In my last post I gave some definitions of Autonomy, Intelligence, and Consciousness and discussed some possible implications with the intent of further discussing artificial intelligence.
In this post, we’re going to dive a bit deeper into consciousness, and set the stage for a discussion of how conscious robots might emerge.
Consciousness in Philosophy
In the study of philosophy of mind there are two main viewpoints, the dualist and the physicalist. The dualist, the primary platform of which lies on Descartes’
philosophy that there is a soul, maintains there is an immaterial part that interacts with the material body resulting in a person’s “mind” [1]. The physicalist maintains that the mind, body, and consciousness arise from physical processes that arise from material.
As with most things in philosophy, the two sides of this debate continue to disagree. However, as neuroscience advances, and we continue to learn more about how the brain works, it is becoming clearer that the brain can be viewed as a complex nonlinear system of systems [2] from which “consciousness emerged as a product of increasing biological complexity, from non-conscious precursors composed of non-conscious components” [3]. As a result, Dennet said about whether or not a machine can be conscious
We have known the answer to this question for a century. The brain is a machine. It is a conscious machine. the brain is a biological machine just as much as the heart and the liver. So of course some machines can think and be conscious. Your brain and mine, for example.[4]
Interestingly, in Mormonism (which I failed to mention last time) Joseph Smith carved out another interesting idea spelled out in D&C 131:7-8
7 There is no such thing as immaterial matter. All spirit is matter, but it is more fine or pure, and can only be discerned by purer eyes;
8 We cannot see it; but when our bodies are purified we shall see that it is all matter.
This view is interesting in that it gives a nod to physicalism, yet (seemingly) reserves that the spirit matter cannot be detected. I think this leaves the believing Mormon with the (unchallengeable) understanding that yes, Mormons are physicalists, but science will not and cannot ever detect what makes up consciousness and the spirit. To me, at least, this puts Mormon theology in the same camp as the dualists, only with a twist!
Consciousness in Psychology
Consciousness in psychology often centers around questions about how memories are formed, retained, and recalled, how emotions arise, and the role of the sub-conscious in a person’s life. There is disagreement as to whether or not memory in humans can be represented by a symbolic system. Nevertheless, William James, an early pioneer in psychology stated:
For practical purposes, nevertheless, and limiting the meaning of the word consciousness to the personal self of the individual, we can pretty confidently answer the question prefixed to this paragraph by saying that the cortex is the sole organ of consciousness in man…My final conclusion, then, about the substantial Soul is that is explains nothing and guarantees nothing. I therefore feel entirely free to discard the word Soul from the rest of this book.[5]
Freud seemed to have a similar sentiment in declaring
The process of something becoming conscious is above all linked with the perceptions, which our sense organs receive from the external world.[6]
Some Thoughts
I have provided a very secular approach to the idea of consciousness. From the vantage point of computer scientists, and scientists in general, this is admittedly the presupposition since there is little to do if we admit a dualist or even Mormon point of view.
My own point of view follows more closely along the secular lines. I think the burden of proof lies with dualists to provide evidence of something immaterial (since there doesn’t seem to be anything else in the universe that is immaterial) that interacts with something material. I think the reasonable null hypothesis should be that there is not anything other than physical processes going on in the human body giving rise to everything we observe in human behavior. For me personally, I find the thought experiments useful, but unconvincing, and certainly science and technology has not supported the dualist point of view.
The Mormon viewpoint espoused by Joseph Smith seems more probable if we make the reasonable assumption that science just hasn’t yet discovered all the possible levels, states, or kinds of matter. Given the hunt for the elusive Higgs boson, I’m more open to the possibility. However, if we insist, in Mormonism, that the spirit matter cannot be detected at all, but only with “purer eyes” then I find it as unconvincing as the dualist position.
As I stated previously, no serious scientist or engineer believes we can create the correct set of circuits, processors, and algorithms and “turn on” consciousness. Rather, consciousness in robots is viewed as a possibly emergent behavior, much like many scientists believe consciousness emerged in biological bodies.
