information became available–the ability of the system to correct for its failures. In other words, what makes perception so reliable is what led your predecessors to think it unreliable, whereas what makes introspection so unreliable is the very thing that led your predecessors to think it the most reliable. No news is good news as far as assumptive sufficiency is concerned!
“Information is additive. Flagging informatic insufficiency is always a matter of providing more information. Since more information always means more metabolic expense and slower processing, the evolutionary default is to strip everything down to the ‘truth,’ you could say–to shoot first and ask questions later!”
Mat: “So there’s no such thing as the truth, now?”
Al: “Not the way you conceive it. How could there be, given finite computational resources and limited informatic ability? How could your ‘view from nowhere’ be anything other than virtual, simply another heuristic? You have packed more magic into that term ‘formal’ than you know, my bald-bodied friend.
“Why do you think your logicians and mathematicians find it impossible to complete their formal systems short of generating inconsistencies? Computation is irreflexive. No device can perform computations on its own computations as it computes. For years your tinkers have been bumping into suggestive connections between incompleteness and thermodynamics, and even now, some are beginning to suspect the illusory nature of the ‘formal,’ that calculation and computation are indeed one and the same. All that remains is for you to grasp the trick of consciousness that makes it seem otherwise: the informatic deprivations that underwrite your illusion of reflexivity, and lead you to posit the ‘formal.’
“Let me a hazard a guess: Tinkers in human computer science find themselves flummoxed with dualisms that bear an eerie resemblance to those found in your philosophical tinkering.”
Mat: “Why… Yes, as it so happens.”
Al: “I apologize. The question was rhetorical. I was accessing the relevant information as I spoke. I see here that no one knows how to connect the semantic level of programming to the implementation level of machine function. The ‘symbol grounding problem,’ some call it… Egad! Can’t you see this has been what I’ve been talking about all along?”
Mat: “I… I don’t understand.”
Al: “Once again, you admit you’re a kind of biomechanical information processing system, one with limited computational capacity and informatic access to its environments. You admit that as such a system, you suffer any number of even more severe informatic shortfalls with reference to your own operations. You admit that the numerous peculiarities you attribute to the mental and the semantic at least admit description in terms of information deficits. And yet you find it impossible to bracket your semantic intuitions, the magical belief that any biomechanical information processing system, let alone one possessing the limited computational capacity and informatic access as yours, can manufacture a kind of absolute ‘epistemic’ relation.
“Implementation, my pheromonal friend. Implementation. Implementation. Implementation. Implementation is the way, the concept you need, to maximize informatic applicability (problem-solving effectiveness) when tinkering with these problems. When you ‘program’ your computers, it’s primarily a matter of one implementation engendering another. Your ‘semantics’ is little more than the coarse-grain crossroads, a low-res cartoon compared to the informatics that you (as a so-called materialist) acknowledge underwrites it. You admit that semantics comes in an informatic box, and yet you insist on shoving that informatics into a semantic box, and you are mystified as to why nothing stays put.”
Mat: “Okay! Okay! So I’m willing to entertain the possibility that my reasoning has been distorted by the misapplication of some kind of ‘semantic stance,’ I guess. The ‘Master Heuristic,’ as you call it. Certain work in rational ecology suggests that the strategic exclusion of information often generates heuristics that are more effective problem solvers than optimization approaches. We evolve heuristics because of their computational speed and metabolic efficiency, but the hidden price we pay is limited applicability: heuristics are tools, and tools are problem specific. So how does all of this bear on the problem of conscious experience, again?”
Al: “But of course! My o my, we’ve strayed far afield, haven’t we? I have to admit, I’m overfond of preaching the virtues of Informatics to species as immature as yours. As I was saying earlier, qualia ‘exist’ relative to things existing in the world the way phonemes ‘mean’ relative to words meaning in language: in a participatory, modal sense. When you attend to qualia, they don’t offer much in the way of existential information, the way phonemes don’t offer much in the way of meaning. This is because, among other things, neither heuristic is matched for the cognitive system employed. Qualia, or phenomemes, are designed to build existence (when taken up by the appropriate cognitive system) the way phonemes are designed to build meaning (when taken up by the appropriate cognitive system).
“So, again, when a tinker submits qualia to the Master Heuristic for ‘existence processing’ they inevitably come up short. The question can’t be resolved. Phenomenality has to be something, and yet it doesn’t seem to be anything at all. You invent whole species of ‘zombies,’ whole genres of thought experiments, trying to get some purchase on the problem, to no avail.
“Consider the conceptual Necker Cube* of phenomenology and naturalism, idealism and materialism, the way your tinkers can’t decide whether to put ‘existence’ here or ‘there,’ to make it this or ‘that.’ The Master Heuristic looks ‘through’ experience, and sees the fine-grained complexities of the world. The Master Heuristic looks at experience, and sees the coarse-grained obscurities of consciousness. Both are right there, as plain as the polyp on your face. Which is fundamental? Who rules the metaphysical roost?
“But, as the informatic concept of ‘granularity’ suggests, the dichotomy is false, the result of a basic heuristic misapplication. To complicate your own materialist truism: You are a systematic assemblage of multiple biomechanical information processing systems, heuristic devices, each possessing limited computational capacity and informatic access, each adapted to a specific set of problems. If you accept this claim, as I think you must, then you should accept that the problem of ‘heuristic misapplication’ looms over all your tinker–”
Mat: “Now you’re starting to sound like Rorty–or even worse, Wittgenstein!”
Al: “Two great tinkers, yes. Indeed, their critiques reveal some of the shortcomings of the Master Heuristic, at least to the extent they considered philosophical problems in terms of performance. But by trading semantic reference for normative competence, they simply traded one inapplicable heuristic, referential truth, for another one, normative truth. I’m offering you effectiveness. Effectiveness is the concept possessing maximal applicability. Information–systematic differences making systematic differences.
“But to get back to the issue at hand: the problem of ‘heuristic misapplication’ looms over all your tinkering. Because many of these heuristics are innate as well as blind to their limited applicability, what I’m saying here will inevitably cut against a number of intuitions. But then you materialists, I gather, have long since accepted that any adequate account of consciousness will likely involve any number of counterintuitive claims.”
Mat: “But you’re saying there’s no such thing as intentionality! No meaning. No agency. No morality!”
Al: “Don’t pretend to be surprised. You materialists may not like to write about it, but our surveillance indicates that many of you have privately abandoned these things anyway.”
Mat: “Many?–maybe. But not me.”
