David Chalmers

What happens when machines become more intelligent than humans? One view is that this event will be followed by an explosion to ever-greater levels of intelligence, as each generation of machines creates more intelligent machines in turn. This intelligence explosion is now often known as the “singularity”. The basic argument here was set out by the statistician I.J. Good in his 1965 article “Speculations Concerning the First Ultraintelligent Machine”: Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an “intelligence explosion”, and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make. The key idea is that a machine that is more intelligent than humans will be better than humans at designing machines. So it will be capable of designing a machine more intelligent than the most intelligent machine that humans can design. So if it is itself designed by humans, it will be capable of designing a machine more intelligent than itself. By similar reasoning, this next machine will also be capable of designing a machine more intelligent than itself. If every machine in turn does what it is capable of, we should expect a sequence of ever more intelligent machines. This intelligence explosion is sometimes combined with another idea, which we might call the “speed explosion”. The argument for a speed explosion starts from the familiar observation that computer processing speed doubles at regular intervals. Suppose that speed doubles every two years and will do so indefinitely. Now suppose that we have human-level artificial intelligence 1 designing new processors. Then faster processing will lead to faster designers and an ever-faster design cycle, leading to a limit point soon afterwards. The argument for a speed explosion was set out by the artificial intelligence researcher Ray Solomonoff in his 1985 article “The Time Scale of Artificial Intelligence”.1 Eliezer Yudkowsky gives a succinct version of the argument in his 1996 article “Staring at the Singularity”: “Computing speed doubles every two subjective years of work..

Hilary Putnam has argued that computational functionalism cannot serve as a foundation for the study of the mind, as every ordinary open physical system implements every finite-state automaton. I argue that Putnam's argument fails, but that it points out the need for a better understanding of the bridge between the theory of computation and the theory of physical systems: the relation of implementation. It also raises questions about the class of automata that can serve as a basis for understanding the mind. I develop an account of implementation, linked to an appropriate class of automata, such that the requirement that a system implement a given automaton places a very strong constraint on the system. This clears the way for computation to play a central role in the analysis of mind.

In article <[email protected]> [email protected] writes: Reminds me of a friend of mine who claims that the number 17 is "the most random" number. His proof ran as follows: pick a number. It's not really as good a random number as 17, is it? (Invariable Answer: "Umm, well, no...") This reminds me of a little experiment I did a couple of years ago. I stood on a busy street corner in Oxford, and asked passers by to "name a random number between zero and infinity." I was wondering what this "random" distribution would look like.

The Matrix presents a version of an old philosophical fable: the brain in a vat. A disembodied brain is floating in a vat, inside a scientist’s laboratory. The scientist has arranged that the brain will be stimulated with the same sort of inputs that a normal embodied brain receives. To do this, the brain is connected to a giant computer simulation of a world. The simulation determines which inputs the brain receives. When the brain produces outputs, these are fed back into the simulation. The internal state of the brain is just like that of a normal brain, despite the fact that it lacks a body. From the brain’s point of view, things seem very much as they seem to you and me.

It is fairly well-known that certain hard computational problems (that is, 'difficult' problems for a digital processor to solve) can in fact be solved much more easily with an analog machine. This raises questions about the true nature of the distinction between analog and digital computation (if such a distinction exists). I try to analyze the source of the observed difference in terms of (1) expanding parallelism and (2) more generally, infinite-state Turing machines. The issue of discreteness vs continuity will also be touched upon, although it is not so important for analyzing these particular problems.

At the April 2006 meeting of the Central Division of the American Philosophical Association, in an author-meets-critics session on Scott Soames' book _Reference and Description: The Case Against Two-Dimensionalism_ , I presented a comment on Soames' book, "Scott Soames' Two-Dimensionalism" . The other critic was Robert Stalnaker. Soames presented his response to critics . Below is a reply to Soames' response to me, for those who were at the session and interested others. Note that this response was mostly written before the session, except for one or two paragraphs where the discussion in the session is mentioned

A number of popular arguments for dualism start from a premise about an epistemic gap between physical truths about truths about consciousness, and infer an ontological gap between physical processes and consciousness. Arguments of this sort include the conceivability argument, the knowledge argument, the explanatory-gap argument, and the property dualism argument. Such arguments are often resisted on the grounds that epistemic premises do not entail ontological conclusion. My view is that one can legitimately infer ontological conclusions from epistemic premises, if one is very careful about how one reasons. To do so, the best way is to reason first from epistemic premises to modal conclusions , and from there to ontological conclusions. Here, the crucial issue is the link between the epistemic and modal domains. How can one reason from theses about what is knowable or conceivable to theses about what is necessary or possible? To bridge the epistemic and modal domains, the framework of two-dimensional semantics can play a central role. I have used this framework in earlier work to mount an argument against materialism. Here, I want to revisit the argument, laying it out in a more explicit and careful form, and responding to a number of objections. In what follows I will concentrate mostly on the conceivability argument. I think that very similar considerations apply to the other arguments mentioned above, however. In the final section of the paper, I show how this analysis might yield a unified treatment of a number of anti-materialist arguments

It's very interesting to see neurophysiological evidence brought to bear on the puzzling question of conscious experience. Many have observed that information-processing models of cognition seem to leave consciousness untouched; it is natural to hope that turning to neurophysiology might lead us to the Holy Grail. Still, I think there are reasons to be skeptical. There are good reasons to suppose that neurophysiological investigation contributes to cognitive explanation at best in virtue of constraining the information-processing structure of cognition. Of course this is a very large and significant role for it to play, but it may be over-optimistic to suppose that it can play some further explanatory role, taking us where information-processing theories cannot. If so, then neurophysiological accounts will be no more and no less successful at dealing with consciousness than information-processing accounts are.