Many people are under the impression that because computers operate on a strictly discrete mathematical basis that there cannot be emergent behavior that can’t be accounted for. Such people likely do not understand the nature of modern day complex systems. When physical systems are put together to act in an uncertain world they often exhibit unpredicted behavior. There is an entire field dedicated to the science of system integration that tries to understand the compositionality and composability of various pieces of a full system. We might idealistically claim that there is always an identifiable cause of some particular system behavior. But the reality is we may not be able to identify it given our technology. The trick is, the exact same thing could be said of ANY physical system, consequence, action, or even human behavior. The practical result is that we don’t know if humans have free will, or are deterministic machines, and although we can insist that robots are completely deterministic, they appear (in some of the very same way humans do) to exhibit non-determinism.
Whether it is from this, or some other by-product of computational power and memory that consciousness emerges is unclear to me. The human brain is the most complex physical system in the known universe, and we are not anywhere near having the needed computational power, efficiency, storage, or sensors to produce that kind of information processing. Nevertheless, I do believe it possible and highly probable that consciousness will eventually emerge in robots.
Next time we’ll discuss what neuroscience and the AI community itself have to say about consciousness in robots. Later we’ll discuss a more specific engineering approach to the development of such a system.
[1] Descartes, R., “Meditations on First Philosophy,” Translated by John Veitch 1901, Reprinted by Prometheus Books, New York, NY, 1989
[2] Kurzweil, R., “The Singularity is Near,” Penguin Group, London, 2005.
[3] Seager, W., “A Brief History of the Philosophical Problem of Consciousness,” Chap 2, edited by P.D. Zelazo, M. Moscovitch and E. Thompson, The Cambridge Handbook of Consciousness, Cambridge Univ. Press, Cambridge, MA, 2007
[4] Dennett, D.C., “The Practical Requirements for Making a Conscious Robot,” Artificial Intelligence and the Mind, Vol.A349 No. 1689, 1994, pp. 133-146.
[5] James, W., “Principles of Psychology,” Holt, Rinehart, and Winston, New York, NY, 1890
[6] Freud, S., Strachey, J., and Gay, P., “An Outline of Psycho-Analysis,” W.W. Norton, New York, NY, 1989
Wheat & Tares 
Of course the real question this leads up to is whether or not a singularity will occur.
I wish to challenge your statement that you make asserting that we don’t know if we have free will or are just deterministic machines.
We know how to practically define something deterministic and something having free will. We know that in machines, as long as they function correctly (no short circuits etc.), they can only do exactly what is programmed by them to do. They inhibit no ability for “choice”. If they came upon a situation like- what flavor of ice-cream to choose, they could only pick exactly what was programmed for them to choose. We may include some complex set of instructions so that to us it may appear like they are actually choosing but in fact they make no choice.
If we are to get to the hard definitions of “choice” and “free will”, we have to assume that we know the opposite of these words, otherwise it is pointless to even debate because it all leads back to issues like- “do we even exist?”. The opposite of choice or free can be termed “automatic”. Machines are “automatic” in that they always do exactly what they are programmed or set out to do. The response of any given input for a machine is always automatic- it always happens exactly as planned barring it doesn’t have some mechanical failure or mishap.
Humans on the other hand do not exhibit this “automatic” behavior at all. We exhibit the opposite- we exhibit real choice- real free will. Take someone on a diet for example and in front of them is placed a bowl of ice-cream. They make a hard choice- a real choice as to whether to eat it or refrain from it. It is not deterministic beforehand of which choice they will make. They are wired to make either choice but at the same time that choice is not hard wired to any deterministic process in their mind or body. Now of course we can’t exactly identify “why” or even “how” this can exist other than the fact that it does in fact exist. We are thus not automatic machines- we do not exhibit the characteristics of something we know to be deterministic on any level.
Your claim seems almost as if it is a claim from ignorance. Merely stating that we cannot be sure if we have free will is not the required grounds to claim that therefore we must be deterministic. One must be able to scientifically show how or why something is exhibiting deterministic behavior. We can do that with machines because we can show scientifically why on a mathematical level how those operations exactly take place. We could program a robot to make an apparent choice on choosing oce-cream, but in reality if we were to look at each individual operation we would find a perfect deterministic cause as to how it chose vanilla over strawberry. It cannot be argued that consciousness is the same exact process.
If we are really just deterministic machines, then our entire life is nothing more than a mere perception and in reality everything that moves on our planet is just a finely tuned machine moving exactly as planned. Everything about our gospel refutes that material explanation!