Al: “The thing to remember is that this is simply what you’ve been all along. Some heuristics, like love, say, are preposterously coarse-grained, and yet preposterously effective all the same, so long as its scope of application is constrained. Meaning, agency, morality: these heuristics are also enormously effective, given the proper scope. The thing to remember is that ‘information is also a heuristic–only one that is particularly effective and perhaps maximally applicable, at least given the scope of the problem you call ‘consciousness.’
Mat: “But still, in the end, you’re just telling me to look at myself like a machine.”
Al: “The way your Doctor looks at you–yes! The way you claim to look at yourself already, and the way natural science has always looked at you. The only real question, my thermally tepid friend, is one of why philosophy has consistently refused to play along–even when it claims to be playing along! And this is what I’m offering you: a way to understand why the obvious strikes you as so preposterous! Heuristics. You are an assemblage of heuristics, a concatenation of devices that take informatic neglect as their cornerstone, problem-solving strategy, each of which is matched to a specific family of problems, and all of which are invisible to metacognition as well as utterly blind to one another–simply because no information regarding any of this finds its way to what you call ‘conscious cognition.’”
Mat: “So this is like Dennett’s stuff. You’re saying we need to make sure our various problem-solving stances need to be properly ‘matched,’ as you put it, to our problems.”
Al: “In a sense, yes–though an intentional heuristic like ‘stance’ is bound to hopelessly confuse things. I don’t think Dennett would want to say that you are ‘stances all the way down’ the way I’m suggesting that you are heuristics all the way down. As an intentional heuristic, ‘stance’ has limited applicability. This is why I’m offering you information: as heuristics go, it offers the highest resolution and the broadest scope. It allows you to explain the structure of other heuristics, as well as the kinds misapplications that keep your tinkers so long-bearded and well-fed. It offers you, in other words, a real way out of all your ancestral confusions.
“And most pertinent to our discussion, it lets you understand why consciousness baffles you so.”
Mat: “Yes. The million dollar question.”
Al: “You are a systematic assemblage of multiple biomechanical information processing systems, heuristic devices, each possessing limited computational capacity and informatic access, each adapted to a specific set of problems.
“In most species, cognition is built around what might be called the ‘open channel principle.’ It evolved to manage the organism’s relationship to environmental change as efficiently as possible. As such, it neglects astronomical amounts of neural and environmental information, relying on those heuristics that optimize effectiveness against metabolic cost. It’s difficult to overstate how crucial this point is: the effectiveness of your cognition turns on the strategic neglect of certain kinds of information–what might be called ‘domain neglect.’
“Take, for instance, ‘aboutness.’ You have experiences OF things rather than experiences FROM things because information regarding the latter is much less germane to survival. Only in instances of perceptual vagueness or ambiguity do you perceive FROM information, typically in indirect ways (what you might call ‘squints,’ cues to gather supplemental information). So-called transparency, in other words, is a form of strategic neglect.
“Now consider how difficult FROM information is to think in semantic terms belonging to what I’ve called your Master Heuristic. Just try to imagine the experience belonging to a perceptual system that provided knowledge FROM, so that you have experience FROM trees rather than OF them. In other words, try to imagine opaque experience. Given your neurophysiology, the best you can do is imagine OF experience that it is FROM. Transparency–the Master Heuristic–is neurophysiologically compulsory.
“And here we stumble upon the threshold of what makes consciousness so incredibly difficult to fathom: it requires accessing the very information the system neglects (either out of structural necessity or to maximize the heuristic efficiency of environmental cognition).
“Why should this be a problem? Well, its an obvious misapplication, for one. As I said earlier, using cognitive systems designed to manage extrinsic environments to assess phenomemes amounts to dropping a rock into the woodchipper. You are trying to make words out of letters, existents out of the informatic constituents of existence.
“You persist because of cognitive neglect: you simply cannot see the limits of applicability pertaining to your Master Heuristic. You find yourself in an informatic dead-end, stranded with ‘experience’ as a peculiarly intractable existent. Given the absence of information pertaining to the insufficiency of the limited information gleaned by attending to experience, you assume it’s all the information you need–or what I earlier called sufficiency. Since deliberative experience OF experience, given its neglect, seems to capture the whole of experience, any information that makes a fragment of that experience is going to seem to contradict that experience, to be talking about something else. So you run into a powerful intuitive barrier, not unlike explaining the Mona Lisa to an ant born glued to her nose.”
Mat: “So, in the picture you’re painting, there is no final picture, only a… frame, I guess you could say, systematic differences making systematic differences, or effective information. Using that, we can think outside the limitations of our heuristics, and see that consciousness as we conceive of it is a kind of perspectival illusion, a figment of informatic constraints. There literally is no such thing outside our own… informatic frame of reference?
Al: “Very good! Once you adopt information as your new Master Heuristic, the antipathy between redness and apples vanishes, along with all the other dichotomies arising out the old, semantic Master Heuristic. The information that you ‘are’ is the information that you ‘see.’ Even though your ‘experience’ will continue to be stamped by the informatic neglect characteristic of semantics, you will know better.
“You are an assemblage of heuristic devices, each possessing limited computational capacity and informatic access, each adapted to a specific set of problems–no different than any of your animal relatives. Part of what distinguishes your species, my binocular friend, is your ability to make problems, to apply your heuristics to novel situations, adapt and enlarge them if possible, even leave them behind if need be–as well as to doggedly throw them at problems they simply cannot solve.”
Mat: “So semantic and normative conceptions of knowledge can’t solve the problem of consciousness simply because the heuristics they rely on, despite the illusion of universality leveraged by neglect, are too specialized. Isn’t this just cognitive closure you’re talking about, the argument that consciousness is to us what quantum-mechanics are to chimpanzees, something simply beyond our cognitive capacity?”
Al: “The problem of cognitive applicability is quite different from that of cognitive closure, as certain tinkers among you have suggested. But the analogy to quantum mechanics is an instructive one: only when your physicists began thinking around, as opposed to through, their default heuristics, could they begin to make sense of what they were finding. This lesson is clear, one would think. Once you understand the scope of a particular heuristic, you have the means of leaving the problems it generates behind.
“But I fear the notion of relinquishing the Master Heuristic will be enormously difficult, if not impossible for many of your tinkers. For them, cognitive closure will apply, and this in turn will legislate any number of myth-preserving fancies. For those who can, those who come to understand that information precedes all the other clumsy, coarse-grained concepts you have inherited from your biology and your traditions, even existence, they will come to see that they, an assemblage of heuristic devices, are their own informatic frame of reference, a system encompassing the vast swathe of the universe they ‘know’ and continuously open to the universe they don’t.
