Patricia Churchland

Reflections on the Chinese Room

pnrj cognitive science , , , , , , , , , , ,

Jul 12 JDN 2459044

Perhaps the most famous thought experiment in the philosophy of mind, John Searle’s Chinese Room is the sort of argument that basically every expert knows is wrong, yet can’t quite explain what is wrong with it. Here’s a brief summary of the argument; for more detail you can consult Wikipedia or the Stanford Encyclopedia of Philosophy.

I am locked in a room. The only way to communicate with me is via a slot in the door, through which papers can be passed.

Someone on the other side of the door is passing me papers with Chinese writing on them. I do not speak any Chinese. Fortunately, there is a series of file cabinets in the room, containing instruction manuals which explain (in English) what an appropriate response in Chinese would be to any given input of Chinese characters. These instructions are simply conditionals like “After receiving input A B C, output X.”

I can follow these instructions and thereby ‘hold a conversation’ in Chinese with the person outside, despite never understanding Chinese.

This room is like a Turing Test. A computer is fed symbols and has instructions telling it to output symbols; it may ‘hold a conversation’, but it will never really understand language.

First, let me note that if this argument were right, it would pretty much doom the entire project of cognitive science. Searle seems to think that calling consciousness a “biological function” as opposed to a “computation” can somehow solve this problem; but this is not how functions work. We don’t say that a crane ‘isn’t really lifting’ because it’s not made of flesh and bone. We don’t say that an airplane ‘isn’t really flying’ because it doesn’t flap its wings like a bird. He often compares to digestion, which is unambiguously a biological function; but if you make a machine that processes food chemically in the same way as digestion, that is basically a digestion machine. (In fact there is a machine called a digester that basically does that.) If Searle is right that no amount of computation could ever get you to consciousness, then we basically have no idea how anything would ever get us to consciousness.

Second, I’m guessing that the argument sounds fairly compelling, especially if you’re not very familiar with the literature. Searle chose his examples very carefully to create a powerfully seductive analogy that tilts our intuitions in a particular direction.

There are various replies that have been made to the Chinese Room. Some have pointed out that the fact that I don’t understand Chinese doesn’t mean that the system doesn’t understand Chinese (the “Systems Reply”). Others have pointed out that in the real world, conscious beings interact with their environment; they don’t just passively respond to inputs (the “Robot Reply”).

Searle has his own counter-reply to these arguments: He insists that if instead of having all those instruction manuals, I memorized all the rules, and then went out in the world and interacted with Chinese speakers, it would still be the case that I didn’t actually understand Chinese. This seems quite dubious to me: For one thing, how is that different from what we would actually observe in someone who does understand Chinese? For another, once you’re interacting with people in the real world, they can do things like point to an object and say the word for it; in such interactions, wouldn’t you eventually learn to genuinely understand the language?

But I’d like to take a somewhat different approach, and instead attack the analogy directly. The argument I’m making here is very much in the spirit of Churchland’s Luminous Room reply, but a little more concrete.

I want you to stop and think about just how big those file cabinets would have to be.

For a proper Turing Test, you can’t have a pre-defined list of allowed topics and canned responses. You’re allowed to talk about anything and everything. There are thousands of symbols in Chinese. There’s no specified limit to how long the test needs to go, or how long each sentence can be.

After each 10-character sequence, the person in the room has to somehow sort through all those file cabinets and find the right set of instructions—not simply to find the correct response to that particular 10-character sequence, but to that sequence in the context of every other sequence that has occurred so far. “What do you think about that?” is a question that one answers very differently depending on what was discussed previously.

The key issue here is combinatoric explosion. Suppose we’re dealing with 100 statements, each 10 characters long, from a vocabulary of 10,000 characters. This means that there are ((10,000)^10)^100 = 10^4000 possible conversations. That’s a ludicrously huge number. It’s bigger than a googol. Even if each atom could store one instruction, there aren’t enough atoms in the known universe. After a few dozen sentences, simply finding the correct file cabinet would be worse than finding a needle in a haystack; it would be finding a hydrogen atom in the whole galaxy.

Even if you assume a shorter memory (which I don’t think is fair; human beings can absolutely remember 100 statements back), say only 10 statements, things aren’t much better: ((10,000)^10)^10 is 10^400, which is still more atoms than there are in the known universe.

In fact, even if I assume no memory at all, just a simple Markov chain that responds only to your previous statement (which can be easily tripped up by asking the same question in a few different contexts), that would still be 10,000^10 = 10^40 sequences, which is at least a quintillion times the total data storage of every computer currently on Earth.

And I’m supposed to imagine that this can be done by hand, in real time, in order to carry out a conversation?

