Online research in philosophy

In this first part of a two-part paper, we describe efforts in the early decades of this century to restrict the extent of violations of the Second Law of thermodynamics that were brought to light by the rise of the kinetic theory and the identification of fluctuation phenomena. We show how these efforts mutated into Szilard's (1929) proposal that Maxwell's Demon is exorcised by proper attention to the entropy costs associated with the Demon's memory and information acquisition. In the second part we will argue that the information theoretic exorcisms of the Demon provide largely illusory benefits. According to the case, they either return a presupposition that can be had without information theoretic consideration or they postulate a broader connection between information and entropy than can be sustained.

The classical arguments for scepticism about the external world are defended, especially the symmetry argument: that there is no reason to prefer the realist hypothesis to, say, the deceitful demon hypothesis. This argument is defended against the various standard objections, such as that the demon hypothesis is only a bare possibility, does not lead to pragmatic success, lacks coherence or simplicity, is ad hoc or parasitic, makes impossible demands for certainty, or contravenes some basic standards for a conceptual or linguistic scheme. Since the conclusion of the sceptical argument is not true, it is concluded that one can only escape the force of the argument through some large premise, such as an aptitude of the intellect for truth, if necessary divinely supported.

According to fitting-attitudes accounts of value, the valuable is what there is sufficient reason to value. Such accounts face the famous wrong kind of reason problem. For example, if an evil demon threatens to kill you unless you value him, it may appear that you have sufficient reason to value the demon, although he is not valuable. One solution to this problem is to deny that the demon’s threat is a reason to value him. It is instead a reason to want to value the demon, and to bring it about that you value him. However, many proponents of the wrong kind of reason problem find this solution unmotivated. This paper thus offers a new argument for this solution. The argument turns on the ‘transmission’ of reasons – the familiar fact that there is often reason for one action or attitude because there is reason for another. I observe that putative reasons of the wrong kind transmit in a very different way to other reasons. I then argue that this difference is best explained by the hypothesis that putative reasons of the wrong kind are not reasons for the attitude in question, but are instead reasons to want and bring about that attitude.

In this second part of our two-part paper we review and analyse attempts since 1950 to use information theoretic notions to exorcise Maxwell's Demon. We argue through a simple dilemma that these attempted exorcisms are ineffective, whether they follow Szilard in seeking a compensating entropy cost in information acquisition or Landauer in seeking that cost in memory erasure. In so far as the Demon is a thermodynamic system already governed by the Second Law, no further supposition about information and entropy is needed to save the Second Law. In so far as the Demon fails to be such a system, no supposition about the entropy cost of information acquisition and processing can save the Second Law from the Demon.

In this second part of our two-part paper we review and analyse attempts since 1950 to use information theoretic notions to exorcise Maxwell’s Demon. We argue through a simple dilemma that these attempted exorcisms are ineffective, whether they follow Szilard in seeking a compensating entropy cost in information acquisition or Landauer in seeking that cost in memory erasure. In so far as the Demon is a thermodynamic system already governed by the Second Law, no further supposition about information and entropy is needed to save the Second Law. In so far as the Demon fails to be such a system, no supposition about the entropy cost of information acquisition and processing can save the Second Law from the Demon.

In this paper we show how a form of Maxwellian Demon can be interpreted as a computing automaton. We then point out some ways in which the Demon systems can be generalized, and briefly describe and discuss the properties of some of the corresponding automata. It is shown that a generalized Maxwell Demon system can carry out arbitrary Turing computations. Finally, the association developed between classes of thermodynamic systems and classes of computational systems is employed to suggest approaches to some fundamental problems of the relationships between computation, the information obtained by computation, and energy.

In my master's thesis for physics and philosophy, I take a long and hard look at the debates surrounding Maxwell's Demon and the status of the second law of thermodynamics. I try to clarify the use of Maxwell's thought experiment in understanding the second law; to prove that the second law is contingent, given only classical mechanics and time asymmetry; to argue that the law only holds because of facts about the kinds of particles that exist in our universe; to show that and why the attempts to banish the demon using fluctuations, measurements or information or erasure have been unsuccessful; and I conclude that Maxwell's Demon is alive and kicking.

"Maxwell's Demon", the famous thought experiment of James Clerk Maxwell, has been devised in 1867 as a counter example for the Second Law of thermodynamics. During the 140 years since the Demon was first suggested, numerous attempts have been made to counter Maxwell's argument. The attempts have been to show that Maxwell was wrong, since his Demon cannot work for one reason or another (see Leff and Rex 2003 for details and references). In this paper we show (following an argument by Albert 2000, Ch. 5.) that Maxwell was basically right, in the sense that his thought experiment is compatible with the laws of mechanics as well as with central principles of statistical mechanics. We then derive some (weak) restrictions on the Demon's efficiency. Finally, we prove that the Demon's cycle of operation can be completed (in particular, the Demon's memory can be erased) without increasing the total entropy of the universe. We draw some conclusions about the way to understand the meaning and role of probability in classical statistical mechanics.

The New Evil Demon Problem is supposed to show that straightforward versions of reliabilism are false: reliability is not necessary for justification after all. I argue that it does no such thing. The reliabilist can count a number of beliefs as justified even in demon worlds, others as unjustified but having positive epistemic status nonetheless. The remaining beliefs---primarily perceptual beliefs---are not, on further reflection, intuitively justified after all. The reliabilist is right to count these beliefs as unjustified in demon worlds, and it is a challenge for the internalist to be able to do so as well.

The New Demon World Objection claims that reliabilist accounts of justification are mistaken because there are justified empirical beliefs at demon worlds—worlds at which the subjects are systematically deceived by a Cartesian demon. In this paper, I defend strongly verific (but not necessarily reliabilist) accounts of justification by claiming that there are two ways to construct a theory of justification: by analyzing our ordinary concept of justification or by taking justification to be a theoretic term defined by its role in the theory of knowledge. The former route is not promising because of the splintered nature of our ordinary concept of justification—or perhaps because there is no single such concept. On the other hand, if justification is defined by the role it plays in the theory of knowledge, then there is good reason to think that justification must be strongly truth-conductive since the term was introduced by Edmund Gettier to play the primary role in converting true belief into knowledge. And if that is right, then there will be no justified empirical belief at demon worlds. The real lesson of the demon world is then turned on its head: justification does not supervene on what one shares with one’s deceived doppelganger.