Sunday, April 3, 2016

John Searle on Consciousness in Artificial Intelligence

John Searle, Slusser Professor of Philosophy at the University of California (Berkeley), gives a Google talk below on consciousness in artificial intelligence.

John Searle is a great analytic philosopher in the tradition of Bertrand Russell, and his work on AI and consciousness is particularly interesting. This talk is really insightful.

Searle is summarising arguments from his paper “Minds, Brains, and Programs” (1980) and his book Consciousness and Language (2002).

Searle makes a very interesting distinction between observer-relative and observer-independent objects. This is actually the same distinction made by Karl Popper between World 3 objects and World 1 objects.

Searle also points to two important principles:
(1) syntax is not semantics, and

(2) simulation is not duplication.
Turing machines are things designed to automatically process symbols by means of set syntactic rules or algorithms, but with no understanding of the symbols. Computers, then, are automated syntactical systems manipulating symbols but devoid of semantics.

Searle also rightly notes two senses of the concept of “intelligence”:
(1) an observer-relative sense of intelligence that Turing machines can have when they automatically process symbols by means of set syntactic rules or algorithms to create output from input, and

(2) the kind of intelligence with consciousness, with perception, sensation, and conscious experience, of the higher animal minds.
Sense (2) is observer-independent, intrinsic and internal.

What of computation? Searle argues that computation is not intrinsic to machines. The same distinction between observer-relative and observer-independent phenomenon can be applied to computation. People can engage in observer-independent and intrinsic computation, just as Turing’s human computers. But machine computation is observer-relative. As Popper would say, what the machine does considered as a World 1 process is not computation, but is a set of mere physical World 1 processes. The machine’s functioning becomes computation in World 3 because it takes human beings to recognise it as such and our interpretation of its physical operation. It also lacks observer-independent, intrinsic conscious intelligence.

This is perhaps the weakest point of the argument, for what about natural types of information processing as in DNA? Clearly, there must be types of natural information in World 1 that have emerged by Darwinian evolution.

But that natural information by itself or computation in Turing machines is not a sufficient condition for consciousness.

Searle also reviews the biological naturalist theory of the mind, and he notes that the creation of a truly artificial intelligence like ours would be analogous to the creation of an artificial heart. It does not matter how good your simulation of a human heart is on a computer, it does not pump blood and it is not an actual heart. An artificial system that reproduces what a heart is and does in the human body needs to reproduce its causally necessary physical attributes, even if it is not organic but synthetic. It is the same with the mind. You need to duplicate the human mind, not simulate it. Whether you need the exactly same kind of physical and biological processes in the brain, or whether it can be done with different biochemical processes or even synthetic materials is currently unknown.

In essence, Turing machines and computers are as dead, as unconscious, as senseless as rocks.

By contrast, even the lower animals have some degree of observer-independent and internal conscious intelligence, and certainly the higher animals do, just as this dog at the end of the video below.

Searle, John. R. 1980. “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3: 417–424.

Searle, John R. 2002. Consciousness and Language. Cambridge University Press, New York.


  1. Okay, one of the arguments you must face is that if neurons function as a kind of TM then we aren't talking about simulation at all, right?

    And as your aside about DNA suggests, information and information processing are ubiquitous in biology today. You and Searle's seem quite unaware of how pervasive. And Callahan too of course.

  2. (1) "one of the arguments you must face is that if neurons function as a kind of TM then we aren't talking about simulation at all, right?

    Right -- if all it takes for conscious intelligence is very complex information processing on a TM and this is necessary and sufficient condition for consciousness, no matter what the substrate is.

    However, even if neurons function as a kind of TM, it does not necessarily follow that digital computers can reproduce consciousness, since being a TM may be a necessary but *not* sufficient condition for consciousness.

    Anyway, critics of the computational theory of mind are arguing: consciousness like superconductivity, or digestion, is a physical process, just simulating cannot reproduce it.

    You need the causally necessary physical processes.

    (2) "You and Searle's seem quite unaware of how pervasive. "

    Oh, Rubbish, Ken B. I already made it clear I find that aspect of his thought weak.

    But clearly mere information processing in nature, as in DNA, is
    necessary but *not* sufficient condition for consciousness either. So this doesn't affect the biological naturalist theory of the mind, and certainly doesn't affect, say, the conscious electromagnetic information (CEMI) field theory of mind.

    And if we are talking about ignorance, Ken B, perhaps you ought to have read up on this subject yourself?

    The role of neuronal electromagnetic fields (or local field potentials or synchronous EM field oscillations) in cognition and consciousness is absolutely cutting edge, exiting mainstream neuroscientific research going on right now.

    See the work of David McCormick and Christof Koch and Johnjoe McFadden.

    E.g., the synchronicity of neuronal firing, a process which generates the internal neuronal electromagnetic fields, seems to be correlated with perception in the visual cortex (McFadden 2013: 155).