Note that I am not simply saying that a person in a room is too slow for the Chinese Room to work. You can use an exaflop quantum supercomputer if you like; it’s still utterly impossible to store and sort through all possible conversations.

This means that, whatever is actually going on inside the head of a real human being, it is nothing like a series of instructions that say “After receiving input A B C, output X.” A human mind cannot even fathom the total set of possible conversations, much less have a cached response to every possible sequence. This means that rules that simple cannot possibly mimic consciousness. This doesn’t mean consciousness isn’t computational; it means you’re doing the wrong kind of computations.

I’m sure Searle’s response would be to say that this is a difference only of degree, not of kind. But is it, really? Sometimes a sufficiently large difference of degree might as well be a difference of kind. (Indeed, perhaps all differences of kind are really very large differences of degree. Remember, there is a continuous series of common ancestors that links you and I to bananas.)

Moreover, Searle has claimed that his point was about semantics rather than consciousness: In an exchange with Daniel Dennett he wrote “Rather he [Dennett] misstates my position as being about consciousness rather than about semantics.” Yet semantics is exactly how we would solve this problem of combinatoric explosion.

Suppose that instead of simply having a list of symbol sequences, the file cabinets contained detailed English-to-Chinese dictionaries and grammars. After reading and memorizing those, then conversing for awhile with the Chinese speaker outside the room, who would deny that the person in the room understands Chinese? Indeed what other way is there to understand Chinese, if not reading dictionaries and talking to Chinese speakers?

Now imagine somehow converting those dictionaries and grammars into a form that a computer could directly apply. I don’t simply mean digitizing the dictionary; of course that’s easy, and it’s been done. I don’t even mean writing a program that translates automatically between English and Chinese; people are currently working on this sort of thing, and while still pretty poor, it’s getting better all the time.

No, I mean somehow coding the software so that the computer can respond to sentences in Chinese with appropriate responses in Chinese. I mean having some kind of mapping within the software of how different concepts relate to one another, with categorizations and associations built in.

I mean something like a searchable cross-referenced database, so that when asked the question, “What’s your favorite farm animal?” despite never having encountered this sentence before, the computer can go through a list of farm animals and choose one to designate as its ‘favorite’, and then store that somewhere so that later on when it is again asked it will give the same answer. And then why asked “Why do you like goats?” the computer can go through the properties of goats, choose some to be the ‘reason’ why it ‘likes’ them, and then adjust its future responses accordingly. If it decides that the reason is “horns are cute”, then when you mention some other horned animal, it updates to increase its probability of considering that animal “cute”.

I mean something like a program that is programmed to follow conversational conventions, so when you ask it its name, will not only tell you something; it will ask you your name in return, and stores that information for later. And then it will map the sound of your name to known patterns of ethnic naming conventions, and so when you say your name is “Ling-Ling Xu” it asks “Is your family Chinese?” And then when you say “yes” it asks “What part of China are they from?” and then when you say “Shanghai” it asks “Did you grow up there?” and so on. It’s not that it has some kind of rule that says “Respond to ‘Shanghai’ with ‘Did you grow up there?’”; on the contrary, later in the conversation you may say “Shanghai” and get a different response because it was in a different context. In fact, if you were to keep spamming “Shanghai” over and over again, it would sound confused: “Why do you keep saying ‘Shanghai’? I don’t understand.”

In other words, I mean semantics. I mean something approaching how human beings actually seem to organize the meanings of words in their brains. Words map to other words and contexts, and some very fundamental words (like “pain” or “red”) map directly to sensory experiences. If you are asked to define what a word means, you generally either use a lot of other words, or you point to a thing and say “It means that.” Why can’t a robot do the same thing?

I really cannot emphasize enough how radically different that process would be from simply having rules like “After receiving input A B C, output X.” I think part of why Searle’s argument is so seductive is that most people don’t have a keen grasp of computer science, and so the difference between a task that is O(N^2) like what I just outlined above doesn’t sound that different to them compared to a task that is O(10^(10^N)) like the simple input-output rules Searle describes. With a fast enough computer it wouldn’t matter, right? Well, if by “fast enough” you mean “faster than could possibly be built in our known universe”, I guess so. But O(N^2) tasks with N in the thousands are done by your computer all the time; no O(10^(10^N)) task will ever be accomplished for such an N within the Milky Way in the next ten billion years.

I suppose you could still insist that this robot, despite having the same conceptual mappings between words as we do, and acquiring new knowledge in the same way we do, and interacting in the world in the same way we do, and carrying on conversations of arbitrary length on arbitrary topics in ways indistinguishable from the way we do, still nevertheless “is not really conscious”. I don’t know how I would conclusively prove you wrong.