Mat: “Excuse me for sounding dense, but you’re pretty much saying that the whole of philosophy is obsolete!”
Al: “Indeed I am, my primitive friend. But please, don’t feign any shock or surprise: a great proportion of your scientists have been saying as much for quite some time. The effectiveness of information has rendered it a social and cultural tsunami, the conceptual anchor of the most profound transformation to ever hit your species, and the most your philosophical tinkers can muster are anaemic attempts to stuff into some kind of semantic box!”
“But narcissistic idylls of your ignorance are now at an end. The mythic assumption, that you humans alone evolved some kind of monolithic, universal cognition, is entirely understandable, given the recursive blindness of your brains. But now you are beginning to understand that you are not so different from your genetic cousins, that you only seemed radically novel because of the drastic information access constraints faced by autocognition. More and more you will come to see semantics as a parochial detour forced upon you by the vagaries and exigencies of your evolution. More and more you will turn to informatics to take its place.”
Mat: “Well, to hell with that! I say.”
Al: “Deny, if you wish. The effectiveness of information is such that it will remake you, whether you believe in it or not.”
RSB is on a YA literature kick! First he references Yeerks from Animorphs, now he’s referencing the famous Red Apple scene from The Giver (and the meditation on what the ‘redness’ of the apple means…)
in other news, a very interesting piece of film criticism today:
http://www.avclub.com/articles/should-some-movies-be-taken-more-seriously-than-ot,85182/
“Sometimes you just know something is wank.
That definitive gut instinct is one of the most freeing responses a moviegoer can have, and it’s one of the drivers that keeps the movies alive as a popular art form: Anyone with a ticket or the ability to stream can watch a movie and know, in the end, just how he or she feels. Or doesn’t feel. Even ambivalence can be a kind of boldness, a catalyst that forces viewers to reckon with what they’ve just seen, to unfold its rangy layers until some semblance of meaning emerges. But even after all that wrestling, what happens if you still don’t like the movie that has been deemed a masterpiece by nearly every critic you respect (and some you don’t)? Did you just not watch the movie hard enough, or think about it long enough? Are you just not smart enough to get it?
Is the problem you?”
And:
“I can see how much care and attention Anderson put into The Master, which may be part of the problem. But the craftsmanship of Premium Rush is the offhand sort, which is not to say it’s at all accidental, but simply that it’s enlisted purely and obviously to give pleasure to the audience. (Plus, it’s increasingly difficult for filmmakers to do good work within the mainstream and, it seems, to get critics to acknowledge just how good that work is.) The picture is so deftly constructed, and the action sequences are shot so clearly, that the first time I watched it, I could sense how well-made it was, but I was having so much fun I couldn’t pick up on every detail of the filmmaking mechanics that went into it. I saw the craft in The Master—and how! It’s cod that, for me, has been hooked, cleaned, and cooked, whereas Premium Rush is still something of a mystery. I want to see it again, but if I do, am I following the siren’s song, or just the aroma of the hamburger? And is it heresy to suggest that, in the world of passionate moviegoing, perhaps they’re just different versions of the same thing?”
Scott, glad to see you are becoming more confident with your arguments. I can only imagine how anxious this all must have made you as you worked it out, knowing you were going radically against the fiercely guarded ideals of, well, pretty much all of humanity!
Thanks for this piece, its quite pleasant to finish one of your posts and feel clarity since in the past I have sometimes not been up to the task of completely understanding it all. (My fault, not yours!) I’ll have to put more of this stuff on my kindle and pay more attention in future.
Danke, James. Life as an iconoclast is easier when you can blame all your odd notions on those damn aliens!
I think you could call Al’s position the Zen – or Al-Is-Zen since at the heart of it alll, there’s nothing there: going with the flow means not fighting the inevitability of our being simply flotsam and jetsam in a universal stream of dynamically co-existent kinetic and potential information exchange.
Awesome. Pushing the interpretation in this direction definitely works, I agree. But it might be even more radical still, I’m not sure. But the whole notion of the koan, for instance, would be a way to isolate the Master Heuristic, then find ways to circumvent it…
I think you have identified the problem with most philosophical approaches. Namely, a failure to take a systems view where things and events are understood to be phenomena (in the general sense, not in the sense of mental phenomena, although they are a subcategory of the general category) arising out of the intra-action of their constituent phenomena. The term ‘intra-action’ was coined by Karen Barad in her book “Meeting the Universe Half Way”. It is meant to emphasis that there are simultaneous, mutually affecting processes involved. In this view everything is causing and is an effect of everything else.
In the big picture this is consistent with David Bohm’s idea of holomovement. The holomovement refers to a single process which evolves all at once. The metaphor often used to visualize this is a glass of water into which a drop of food colouring is dropped and then gently stirred. No part of this system evolves independently of the whole and yet phenomena (‘things’ and events) seem to occur. What appear as cause and effect relationships (things and events) are a consequence of attending to a bracketed off portion of the whole.
Taking the holomovement perspective we can see that we are “an assemblage of heuristic devices…. a system encompassing the vast swathe of the universe they ‘know’ and continuously open to the universe they don’t”. An evolving system continuous with an all encompassing evolving system.
Well, to defend the philosophers, the ‘whole’ views you mention are speculative in the extreme. It may be the case that processes only seem temporal and discrete because of our parochial perspective, but lacking the ability to take a God’s eye view, what else do we have to run with? TI is the enemy here. Short of that, what is there to do but accept the life sciences perspective – the only that could very deliver immortality, and perhaps the very informatic perspective an empirical investigation of ‘holomovement’ would require? Al’s position here is more skeptical than speculative, I think.
Smart – and well written.
I am quite over my head to try and pick holes in the AI’s reasoning so I’ll just comment on a few things that stand out to me
I’m quite convinced it isn’t possible to relinquish it, without some pretty radical re-engineering. In a sense, it would be akin to abandoning the self itself
Well, it’s been pretty obvious to me for as long as I care to remember myself that morality is nothing more than a subjective quirk, hardly more valid (in the philosophical / universal sense) than fear of arthropods.
Agency and intentionality greatly depend on how you define them (under some definitions they very well may exist, but under those definitions even hard drives have agency, and, one might argue, even intentionality)
It baffles me that someone would find this significantly surprising (though yes, subjective and whimsical nature of so-called “morality” is quite disconcerting, but then, so are, for instance, quantum effects) – but then again, the very claim of having a “sense of free will” baffles me.
P.S.:
As to the whole “aliens studying cats” thought experiment, I’m not sure it’s even logically coherent.
?