    The research literature reports that the firing of neurons in synchrony, which generate these fields, is the best correlate of consciousness yet discovered (McFadden 2013: 156).

    McFadden, Johnjoe. 2013. “The CEMI Field Theory: Closing the Loop,” Journal of Consciousness Studies 20.1–2: 153–168.

    Also, it seems that our internal neuronal electromagnetic fields are not passive by-products of the activity of neurons.

    (1) The neuronal electromagnetic fields interact with neurons and could allow a type of inter-neuronal communication system and they also increase the synchrony of neurons firing together.

    “The perpetual fluctuations of these extracellular fields are the hallmark of the living and behaving brain in all organisms, and their absence is a strong indicator of a deeply comatose, or even dead, brain," Anastassiou explains.”
    Kathy Svitil, “Neurobiologists Find that Weak Electrical Fields in the Brain Help Neurons Fire Together.” Caltech, 2 February, 2011

    (2) the neuronal electromagnetic fields have feedback effect and seem to induce internal synchronization of neurons and might also have some role during sleep:

    Ferris Jabr, “Cortical Call Out: The Brain’s Electric Field Creates a Feedback Loop That Synchronizes Neural Activity,” Scientific American, July 22, 2010.

  3. LK, that's like saying driving cars isn't travelling but just simulation of the legs. It's as good as a joke. :)

    1. Driving a car is another output from Turing machines automatically processing symbols by means of set syntactic rules or algorithms, but with no understanding of the symbols.

      Can your desktop computer digest food? Or reproduce photosynthesis?

    2. Look back to your response when I took a shot about digestion ...

    3. (1) you simply claimed, without evidence, it is a "prejudice" to think that consciousness requires causally necessary physical/biological processes.

      NOT an answer.

      (2) then you have a confused comment here:

      If a machine has been provided with a special system that actually does digest food, then you prove my point. An ordinary digital computer can simulate digestion in some model, but not *reproduce* digestion. It would require a special system to duplicate the necessary physical/biological processes to cause digestion. We can inspect that internal system and say whether it is duplicating the functions and biological processes of a stomach.

      You haven't refuted me, Ken B.

    4. So you are arguing that only exact duplication of what we have now can have consciousness, as a matter of principle? But my claims about prejudice are unwarranted?

    5. Once again, Ken B, you seem to arguing in bad faith.

      I do not argue that the biological naturalist theory of the mind is 100% true, and cannot questioned, and must be accepted as a matter of faith. If anything, you seem to be displaying a crude faith in the CTM.

      I say:

      (1) the computational theory of mind (CTM) has serious problems, is far from proven in the same sense that the proposition that "the earth revolves around the sun" has been inductively proven, and, moreover, the science is far from settled.

      (2) until we have a proven theory of consciousness as well proven as the proposition that "the earth revolves around the sun" these things are still debatable.

      (3) **if** it is the case that we discover that CTM is wrong and that consciousness requires necessary causal physical processes, then Turing machines can never attain consciousness.

      (4) I have given you an interesting new biological naturalist theory of the mind: conscious electromagnetic field information (CEMI) theory.
      If CEMI is true, or even another biological naturalist theory of the mind (e.g., consciousness emerges by physically necessary patterns of synchronous neural network firing), then CTM is screwed.

      You're the one, it seems, militantly incapable of even considering for one second that your dogma of CTM could be wrong. I am skeptical.

      You scream "heresy," "prejudice," "supernaturalism"!

    6. In earlier posts, you also seem to militantly deny the plain fact that Turing used a naïve behaviourist method for the Imitation Game and was unconcerned by

      (1) the need to define intelligence in clear, sensible ways (and in place if that using the naïve behaviourist Imitation Game), and

      (2) the crucial need to scientifically understand human conscious intelligence before deciding on whether you can declare machines capable of attaining true human conscious intelligence.
      Poor show, Ken B.

    7. Tell me something else, Ken B.

      Do you doubt or have serious criticisms of the current consensus on global warming, or the role and extent of anthropogenic gases in driving warming? (with some evidence provided).

      If so, and I screamed "science-denier!", "this is just right-wing prejudice!", "anti-science!" at you, would you agree:

      (1) that, yes, you are anti-science
      (2) that your views are just a prejudice, and
      (3) that such responses by me actually provide a serious counterargument to your argument?
      I am guessing you'd vehemently claim that my insults are **not an answer** and **not an argument**.

      And yet you use the same shoddy and flawed tactic right here in the comments section. Interesting.

  4. Can anyone explain to me why Searle says consciousness is observer independent? It seems by definition it requires at least one observer: yourself.

    He alluded to this in other places when he says "consciousness can't be dismissed as an illusion because to even have the illusion of being conscious...would be to be conscious by definition."

    But if consciousness is like this, it would seem observer dependent. The observer being yourself.