But I have two things to say about that: One, how do I know you aren’t such a machine? This is the problem of zombies. Two, is that really how you would react, if you met such a machine? When you see Lieutenant Commander Data on Star Trek: The Next Generation, is your thought “Oh, he’s just a calculating engine that makes a very convincing simulation of human behavior”? I don’t think it is. I think the natural, intuitive response is actually to assume that anything behaving that much like us is in fact a conscious being.

And that’s all the Chinese Room was anyway: Intuition. Searle never actually proved that the person in the room, or the person-room system, or the person-room-environment system, doesn’t actually understand Chinese. He just feels that way, and expects us to feel that way as well. But I contend that if you ever did actually meet a machine that really, truly passed the strictest form of a Turing Test, your intuition would say something quite different: You would assume that machine was as conscious as you and I.

The Cognitive Science of Morality Part II: Molly Crockett

pnrj cognitive science , , , , , , , , , , , , , , , , , , , , , , , , , , ,

JDN 2457140 EDT 20:16.

This weekend has been very busy for me, so this post is going to be shorter than most—which is probably a good thing anyway, since my posts tend to run a bit long.

In an earlier post I discussed the Weinberg Cognitive Science Conference and my favorite speaker in the lineup, Joshua Greene. After a brief interlude from Capybara Day, it’s now time to talk about my second-favorite speaker, Molly Crockett. (Is it just me, or does the name “Molly” somehow seem incongruous with a person of such prestige?)

Molly Crockett is a neuroeconomist, though you’d never hear her say that. She doesn’t think of herself as an economist at all, but purely as a neuroscientist. I suspect this is because when she hears the word “economist” she thinks of only mainstream neoclassical economists, and she doesn’t want to be associated with such things.

Still, what she studies is clearly neuroeconomics—I in fact first learned of her work by reading the textbook Neuroeconomics, though I really got interested in her work after watching her TED Talk. It’s one of the better TED talks (they put out so many of them now that the quality is mixed at best); she talks about news reporting on neuroscience, how it is invariably ridiculous and sensationalist. This is particularly frustrating because of how amazing and important neuroscience actually is.

I could almost forgive the sensationalism if they were talking about something that’s actually fantastically boring, like, say, tax codes, or financial regulations. Of course, even then there is the Oliver Effect: You can hide a lot of evil by putting it in something boring. But Dodd-Frank is 2300 pages long; I read an earlier draft that was only (“only”) 600 pages, and it literally contained a three-page section explaining how to define the word “bank”. (Assuming direct proportionality, I would infer that there is now a twelve-page section defining the word “bank”. Hopefully not?) It doesn’t get a whole lot more snoozeworthy than that. So if you must be a bit sensationalist in order to get people to see why eliminating margin requirements and the swaps pushout rule are terrible, terrible ideas, so be it.

But neuroscience is not boring, and so sensationalism only means that news outlets are making up exciting things that aren’t true instead of saying the actually true things that are incredibly exciting.

Here, let me express without sensationalism what Molly Crockett does for a living: Molly Crockett experimentally determines how psychoactive drugs modulate moral judgments. The effects she observes are small, but they are real; and since these experiments are done using small doses for a short period of time, if these effects scale up they could be profound. This is the basic research component—when it comes to technological fruition it will be literally A Clockwork Orange. But it may be A Clockwork Orange in the best possible way: It could be, at last, a medical cure for psychopathy, a pill to make us not just happier or healthier, but better. We are not there yet by any means, but this is clearly the first step: Molly Crockett is to A Clockwork Orange roughly as Michael Faraday is to the Internet.

In one of the experiments she talked about at the conference, Crockett found that serotonin reuptake inhibitors enhance harm aversion. Serotonin reuptake inhibitors are very commonly used drugs—you are likely familiar with one called Prozac. So basically what this study means is that Prozac makes people more averse to causing pain in themselves or others. It doesn’t necessarily make them more altruistic, let alone more ethical; but it does make them more averse to causing pain. (To see the difference, imagine a 19th-century field surgeon dealing with a wounded soldier; there is no anesthetic, but an amputation must be made. Sometimes being ethical requires causing pain.)

The experiment is actually what Crockett calls “the honest Milgram Experiment“; under Milgram, the experimenters told their subjects they would be causing shocks, but no actual shocks were administered. Under Crockett, the shocks are absolutely 100% real (though they are restricted to a much lower voltage of course). People are given competing offers that contain an amount of money and a number of shocks to be delivered, either to you or to the other subject. They decide how much it’s worth to them to bear the shocks—or to make someone else bear them. It’s a classic willingness-to-pay paradigm, applied to the Milgram Experiment.