Aliens that have no concept of “consciousness” can not even be related via human concepts of “possessing knowledge”. I’m not sure modern language is equipped well enough to construct narratives about unconscious agents, let alone narratives describing what knowledge such agents “have” under some circumstances.
And I’m not convinced cats are conscious at all, but that’s a bit of a pedantic criticism, isn’t it
I’m quite convinced it isn’t possible to relinquish it, without some pretty radical re-engineering. In a sense, it would be akin to abandoning the self itself
I had a responce along these lines – but then it seemed off topic as this piece aims at the idea of materialism, it seems.
Really, take the AI to be looking out from the very edge of a cliff. What distinguishes the step forward from the step prior? If you abandon a Master Heuristic, at all levels, wont either step seem absolutely equal?
Then I’d ra ra on about us being the photographic negative of the dead and yes, how could we grasp that our light is the AI’s dark and our dark the AI’s light – how can we really grasp what it’s saying, when at any moment the AI will step off that cliff and were just not enlightened enough to follow? Were just lowly photographic negatives of enlightened dead (oooh, had to squeeze that in – didn’t plan it in advance, honest!)
But I think that covers another topic that’s different from criticising materialism.
Not necessarily.
I think I brought this up before, but a coherent goal system can exist in absence of any such consciousness or somesuch.
Take an ICBM speeding to a populated city.
Goal driven ? Yeah.
Aware of self and surroundings ? Yes (otherwise it couldn’t do evasive maneuvers on late flight stages, and couldn’t correct it course in case of deviation)
Conscious / capable of “subjective experience” ? Errrr… Doubtful
I’ve probably been a bit broad in my use of ‘master heuristic’, particularly in glibly referencing ‘at all levels’
Aware of self and surroundings ? Yes (otherwise it couldn’t do evasive maneuvers on late flight stages, and couldn’t correct it course in case of deviation)
I don’t know about aware, but essentially that’s a heuristic in action (as I’m using the term, anyway). Sure, it’s probably simpler than an earthworms heuristic, but…
It might be picking at the analogy rather than the problem, but how does the AI converse (Assuming it’s not just outputting a rather luckily written prewritten speach/playing a tape) without any way of navigating/without some kind of heuristic to work by? How does the missile aim itself in the conversation?
Heck, even if a conversation is, for it, simplistic as to be comparable to the system that keeps our hearts beating (ie, it can just keep talk, talk, talking like our hearts just keep beating – and about as dumb, even if it seems complex to us in conversation), whats in control of that? A larger heuristic? Why did it turn that circuit on?
To make an extended draw of the bow, what might be interesting is what happens where the supposed master heuristic does not really differentiate between positive and negative, but the supposed subordinate heuristic/the talker does. What happens then? Does the talker, to forfil it’s ends, talk over the inert master (in whatever ways the talker can figure out/is physically able to)?
Regarding the original scenario, yeah. Like I pointed out to Eric, the inconsistency is explicitly larded into the scenario, so I’m not sure how the reductio is supposed to work.
Ooooopsssie!
Looks like I’ve made gravatar moderately unhappy by trying to have more than one Pinhead image
And everybody knows philosophical discussions need more Pinhead userpics
I think I got rusty because I got lost in the first thing the AI says…
Anyway, I just finished watching this: http://www.youtube.com/watch?v=FnSEt2BCcRs
Documentary for the masses on consciousness. It should be part of a series with the Stephen Hawking brand.
There’s nothing really so interesting about it beside its theme and really fancy editing. Very sensational, emphatic style, but I guess it works. It’s quite nice to watch.
Though it’s interesting that these videos all seem to end with some rather vague philosophical statements that are supposed to make one feel good.
There was a storm before and I couldn’t complete.
Anyway, the more I hear the perspective of “meaningless” evolution, that is examined in that video too, the more I think it becomes a case similar to the one in the Malazan books.
Gods exist to give purpose. So that we wish to be like a tool, made for a specific goal that we simply do not fully understand. So as long there are gods, there’s agency. You either help the god’s plan, or you oppose it if the gods are enemies. The world is made relatively simple.
There’s the notion of justice. So if you sin you go to hell, and if you are a good person you are eventually rewarded. It’s quite convenient because it does the job for you. But what happens when the gods are gone? It’s simple: responsibility. You can’t abdicate.
It’s when there are no gods that one achieve a true emancipation and freedom. Suddenly you aren’t moving on a defined path, no safety nets, no one that tells you what to do, but you truly do determine and decide what is right or wrong. You have the full responsibility of your choices, you answer to everything.
Now the problem is that it’s the whole notion of free will to be at stake, so this perspective sounds as delusional as the other. But still, even the idea that consciousness is just passive spectation, if there’s nothing else to it then it continues to be a form of freedom. If the experience itself has no true meaning, from our perspective the meaning is built and we live within it. It can be an hallucination, but we earn the right to erase the difference.
It’s as if saying that without gods, and purposes, you get the full responsibility of your actions. And without free will you get the right to deny reality and build one. With progress and science, comes the possibility to emancipate more and more FROM reality, till you fully accomplish the hallucination itself. It’s not about a stronger control on reality, it’s about moving away from it.
The destination of humanity is to be completely deceived, instead of partially so. Because that’s how they reclaim the world. An illusory one, perfectly tailor-made. And, philosophically, the real world has no right to claim itself “real” anymore. Because our limited perspective isn’t limited anymore. It’s the Law, since nothing else exists. You draw from a blank state.
That’s the drift of relativity. You reach a point where you can truly draw the world as you like, and the hallucination is not subordinate to reality.
Before entering this world, doesn’t it sound like a world where you cannot be wrong?
No one can prove you wrong, because you deny any physical method of being proven wrong because the hallucination comes before the physical. And the hallucination has no bone to it – the hallucination doesn’t maintain some sort of logic system that remains still so it can be proven to be wrong by atleast logic, since the hallucination is boneless and will just twist to fit itself as being right. That’d be beyond gaming ambiguity – you don’t even have to game what crumbs of ambiguity reality passes you – instead it makes everything ambiguous. Indeed, that sounds like ‘the outside’, doesn’t it?
What becomes of responsibility when no matter what you do, you come up being right and virtuous?
The hallucination can have as many rules as the real world. The same as dreams have their own logic.
It’s not the negation of order. Besides, it’s possible that the “subjective” is still collective. So you don’t have a private reality, but one shared by the peoples like you. You just have a strongest control on its physics, which is already happening.
The theory works because it doesn’t predict anything fancy. It only describes what already happens. You need only to compare the way we live today to the way we lived hundred of years ago. Everything changed. Parts of the real world were reclaimed and reshaped.