I believe your only chance at creating consciousness in a robot is to actually make it be alive with biological features. But, at that point you don’t have a robot (machine) at all but have become like God himself and created “life”.
Re Rob-
I you can provide a detailed test to check your definitions (beyond ice cream anecdotes), I’d love to see it. We can see if all humans pass it, and no computers/robots pass it. I assume you’re familiar with the turing test?
There is no way to prove that a machine will function correctly, ever, period. Are you familiar with undecidable problems in computation? The research field of verification and validation of a machine attempts to demonstrate that a machine will function correctly most of the time. But at the end of the day it’s an inference. Even if all problems in CS were decidable, uncertainty in any system will prevent any propositional logic proof of correctness.
I’m not making the claim that we’re deterministic. I included quotes that claimed we are machines, but that doesn’t mean we’re deterministic. This is a false dichotomy. You seem to be claiming, if something is a machine then it’s deterministic. A robot is a machine and therefore deterministic. Humans are not machines, therefore are not deterministic. But this ignores a great deal of philosophical thought. Dennett, whom I quoted in the post is a compatibilist (as are many other philosophers as I understand it), and yet believes we are machines. Ergo, being a machine does not imply there is no free will. I also am a compatibilist. I’m definitely not a libertarian or determinist (in the philosophical sense).
But you don’t require this for other areas of the physical world. Can you scientifically show how or why the weather behaves as it does, identifying each variable that contributes to the weather on any given day? No, yet surely you don’t believe that the earth, or weather has free will? You assume (even though it’s not scientifically understood) that there are physical laws and causes that create the weather in it’s entirety. I agree with that viewpoint (the naturalist viewpoint) but I’m willing to apply it to human consciousness, and it seems like you’re not.
No one is saying it is the exact same process. No one is claiming that humans are deterministic, and therefore we can create a robot. That’s not the claim. If that were the claim, it would only be a matter of time before we understood all the details of the mind/brain and wrote the correct algorithm. But as I’ve mentioned, people who believe we’ll develop conscious robots believe consciousness will be emergent behavior as it was for biological organisms. I’m not sure why you’re so unwilling to admit that possibility.
Re Stephen-
Yeah, that is an interesting question. It certainly appears that AI will be a part of the process, though I imagine what role it would take is anyone’s guess. It seems pretty far out there, even further than conscious robots. I haven’t given any real time to research or contemplation on that topic, though it is fascinating!
Interestingly, Descartes’ dualism is by far the weakest form of dualism proposed by philosophers, yet it is the most-often used in articles like this. And even then it is misrepresented – Descartes didn’t ever say that the material was unimportant or not even central to the workings of the mind. He merely stated that there is an immaterial component to the mind that interacts with the material. If you take the position that there can be anything “emergent” from matter (whatever that means) that is unexplainable by the physical facts of the system, then you’re a dualist too.
I also have a problem with your seeming to use “physicalism” and “materialism” interchangeably. They are two similar but not identical positions. You also say there are two main positions, physicalists and dualists. This is not only inaccurate, it’s just plain wrong. Someone could be both a dualist and a physicalist and a person can be neither a physicalist nor a dualist. They’re not two opposite sides, they’re two different issues.
This is why it is imperative for scientists to study philosophy. They might take the position of Hawking and say that philosophy is dead and useless, but they don’t realize that this statement in itself is a philosophy. Every time you make a theory you’re using philosophy. Every time you make a positive statement about reality you are using philosophy. Every time you say, “I think science is descriptive of the universe but can never fully explain it,” that is a philosophy. The scientific method is itself a philosophy.
When I was accepted into grad school for experimental psychology I realized this and have been training myself very hard in philosophy, because I realized that a scientist who doesn’t know philosophy is like a public speaker who hasn’t studied English. I would really, really recommend you read Philosophy of Mind by Feser to get a survey of the issues. You might be surprised what an acceptance of eliminative materialism like Dennett truly entails – no subjective, no unity of consciousness, no first-person, no continuity of self, no qualia, no libertarian free will. Strict materialism simply doesn’t have the tools to grant any of these things, and if you had studied the philosophy behind what Dennett is saying I doubt you’d jump so readily behind what he’s saying.