What Crockett found did not surprise me, nor do I expect it will surprise you if you imagine yourself in the same place; but it would totally knock the socks off of any neoclassical economist. People are much more willing to bear shocks for money than they are to give shocks for money. They are what Crockett terms hyper-altruistic; I would say that they are exhibiting an apparent solidarity coefficient greater than 1. They seem to be valuing others more than they value themselves.

Normally I’d say that this makes no sense at all—why would you value some random stranger more than yourself? Equally perhaps, and obviously only a psychopath would value them not at all; but more? And there’s no way you can actually live this way in your daily life; you’d give away all your possessions and perhaps even starve yourself to death. (I guess maybe Jesus lived that way.) But Crockett came up with a model that explains it pretty well: We are morally risk-averse. If we knew we were dealing with someone very strong who had no trouble dealing with shocks, we’d be willing to shock them a fairly large amount. But we might actually be dealing with someone very vulnerable who would suffer greatly; and we don’t want to take that chance.

I think there’s some truth to that. But her model leaves something else out that I think is quite important: We are also averse to unfairness. We don’t like the idea of raising one person while lowering another. (Obviously not so averse as to never do it—we do it all the time—but without a compelling reason we consider it morally unjustified.) So if the two subjects are in roughly the same condition (being two undergrads at Oxford, they probably are), then helping one while hurting the other is likely to create inequality where none previously existed. But if you hurt yourself in order to help yourself, no such inequality is created; all you do is raise yourself up, provided that you do believe that the money is good enough to be worth the shocks. It’s actually quite Rawslian; lifting one person up while not affecting the other is exactly the sort of inequality you’re allowed to create according to the Difference Principle.

There’s also the fact that the subjects can’t communicate; I think if I could make a deal to share the money afterward, I’d feel better about shocking someone more in order to get us both more money. So perhaps with communication people would actually be willing to shock others more. (And the sensation headline would of course be: “Talking makes people hurt each other.”)

But all of these ideas are things that could be tested in future experiments! And maybe I’ll do those experiments someday, or Crockett, or one of her students. And with clever experimental paradigms we might find out all sorts of things about how the human mind works, how moral intuitions are structured, and ultimately how chemical interventions can actually change human moral behavior. The potential for both good and evil is so huge, it’s both wondrous and terrifying—but can you deny that it is exciting?

And that’s not even getting into the Basic Fact of Cognitive Science, which undermines all concepts of afterlife and theistic religion. I already talked about it before—as the sort of thing that I sort of wish I could say when I introduce myself as a cognitive scientist—but I think it bears repeating.

As Patricia Churchland said on the Colbert Report: Colbert asked, “Are you saying I have no soul?” and she answered, “Yes.” I actually prefer Daniel Dennett’s formulation: “Yes, we have a soul, but it’s made of lots of tiny robots.”

We don’t have a magical, supernatural soul (whatever that means); we don’t have an immortal soul that will rise into Heaven or be reincarnated in someone else. But we do have something worth preserving: We have minds that are capable of consciousness. We love and hate, exalt and suffer, remember and imagine, understand and wonder. And yes, we are born and we die. Once the unique electrochemical pattern that defines your consciousness is sufficiently degraded, you are gone. Nothing remains of what you were—except perhaps the memories of others, or things you have created. But even this legacy is unlikely to last forever. One day it is likely that all of us—and everything we know, and everything we have built, from the Great Pyramids to Hamlet to Beethoven’s Ninth to Principia Mathematica to the US Interstate Highway System—will be gone. I don’t have any consolation to offer you on that point; I can’t promise you that anything will survive a thousand years, much less a million. There is a chance—even a chance that at some point in the distant future, whatever humanity has become will find a way to reverse the entropic decay of the universe itself—but nothing remotely like a guarantee. In all probability you, and I, and all of this will be gone someday, and that is absolutely terrifying.

But it is also undeniably true. The fundamental link between the mind and the brain is one of the basic facts of cognitive science; indeed I like to call it The Basic Fact of Cognitive Science. We know specifically which kinds of brain damage will make you unable to form memories, comprehend language, speak language (a totally different area), see, hear, smell, feel anger, integrate emotions with logic… do I need to go on? Everything that you are is done by your brain—because you are your brain.

Now why can’t the science journalists write about that? Instead we get “The Simple Trick That Can Boost Your Confidence Immediately” and “When it Comes to Picking Art, Men & Women Just Don’t See Eye to Eye.” HuffPo is particularly awful of course; the New York Times is better, but still hardly as good as one might like. They keep trying to find ways to make it exciting—but so rarely seem to grasp how exciting it already is.