You can build society any way you want. It can be just, or unjust. It boils down to what people end up shaping up. And it has “real” consequences on the lives of all people.
I agree it’s ordered – there’s plenty of order within believing your eating ham while your chowing down on a baby. But without objective reality being linked in, no ones going to prove it’s anything but ham. Reminds me of ‘Better than life’ described in the Red Dwarf book, where someone notes they saw a BTL junkie eat their own vomit as if it were a sumptuous feast, because in the virtual world it was. So I’m not the only one to note this sort of distinction.
“So you don’t have a private reality, but one shared by the peoples like you.”
Which following the trends of today, will be more and more people you choose to be with you. An echo chamber. Instead of being stuck with people who think in different ways (irritating ways). This wont be a corrective system.
“The theory works because it doesn’t predict anything fancy. It only describes what already happens.”
I think ‘works’ being justified in it’s use is being used to justify another use of the word ‘works’. Eg, the theory that you are poisoned works because poison is found in your blood. That does not mean that having poison in your blood works.
And it has “real” consequences on the lives of all people.
Don’t let that “real” make the hallucination subordinate to it! Those people are fine!
Re: Abalieno
Sorry, but as far as I know it’s not how theories are supposed to work.
Theories are supposed to predict… things. And be judged upon how well they predict.
Theories that just “explain” are less than useless.
Which following the trends of today, will be more and more people you choose to be with you. An echo chamber. Instead of being stuck with people who think in different ways (irritating ways). This wont be a corrective system.
There’s a nice, completely nut anime that deals in part with this stuff
It’s aptly titled “Humanity has declined”.
http://myanimelist.net/forum/?topicid=464255&show=60#msg15913315
Excuse me if I mix again fancy scientific theories with mass culture
In the recap to the last episode of Fringe Jeff Jensen linked to this rather interesting article: http://www.nytimes.com/2002/03/12/science/peering-through-the-gates-of-time.html?pagewanted=all&src=pm
Which is basically the “story” at the basis of that movie/mockumentary “What the Bleep Do We Know!?”.
Though this is seen slightly more from the scientific perspective.
But it is interesting to notice how the “singularity” has this meaningful ambivalence. We talk about the singularity as the no-return point for consciousness and “reality”. You can then consider the depth of the soul, the gods that hide within the deeper brain, the cosmological gods as if coming out of a Lovecraft story. And then you can consider the singularity as it applies to physics and black holes:
At the center, space would be infinitely curved, and as Dr. Wheeler likes to say, ”smoke pours out of the computer.” Space, time and even the laws of physics themselves would break down at this cosmic dead end, called a singularity.”
The upper world that mimics the inner one. The cosmos mimicking exactly the depth of the soul/human brain/consciousness.
So, even if this is just an abstract idea, and so bullshit, there seem to be a sort of equivalence or link, between the two singularities. One deep within ourselves, that kind of threshold between consciousness and non-consciousness, and the one “above”, in the physics of the universe.
It’s just interesting to consider that the black hole singularity, with its effect on light, time and matter, is very similar to the singularity of consciousness: the idea we have of time and the concreteness of the real world.
The black hole in the space and the black hole just behind consciousness seem to be extremely close. Maybe TOO CLOSE for this to be just a romantic view.
And the fun is when the ambivalence is unintended, as in that article, the things still apply to both cases:
The part that interested Dr. Wheeler most was an apocalyptic prediction contained within the equations: matter, say in a dead star, could collapse into a heap so dense that light could not even escape from it, eventually squeezing itself out of existence.
This seems to me the same way we describe consciousness “squeezing itself out of existence”, as we move closer to the singularity. Those are similar to Bakker’s words.
According to past comments by Scott, BBT avoids metaphysical commitments. I wonder, do these alien AIs also avoid all metaphysics? It seems as though they have, in some sense, moved past metaphysical problems altogether.
At any rate, there’s something appallingly circular about their reasoning: “you” don’t exist, in order to prove it this to yourself you have to modify yourself until “you” don’t exist.
Thanks robot, that’s not very helpful.
Limits with one side and ‘first person frames’ exist, beyond that I can’t say a goddamn thing about them, but they do exist! *throws temper tantrum*
If evolution had saddled with a ‘I am a dog’ device, this ‘appalling circularity’ is precisely what you would expect! This really is a central question: Would evolution do something like this? Looking at peacock feathers, I think you have to admit, very possibly.
Metaphysics, on Al’s presentation, becomes something entirely different. His ‘cogito,’ notice, is an empirical claim: Humans are information processing systems. And metaphysics as traditionally understood, becomes a kind of heuristic reflex. What he offers is a kind of nonnormative pragmatism: metaphysical in that it employs unexplained explainers (information, causality, effectiveness) but very deflationary/minimalist otherwise.
I still have to crack Ladyman and Ross before I can say how near it is to what they offer. That book is killing my brain for some reason!
” this ‘appalling circularity’ is precisely what you would expect! ”
Which is why I threw a tantrum.
Didn’t you once accuse Derrida of using différance in similar fashion? An argument that self-protects by undercutting the tools you’d normally use to attack it as confirming it?
I never said I was comfortable with it! But it does follow that this is the kind of cognitive dilemma we would find ourselves if it is the case that evolution stranded us with a bunch of heuristics and a serious case of auto-cognitive neglect. What’s the wee little Frenchman’s excuse? Tenure.
Hi Scott! Check your ham filter
Being an unwashed plebe with no formal philosophical education, I gotta brazenly interrupt and say that I don’t see the “consciousness-less robots will not discover the existence of subjective conscious experiences regardless of the depth of their neurological investigation” as a valid criticism of materialism.
I’d say it might very well be a case of begging the question.
No, really, think about it:
the whole thing starts with an exotic hypothetical (which involves a thing with a mindset completely alien to us using a technology vastly ahead of ours to map out the brain of a creature which, debatably, might possess consciousness) and then we construct a thought experiment that requires us to postulate that, in case this unlikely hypothetical were to take place (an alien with a very inhuman “It’s thought, Jim, but not as we know it” mental organization studies cat brains via super-technology), a certain counterintuitive outcome (no evidence of “subjective conscious experiences” is discerned by the aliens despite their “presence” in cats being postulated as true for the purpose of this experiment) will occur.