Philosophy. I say the he — no, wait, that’s spinach. 😀
From what little philosophy I have studied, it seems monism and dualism are constantly reemerging in philosophical debate because both have problems which start to annoy us when either philosophy becomes prominent.
Because I’m a physicist, I find myself drawn to the idea of duality, which is something different from dualism. It says that a single thing can be described in two (or more) seemingly incompatible ways. The descriptions, when they can be made, always make the same predictions about what we will experience, but ascribe the experience to different mechanisms.
I also think free will is meaningful only in the sense that we are compelled to behave as if we actually had it, because we are never compelled to behave in a way that is foreign to our own natures or desires. To have the ability to choose something I would never choose anyway is meaningless. The potential to realize my desires, and only my desires, makes the question of whether I have “free will” purely a matter of description choice.
FireTag, I believe your concept of duality, which is a very valid position to hold, falls under the term “property dualism” in philosophy.
And yes, both monism and dualism are both annoying, but I think what’s more annoying right now is that we’re in a period of physicalist monism being portrayed as the ONLY reasonable scientific framework, and people don’t even know that it has major problems. They basically proof-text strict materialist monism into all our past scientific discoveries and assume that will be the future, too. They don’t realize that idealist monism, platonist dualism, property dualism, etc., ALL fit the data we have.
It would be as if in 100 years the many-worlds hypothesis was “accepted” as the only true interpretation of quantum mechanics, and anyone who holds a different position is unreasonable. Barring some new scientific discovery, the many-worlds hypothesis is just one of several possible theories that fit the data we have exactly, and picking one over the other is really more a matter of temperament and personality and not necessarily data.
jmb 275,
Emergent behavior? Emerging from what though? Do you honestly belive thta the more sophisticated we make machines and computers that at some point they will begin to emerge consciousness?
The turing test is a load of BS. We could make a machine that could look and act very similar to a human even to the point where someone would have a hard time knowing if it was human or not. But does that mean it is conscious? By no means.
A machine will operate only exactly as it is programmed and made. Now for us it may seem it doesn’t act in this manner but it has to. Machines have to obey exactly what they are made to do. Now if there is an error of input on our part that is our fault and not the machine. If an operator programs it to run a math operation that has no real answer then of course it can’t answer that operation. But, it will still operate exactly according to the input it is given every time.
I think we can all agree that our bodies act like machines in their basic operations- heart pumping, muscle movement, etc. But to imply that this goes right down to the explaining consciousness level is unbelievable. Where is there proof?
This whole thought experiment on robotic consciousness is interesting but at the very heart of the matter is the practicality of what we know as laws and principles within our system. Those laws are pretty well understood when it comes to machinery, computers, etc. From the myriad of tests and experiments, we know that computational language coupled with nonliving contraptions such as man-made machinery have no level of detectable consciousness. We also know that this same principle does not explain how humans make decisions. Our unique ability to make decisions goes beyond the capacity of any computational language we can program into a computer. We can then infer from that principle alone, that the current way that machines and comptuers are made and work upon, does in no way explain how humans make decisions. Which leads back tot he very basic question to begin with-
How can we possibly develop consciousness in a machine or any other man-made contraption, if we have no clues whatsoever as to how humans do it? Obviously, all the computational properties of machinery to this point have gotten us no closer than we have ever been to finding that answer.
This leads me back to the beginning- How is it even possible that consciousness could emerge from something non-living? Certainly sophistication in machinery has gotten us no closer. Perhaps through some other medium besides computers and man-made machinery?
Syphax:
I’m largely in agreement with you. This is a monist period, and part of the reason I like duality is that it says there is always going to be more than one way to describe these issues.
Rob: How does wetness emerge as a property of groups of water molecules that are individually not wet? Yet, you certainly don’t deny that water is wet?
Suppose that the potential for consciousness is inherent in everything, just as every piece of matter as the potential to hold charge or energy. You’re being awfully arbitrary about what can and cannot be.