This paper offers both a theoretical and an experimental perspective on the relationship between connectionist and Classical (symbol-processing) models. Firstly, a serious flaw in Fodor and Pylyshyn’s argument against connectionism is pointed out: if, in fact, a part of their argument is valid, then it establishes a conclusion quite different from that which they intend, a conclusion which is demonstrably false. The source of this flaw is traced to an underestimation of the differences between localist and distributed representation. It has been claimed that distributed representations cannot support systematic operations, or that if they can, then they will be mere implementations of traditional ideas. This paper presents experimental evidence against this conclusion: distributed representations can be used to support direct structure-sensitive operations, in a man- ner quite unlike the Classical approach. Finally, it is argued that even if Fodor and Pylyshyn’s argument that connectionist models of compositionality must be mere implementations were correct, then this would still not be a serious argument against connectionism as a theory of mind.

This paper explores how an evolutionary process can produce systems that learn. A general framework for the evolution of learning is outlined, and is applied to the task of evolving mechanisms suitable for supervised learning in single-layer neural networks. Dynamic properties of a network’s information-processing capacity are encoded genetically, and these properties are subjected to selective pressure based on their success in producing adaptive behavior in diverse environments. As a result of selection and genetic recombination, various successful learning mechanisms evolve, including the well-known delta rule. The effect of environmental diversity on the evolution of learning is investigated, and the role of different kinds of emergent phenomena in genetic and connectionist systems is discussed.

It is widely accepted that conscious experience has a physical basis. That is, the properties of experience (phenomenal properties, or qualia) systematically depend on physical properties according to some lawful relation. There are two key questions about this relation. The first concerns the strength of the laws: are they logically or metaphysically necessary, so that consciousness is nothing "over and above" the underlying physical process, or are they merely contingent laws like the law of gravity? This question about the strength of the psychophysical link is the basis for debates over physicalism and property dualism. The second question concerns the shape of the laws: precisely how do phenomenal properties depend on physical properties? What sort of physical properties enter into the laws' antecedents, for instance; consequently, what sort of physical systems can give rise to conscious experience? It is this second question that I address in this paper.

This appeared in Philosophy and Phenomenological Research 59:473-93, as a response to four papers in a symposium on my book The Conscious Mind . Most of it should be comprehensible without having read the papers in question. This paper is for an audience of philosophers and so is relatively technical. It will probably also help to have read some of the book. The papers I’m responding to are: Chris Hill & Brian McLaughlin, There are fewer things in reality than are dreamt of in Chalmers’ philosophy Brian Loar, David Chalmers’ The Conscious Mind Sydney Shoemaker, On David Chalmers’ The Conscious Mind Stephen Yablo, Concepts and consciousness Contents.

Inspired by Rudolf Carnap's Der Logische Aufbau Der Welt, David J. Chalmers argues that the world can be constructed from a few basic elements. He develops a scrutability thesis saying that all truths about the world can be derived from basic truths and ideal reasoning. This thesis leads to many philosophical consequences: a broadly Fregean approach to meaning, an internalist approach to the contents of thought, and a reply to W. V. Quine's arguments against the analytic and the a priori. Chalmers also uses scrutability to analyze the unity of science, to defend a conceptual approach to metaphysics, and to mount a structuralist response to skepticism. Based on the 2010 John Locke lectures, Constructing the World opens up debate on central philosophical issues involving language, consciousness, knowledge, and reality. This major work by a leading philosopher will appeal to philosophers in all areas. This entry contains uncorrected proofs of front matter, chapter 1, and first excursus.

Where does the mind stop and the rest of the world begin? The question invites two standard replies. Some accept the demarcations of skin and skull, and say that what is outside the body is outside the mind. Others are impressed by arguments suggesting that the meaning of our words "just ain't in the head", and hold that this externalism about meaning carries over into an externalism about mind. We propose to pursue a third position. We advocate a very different sort of externalism: an _active externalism_, based on the active role of the environment in driving cognitive processes

Frank Ramsey (1931) wrote: If two people are arguing 'if p will q?' and both are in doubt as to p, they are adding p hypothetically to their stock of knowledge and arguing on that basis about q. We can say that they are fixing their degrees of belief in q given p. Let us take the first sentence the way it is often taken, as proposing the following test for the acceptability of an indicative conditional: ‘If p then q’ is acceptable to a subject S iff, were S to accept p and consider q, S would accept q. Now consider an indicative conditional of the form (1) If p, then I believe p. Suppose that you accept p and consider ‘I believe p’. To accept p while rejecting ‘I believe p’ is tantamount to accepting the Moore-paradoxical sentence ‘p and I do not believe p’, and so is irrational. To accept p while suspending judgment about ‘I believe p’ is irrational for similar reasons. So rationality requires that if you accept p and consider ‘I believe p’, you accept ‘I believe p’.

I would like to thank the authors of the 26 contributions to this symposium on my article “The Singularity: A Philosophical Analysis”. I learned a great deal from the reading their commentaries. Some of the commentaries engaged my article in detail, while others developed ideas about the singularity in other directions. In this reply I will concentrate mainly on those in the first group, with occasional comments on those in the second. A singularity (or an intelligence explosion) is a rapid increase in intelligence to superintelligence (intelligence of far greater than human levels), as each generation of intelligent systems creates more intelligent systems in turn. The target article argues that we should take the possibility of a singularity seriously, and argues that there will be superintelligent systems within centuries unless certain specific defeating conditions obtain