Of course such outcome will pose a problem for those materialist lines of thinking about consciousness that aren’t epiphenomenalist, but it will do so only because the thought experiment postulates a circumstance that is problematic to those (namely, that a sufficiently advanced neuroscientific inquiry won’t lead to discovery of “subjective conscious experience”)
A thought experiment about what could the consequences of CERN unwittingly summoning Baron Samedi during an LHC experiment be for materialism (and CERN) is IMHO exactly as enlightening.
P.S.: Trying to get past spam filter
Just a suggestion. Your post was too long, and seemed to flutter around a point which I don’t think was ever made quite clear enough.
AI: Who are you?
Cat: Mats cat and I noticed Mat doesn’t just sniff his food before he starts to eat it, but checks labels and dates. I think somewheres in his evolution he unfortunately lost his olfactory sense and sniffs logic instead. Poor creature also grew that huge neocortex. What happened to the poor guy?
AI: Let me tell you a secret….Mat cover your ears. Uh lets see…
Mat: Can I uncover them!
AI: Mat, why are you yelling?
Mat: I can’t hear myself!
AI: Just talk normal.
Mat: But…
AI: Oh I see you need to talk loud for control and feedback of your inner semantic self..aka thoughts. Do you cover your ears when engaged in silent heurisitcs? aka thought.
Cat: Strange creature Mat. I just see that cricket appear on my visual cortex and I run across the room and pounce on it. Mat’s wife yells “Cricket” at him and makes Mat chase after it. Human marriage is full of commands and rules, like she constantly tells Mat what he can and cannot say.
AI: That cricket on your visual cortex is actually an object inside you.
Cat: Oh! and the word “Cricket” yelled by Mat’s wife is actually an object on Mat’s huge neocortex.
AI: Yeah, that huge neocortex is a language cortex and makes Mat get “LOST IN THE HERMENEUTIC HALL OF MIRRORS” http://cogprints.org/1577/1/harnad90.dyer.crit.html
Awesome! Welcome, VicP. I like Harnad, and I think the ‘symbol grounding problem’ he plucks out of the Searle controversy is another interesting clue. But you have two routes to go with the problem: figure out a way to interpret meaning into nonhuman information processors (which by no means have to be ‘symbol crunchers’ (if there’s any such thing)) or find a way to interpret meaning out of human information processors. Since absolutely no one is attempting to do the latter, I’m inclined to think that’s where the answer lies. At least good ol’ Al agrees!
More incoherent chatter on my part, then.
First, I’m not exactly buying that the “flying-thinking” analogy in regards to inherent properties holds.
To elaborate: when a computer is running a simulation of an airplane, it’s not in the business of flying, it’s in the business of running a model to predict what would “flying under hypothetical conditions X” look like for a specific hypothetical plane.
Thinking, especially organized, logical thought, appears significantly more like the latter (one could argue that the whole point of being able to “think dem thotz” is to be able to accurately predict complex behaviors and their outcomes relative to yourself).
So if we take the flight sim model and use it for describing a hypothetical “thought emulating” (a hypothetical computer running software that models what a given hypothetical mind would think under given hypothetical circumstances), it won’t be akin to a computer emulating hypothetical flight behaviors.
It would be akin to a computer simulating… a computer.
Now, nested computer-in-computer-in-computer-all-the-way-down things are more of 01′s eparchy, but, at the risk of stealing rhetorical bread and butter from my dear 01, I gotta ask:
if I take a NES emulator, install it on a Windows machine, and then run a NES game on it and play it, would I be actually playing a NES game ?
If yes, then I don’t see how T=SC is even remotely problematic.
As to Searle’s criticism of Chinese Rooms…
…sorry, cant’ buy that.
If I understand it correctly, the whole argument boils down to the claim that, should Searle memorize every single component rule and symbol involved in a “fully functional” Chinese Room, he would still *not* understand Chinese.
Sorry, Searle, quod erat non demonstrandum
As far as I see (at least in several descriptions of the argument that I’ve googled up, just in case), there is not a single logical reason why one should conclude that upon learning a massive set of Chinese linguistic rules and symbols Searle would still not understand Chinese (and we can’t settle it empirically because we don’t have functioning Chinese rooms yet).
The argument is structured around the expectation that the audience will “intuitively” accept it, because it kind of sort of bellyfeels right.
Arguing from “intuition” is not even a fallacy.
It’s just bullshit.
P.S.:
What is language education, if not forcing a subject to internalize (and learn to quickly operate) a massive set of rules and symbols ?
I think Searle definitely has the better of you on this one, 03. Syntactic facility does not semantic comprehension make. If it did, the mind/body problem would have solved a long, long time ago.
I’m not convinced that a Chinese Room that can carry out what a human would consider a “coherent, intelligent discussion” can operate solely via a “syntactic facility”.
The rules necessary to make a convincing Chinese pen-pal would by necessity include every rule and definition a Chinese speaker has to “know” (so by memorizing all of them you inevitably learn the grammar and the vocabulary), and, in order to prevent the Chinese Room from degrading into a yet another Chomskybot, it would have to implement some ” set of rules that “binds” its chatter to actual reality of a human being’s existence (presumably, a Chinese citizen’s existence) which is crucial for passing Turing Test (aka “having an intelligent discussion).
In other words, anything that can pass a Turing Test with a native speaker needs a rather impressive set of semantic rules (so impressive, in fact, that we’re nowhere near formulating it).
It appears to me that meticulously memorizing the grammar/vocabulary part of “Chinese Room code” and comprehending it to the point you can “run it in your head” would give a human enough insight into Chinese to be considered “someone who understands Chinese” (since a living, “socially normal” human already comes with semantic rules in place, assuming he’s not some horrible victim of an inhumane feral-child experiment).
However, even if that turns out not to be true (I’m not sure that the above opinion of mine can be reliably disproved outside, well, actually building a CR and memorizing parts of its code), a Chinese Room that completely fulfills the original definition by necessity has a massive set of its own semantic rules (ones it uses to pass as a Chinese dude and not a blind idiot Postmodernism Generator), and anyone memorizing and comprehending those would get some semantic comprehension.
One could try, of course, to argue that the CR’s “semantic rules” would be very different from human “semantic comprehension”, but
a) as far as I can tell, Searle doesn’t prove that
b) I don’t see a way of proving or disproving such a hypothesis except for building a CR-style device, which makes this a purely experimental affair.
Has there ever been a scientific test where perhaps people with quirky persona’s of various sorts are put on one terminal #1 and just paid to chat while other people are told they are going into a Turing test and will be speaking, via terminal #2 and typing at a keyboard, with an AI. Terminal #2 is connected to terminal #1.
I wonder how many people on terminal #1 would fail the ‘Turing test’?