Re Syphax-
I feel like we talked a bit about this last time. I’m convinced this is a perspective issue, and the fact is, you just disagree with me, and equate that with either my ignorance or wrongness. Although I have not read Feser, I’ve read and studied the philosophy of mind at some level. It’s certainly not wrong to claim that dualism and physicalism are the two main schools of thoughts. Perhaps you prefer monism instead of physicalism. Monism is a loaded term in philosophy so I choose not to use it. Nevertheless, I know of prominent philosophers who frame the philosophy of mind problem in the dualism vs. physicalism framework. You can disagree with me, and so can Feser, but as far as I can tell, in philosophy that doesn’t necessarily make me wrong.
Your point about the interchangeable use of physicalism and materialism are duly noted, though I don’t think I used the term materialism anywhere, but rather refer to “material.”
FWIW, Wikipedia has this to say
This statement matches my experience in my own independent research into the issue. I don’t doubt others see it differently.
Perhaps if I can move beyond Feser’s arrogance I’ll give it a whirl. I think you’ve misrepresented Dennett, though I admit to not having studied his ideas much. Dennett claims to be a compatibilist, so he does not reject free will. I don’t know about you, but I definitely don’t hold to libertarianism exclusively. I join the ranks of the compatibilists. Oh, and ouch, I really feel like given my career choice I’ve given philosophy of mind a fair amount of my time (books, classes, papers, etc.). So it hurts a bit suggest that I haven’t studied it. Nevertheless, I don’t claim to be a philosopher.
As a rant on philosophy, I’ll say that the community and language surrounding philosophy is almost enough to drive me away from even pursuing it as a hobby. The underlying propositional logic upon which all philosophy rests creates a culture where perspective differences are framed as right or wrong. And yet, from what I can tell, philosophers have all kinds of mental gymnastics and thought experiments they can use to believe just about anything. Strangely, in the engineering fields I don’t find the same culture despite the underlying mathematical foundation. I suppose it is because our hypotheses are more testable by impartial mechanisms.
firetag,
No I am being intellectually honest. We know that consciousness is only observed in living biologic forms, not man-made machinery. No emergence of consciousness has ever been observed in all the billions of man-made contraptions ever invented, even in highly sophisticated contraptions.
Thus I am only stating the facts of what hasn’t ever been observed. If we are going to make conscious robots, we have to take an entirely different route because nothing we have done in the robotic technology department so far has brought us any closer than “0” to bringing about consciousness in man-made machines.
Re Rob-
Thanks for commenting. I think there are plenty of people who see things the way you do. I have no intention of convincing you that I’m right. Good luck!
Re Syphax
I completely agree. It is sad. I don’t hold that physicalist monism is the ONLY reasonable solution. But I think there are a few things to point out. Scientists have little to work with if we start with the hypothesis espoused by platonic dualism. It’s hard to measure immaterial things, much less ascertain their interaction with material things.
Personally, of the more nuanced positions on philosophy of mind, I think emergent materialism fits the data the best, as it appropriately takes into account evolution, which is a sound scientific theory.
Besides that, as I said in the post, it’s hard for the concept of conscious robots to even get off the ground if we don’t admit a materialist POV.
Maybe you should look up platonic dualism. Science would collapse completely if it didn’t deal with platonic objects such as perfect triangles, math and physics equations, etc. Ask yourself, “where” in space and time is the number 7? Where does a perfectly straight line exist anywhere in the universe? Nowhere – these are imaginary, mathematical entities. Mathematics is a completely esoteric language that doesn’t exist in space and time, and yet we can manipulate these objects in our minds, add them, subtract them, find their square roots, etc. This is the Platonic realm. So to say that science wouldn’t get off the ground if we postulate the Platonic realm seems backwards: science wouldn’t get off the ground WITHOUT the platonic realm. So in this case, it’s easier to measure an immaterial thing than a material thing! The platonic number 7 is always exactly 7 and will never be anything else. Everything in the material world is messy and harder to measure.
If you can take measurements of matter, move them into the world of mathematical equations, manipulate them and change them in your mind, and then plug them back into the world of matter, you have demonstrated that you have used an immaterial realm to influence the material realm.
How can we possibly develop consciousness in a machine or any other man-made contraption, if we have no clues whatsoever as to how humans do it? Excellent question. We could program a robot to make an apparent choice on choosing oce-cream, but in reality if we were to look at each individual operation we would find a perfect deterministic cause as to how it chose vanilla over strawberry. If you were able to know your human conscious plus unconscious decision making logic you would discover how you made the vanilla vs. strawberry choice (magic is not involved) and at that point it would be easily programmable except for an emotion generator which with the exception of empathy is probably an undesirable feature.