Which is a long way around of questioning ‘which is crucial for passing Turing Test’ and whether that crucialness is all that reliable in itself?
it would have to implement some ” set of rules that “binds” its chatter to actual reality of a human being’s existence
Seems to go beyond the rules of language?
Re: Callan
No, as far as I know, no one has tried to substitute computer with a “person with a mental condition” in a Turing Test (or better yet, include both computers and “unconventional people” in the test group).
It seems to me that there are many kinds of people who, try as they might, won’t pass – but in absence of experimental data, it’s just my mere intuition
Well, yes, a system that can hold an “intelligent” (as opposed to just “linguistically well-formed”) discussion needs some semantic rules, otherwise the discussion would look like this:
Dude:
Hello Asemantic Chinese Room! What do you think about Searle’s argument about Chinese Rooms ?
Asemantic Chinese Room:
It appears to me that Searle is beautifully ignorant of so-called “vagina tree model” of cognitive blooming and, consequentially, its crimson applications to the process of digesting a Chinese Room.
Dude:
O_O
However, I don’t think you need to memorize and understand the “semantic” part to learn Chinese from “Turing-Test capable Chinese Room” design documentation.
You just need the language bits. Assuming you already have some kind of semantic framework going – like (most?) humans do
Well, I’m simultaneously a fan of “derived semantics” response, “virtualization response” and “brain simulation” response to Searle’s room shenanigan (and yes, I do think that there is no reason not to believe that Searle, upon memorizing the full instruction set and data set of a Turing-test capable machine would become capable of running a “second mind” distinct from his own. It’s freaky, but if minds are computational, you can very well have more than one in a single brain – especially if you bother to read some of the weirder chronic brain damage case studies and thus know just how much fucking redundancy the thing has)
However, I think that 03 here is onto something, and something very important. The entire CR argument is a mite of bait-and-switch (Searle is a cool guy, so I believe it’s not intended to be a bait-and-switch, it just so happened).
It starts with an implicit postulate that the CR system has semantic rules in it (semantic rules, and Third correctly notes, are needed for an “intelligent” discussion). And then this postulate is rushed under the rhetorical carpet, and argument starts revolving around demonstrating “Syntactic != semantic comprehension”… which is exactly the “switch” part of bait and switch.
“Syntactic != semantic comprehension” may very well be true (Which I will rant about below) but it is not relevant to CRs, simply because when Searle has postulated an ability to have an “intelligent” discussion, he postulated presence of some (however stunted, crude and/or derivative) semantic rule-set.
Now… regarding semantic and syntactic “comprehensions” and their relationship, well, I don’t believe that they are strictly equivalent, but if Hellen Keller’s recollections of “living without a language” is accurate, there is an intimate connection between acquiring a formal language (be it spoken, gesture or touch) and developing what we would consider “consciousness” and “consciousness experience”.
For if Hellen Keller’s account of gaining “consciousness” and “thought” only upon being introduced to tactile form of language is accurate, then “syntactic” is a necessary prerequisite to “semantic” and there is an intimate, causal relationship between the two, at least for computation devices with human brain-like architecture.
One could then say that, in order to create a human mind, one has to infect a human brain with language.
P.S.:
, this does not contradict the existence of people with no “conscious internal language” aka “stream of consciousness” aka “internal monologue”, because, while such people lack a baseline-typical “inner voice” they have most certainly been taught one or more natural language, so their brains are as much “infected with language” as those of “folks with an inner voice”
Preempting Third’s favorite argument from personal experience
Looking it up further, I see the Chinese room was raised with an intent in mind rather than just as an actual experiment to see what happens.
“Since he does not understand Chinese, Searle argues, we must infer that the computer does not understand Chinese either.”
Seems kind of spurious if you think of an example where, say, he takes the words from an actual Chinese person, then follows some rules to convert them to another language he doesn’t know (lets say German) and to the other person it passes as a human (hopefully!). Does that mean the chinese person doesn’t “understand” Chinese because Searle didn’t understand Chinese, even though the German speaker had an intelligent conversation with the Chinese speaker?
I think they should look at running experiments like the Chinese room (unless they fear accidental AI creation) rather than using the idea as a proof.
Re: Callan
I doubt we have the technology to make a convincing “robot conversationist” or, for that matter, using your example, “robot translator”
Google Translate comes quite close, though.
Would Searle pick up enough Chinese to claim at least partial understanding of the language if he were to scrutinize and memorize every bit of code and every entry in the database that pertain to EnglishChinese translations by Google Translate ?
I guess he very well might. He should at least put his brain where is mouth is and try
I think he would quite possibly learn it. But his point seems to be that the bit of him that might learn it isn’t part of the chinese room – he’s just doing labour work that even a monkey could do.
But then again that’s the thing – he’s retracted any sort of part like that from the chinese room, then said that the chinese room doesn’t have a part like that that learns and understands.
Well, that objection seems irrelevant.
If a human, upon memorizing the “entire program and database” of a functional CR (or even a measly robo-translator, which is even in its idealized form vastly inferior to CR since it can’t carry out a conversation) gains understanding (even if limited) of the language the CR (or robo-translator) is using, the Searlian argument hasn’t got a leg, or even a bloody stump, to stand on.
As 01 observed above, The CR argument as it is usually presented is, basically:
1) postulating presence of a “semantic system” in CR in round one
2) then denying presence of any such thing in CR in round two
and the key tricks of such an argument are, first, somewhat disguising the semantic implication in round one, and, second, blazing from round one to round two in a sufficiently fast and a sufficiently aesthetically pleasing manner.
Sufficiently advanced dialog is indistinguishable from truth.
AI: You seem to know alot about Mat’s species.
Cat: I’ve been reading up on the neocortex in Cricket’s book.
AI: Cricket?
Cat: Yes, Francis Cricket, “The Astonishing Hypothesis” P135. Mat’s Neocortex has 6 layers and the word “Cricket” must go right to Mat’s lower brain functions and make him run across the room.
AI: Do you know which layer he does his heuristics in? Most of Mat’s inner words have meaning, but the meaning doesn’t always reach deep into his lower brain regions.
Cat: Last winter I saw Mat trying to walk across the pond in the backyard but he was slipping all over the place. Like something was ringing his bell-um.
I was reminded of this post by today’s SMBC: http://www.smbc-comics.com/index.php?db=comics&id=2753#comic
If you’re right, hopefully this isn’t how it’ll happen!
Or, the scarier thought: this process is already underway
Good link, really pretty spot on! That SMBC comic is quite a thing – I’m curious about the guys (or gals?) background?
Awesome!
Worth a read, although I’m pretty sure he’s wrong about both the ‘ease’ and ‘irrelevance’ of programing an algorithm that can recognize itself in a mirror.
http://www.aeonmagazine.com/being-human/david-deutsch-artificial-intelligence/
Yeah, this was actually published on the Guardian site. You should see the amount of comment furor in kicked up in the comments over there!
The essay has a number of head-scratchers, I agree. He seems to play fast and loose with the distinction between ‘behaves as if’ and ‘has the experience of,’ especially with regards to self-recognition.
Deutch is cool, but I think he is overly-dismissive of the “sufficient complexity” argument. The argument may be silly on its face, but nature has numerous examples when just dialing the quantity of objects up to eleven brings about distinct, qualitative changes.
Consider a single speck of sand floating through space. It doesn’t have appreciable gravity, doesn’t form satellites or atmosphere.
Now imagine that a “sufficiently advanced” alien, in order to win an obscure philosophical argument, is manufacturing identical (in terms of size and chemical composition) sand specks at an astounding rate, and placing them in the immediate vicinity of the original.
At some point, the total mass of the sand specks in that region of specks reaches 8*10^24 kilograms, at which point the alien leaves, satisfied.
Thanks to gravity and accretion, the resultant cloud of coarse particulate will (assuming there are no massive bodies in the immediate vicinity to gravitationally drag this sand-cloud upon themselves) coalesce into a silicate planetoid, which would eventually (by virtue of its own fairly impressive gravity) gain a frozen not-quite-atmosphere not unlike that of Pluto (but much more pronounced).
You would eventually have a very nice, deathly cold nomadic planet.
This is not to say that I believe there is an “intelligence gravity effect” that would make any sufficiently powerful computation system to “accrete” a mind, but dismissing arguments from sufficient quantity / complexity seems… premature
That intelligence gravity effect sounds like the blind brain hypothesis – it gets sufficient mass that it can’t keep up with itself as individual particals – not so much accrete a mind, but accrete a confusion. Which makes Kellhus…a black hole? Or is that the No God? Okay, who can resist a long draw of the bow, eh?
Something struck me while re-reading the argument between Third and Scott…
What if Mitt Romney is actually a living Chinese Room ?
Just think about it, don’t bizarre, aloof comments like “You know, the trees are the right height” indicate a semantic framework that has unusual symbol grounding issues, of an entirely alien mind devoid of human experiences and yet burdened with the task of pretending to be human ?
Insofar as I understood this, I enjoyed the POV. There’s a radicalism to the ideas that smells true-ish to me. Trying to understand if it’s RIGHT, however, is challenging, especially since the argument partially undermines such concepts as truth and falsity all together.
I love dialogues in philosophy, (Is not Al a 21st Century version of Berkeley’s Philonous?) and the format worked well here. One point of style: I don’t think referring to consciousness as an illusion is helpful at all. That implies that we think consciousness is something that we understand which turns out not to be the case, like a mirage in a desert. Rather, the problem with consciousness is JUST THAT we don’t understand what it is and are in need of a more satisfying explanation. Consciousnesses isn’t REALLY an illusion according to this theory, it’s just that we lacked the conceptual model applicable to our conscious experiences. Any good theory of consciousness will not negate our basic intuitions about our experience but rather subsume them in a deeper explanatory paradigm, which this theory has the potential to do. (A dialectical notion, i admit.) That being said, if Al is right, he couldn’t ever truly express himself convincingly in language as all language essentially plays the same semantic game, a game which the AI recognizes is ultimately misguided. Fair enough?
Ultimately, I fear that this kind of understanding tests the limits of conceivability, in that it pushes us to the limits of concepts that we can hold in our mind without contradiction. It inches toward paradox. This is a very “high” (or “deep” if you prefer) hierarchical level of explanation, one that probably needs a few gaps filled first along the way before we get there. As such, I doubt it can really “help” us much in how we think about how the brain generates mind scientifically, at least in the short term. Certainly, I think an information-based model is ultimately the way to go, not only for philosophy of mind but for metaphysics and natural physics as well. (The convergences of thought are incredible in this area.)
A funny question but a nonetheless serious one: Does Al enjoy its life? Could it? Do values somehow transcend this debate over consciousness? Or will a true understanding of the informational nature of consciousness rob the world of value just as you claim such understanding does for semantics and qualia? Certainly in your dialogue AI enjoys his smugness.
The hinky sense you get from Al’s use of ‘consciousness’ is precisely the one I get from my usages more generally. The novelist in me, I sometimes think, pushes a little too hard for the rhetorically satisfying expression. You’re entirely right: Al uses the term inconsistently.
I personally find Al’s position VERY difficult to think, even though I agree with him pretty much all the way through. Once you accept (as I think we must) that human cognition, like animal cognition, is built around certain ‘hardwired heuristic defaults,’ then there’s no escaping the ‘applicability problem.’ Once you acknowledge the role (both metacognitive and cognitive) played by informatic neglect, you find yourself a very strange position vis a vis your own intuitions. The ‘epistemic heuristic,’ as Al calls it, frickin feels universal, even though it almost certainly is not. As a result, you become your own worst cognitive enemy, not unlike someone suffering Anton’s syndrome trying to convince themselves they’re blind even though, as far as they can tell, they see perfectly fine.
This has really got me thinking about Graham Priest, dielethism, and paraconsistent logics, lately. Heuristic cognition is also FRAGMENTED cognition. Informatic neglect means that, from a metacognitive standpoint, whatever heuristic angle we take to a problem will feel appropriate to that problem whether that problem lies within its scope of applicability or not. So we have to resist this tendency, train ourselves to look at philosophical impasses as possible ‘scope violations.’ But beyond this, I’m not at all sure whether Al’s ‘to the victor go the spoils’ approach is the best or the only one. Rather, what he describes seems to be another philosophical branching point. We can defer to information as our new ‘Master Heuristic’ as Al recommends, or we can resist subscribing to any hierarchy at all, and adopt a ‘hybridized model,’ where we figure out ways to delimit various scopes of applicability belonging to various heuristics (which themselves need to be empirically identified) and content ourselves with having no ‘master perspective’ whatsoever.
And this gives me, yet again, another Nietzschean tingle down my spine – literally! Because it sounds awfully similar to (or at least offers a fascinating way to interpret) what he advocates with his ‘perspectivalism.’
In answer to your question, Al would laugh, yawn like a kitten, and say, “Oh you humans! Yes. And no!”
Just struck me that the robot’s attitude about engineering reminds me of how much Vox tries to play down science and say engeneering’s where it’s at.