Interesting post. I think these are question we really may one day face. My question is when will we baptize robots.
You may be interested in this paper I wrote on consciousness, especially some of the sources I draw on.
Howard,
This is really where free will comes into play. I don’t believe in destiny on things. How I make decisions is not something that can be written. This is what makes robot consciousness in my opinion- impossible. Intelligence is not any form of written equation.
“My question is when will we baptize robots.”
Funny stuff indeed!
Re SteveP-
Thanks for stopping by! That article looks great. I’m anxious to read it.
Re Syphax-
I’m sorry I’m not as trained in philosophy as you appear to be. I have devoted my career to research in computation and engineering, so philosophy is just sort of a hobby.
I looked for dualism, and more specifically platonic dualism here (which seems to be a decent source). Platonic dualism necessarily implies (to me anyway) that the soul and body are separable, that the soul existed before birth, etc. In my #13 I suggested that it was hard to measure immaterial things. How would you propose we scientifically demonstrate that the soul existed before birth? What would you say is the implication for platonic dualism in terms of material? Surely it’s not reasonable to conclude that platonic dualism implies that the mind and brain are the same (no dualism)? Does platonic dualism admit the possibility that the mind/body are both material but still separable, one of which existed before birth and lives on after death? If not, why was my statement incorrect?
Again, I didn’t say this. You said this. I started with platonic dualism, which appears to me to imply a separable mind/body situation. But you didn’t confront that issue. Instead you launched into a discussion about platonism generally (which is not the topic at hand) and used that to imply that I don’t know what I’m talking about since mathematics rests upon platonic ideas. But it’s strange because I would agree exactly with what you said about platonism and mathematics. I’m the first one to harp on mathematicians and physicists who believe that mathematics can or do describe the world perfectly.
Now this is interesting. But I think it’s worth challenging a bit. Materialists wouldn’t argue that the mind (collection of thoughts, emotions, etc. both cognitive and subconscious) is immaterial. What they argue is that it arises from the brain. The dualist argues that the mind/soul is separate from the brain (i.e. does not arise from the neurons and synapses firing) but is actually a separate thing (in fact a different “substance” if you’re a substance dualist). As a result, of course the immaterial (thoughts, etc.) influence the material realm.
It’s worth mentioning, that in this light, computers do this all the time. They take measurements of matter, make calculations based on measurements (that assume platonic entities), and plug them back into the world of matter, possibly even by moving an actuator of some kind. But I doubt this will convince you that robots are dualistic in nature.
In SteveP’s paper he made this statement which I find very relevent- very well worded-
“Our human in telligence is biased away from understanding consciousness. It is not that consciousness is objectively any
more complex than the things we can understand; it’s just that our faculties
are not cut out to penetrate to its un derlying nature.”
I believe that pretty much nails it on it’s head. This is where I see technology directed at the material things and scientists using these material observable objects and trying to apply them to an understanding of consciousness. The problem is that the cause of consciousness doesn’t seem to be a product of the typical areas of material objects and speculation. We tend to want to look at the material things we produce as being somehow capable of conscious abilities and thus we get biased in believeing that consciousness can indeed be achieved in ways similar to how computers or machinery operate. It is a bias for sure.
I have just converted to arguing that machines will be conscious oneday. When I was changing my mind this is how it felt… I thought that I had been indoctrinated by the culture of dualism (from previous centuries) and also indoctrinated by the old-fashioned view of machines. (Frankenstein and infinite less famous depictions of conscious machines in western culture). Moreover I felt that our opposition to the possibility of machine consciousness is like (and will be seen to be like) peoples doubts about the reliabiity of microwave ovens in the very early 1980s. They cause cancer people said. This latter point about microwaves is analogous to uploading consciousness into a machine today. We will die many of us think now. But thats todays perspective which is ignorant relative to the year 2045’s perspective. Moreover it is an old fashioned view of technology which is drowned in dualism. Even now that I have converted I think that I need to mature my views on this… its hard to just entirely shrug off out-dated views that you have consciously held for all of your life.