Searle's abstract argument against strong AI

Discussion of Searle's case against strong AI has usually focused upon his Chinese Room thought-experiment. In this paper, however, I expound and then try to refute what I call his abstract argument against strong AI, an argument which turns upon quite general considerations concerning programs, syntax, and semantics, and which seems not to depend on intuitions about the Chinese Room. I claim that this argument fails, since it assumes one particular account of what a program is. I suggest an alternative account which, however, cannot play a role in a Searle-type argument, and argue that Searle gives no good reason for favoring his account, which allows the abstract argument to work, over the alternative, which doesn't. This response to Searle's abstract argument also, incidentally, enables the Robot Reply to the Chinese Room to defend itself against objections Searle makes to it.     

Artificial intelligence and personal identity

Considerations of personal identity bear on John Searle's Chinese Room argument, and on the opposed position that a computer itself could really understand a natural language. In this paper I develop the notion of a virtual person, modelled on the concept of virtual machines familiar in computer science. I show how Searle's argument, and J. Maloney's attempt to defend it, fail. I conclude that Searle is correct in holding that no digital machine could understand language, but wrong in holding that artificial minds are impossible: minds and persons are not the same as the machines, biological or electronic, that realize them.     

The Wrong Stuff: Chinese Rooms and the Nature of Understanding

Searle's Chinese Room argument is a general argument that proves that machines do not have mental states in virtue of their programming. I claim that the argument expresses powerful but mistaken intuitions about understanding and the first person point of view. A distinction is drawn between a competence sense and a performance sense of 'understanding texts'. It is argued that the Chinese Room intuition looks for a special experience (performance) of comprehension, whereas artificial intelligence is attempting to explain the knowledge (competence) required to understand texts. Moreover, a dilemma is sketched for the argument: either Searle hasn't identified the appropriate subject of understanding or he may understand after all. Finally, I question the underlying assumption that the general definition of mental states requires a projectable-by-us first person point of view.     

A New Science of Criticism

John R. Searle has recently observed that something might instantiate a Chinese‐‘understanding’ computer program without having any understanding of Chinese. He thinks that this implies that instantiating such a program is ‘never by itself a sufficient condition of intentionality’. I show that this phrase is incoherent, and that all that follows is that instantiating such a program is not in every case a sufficient condition for the given intentionality. But the conclusion to Searle's argument, thus revised, is neither new nor significant; Searle's arguments merely raise old issues in new clothing.     

Is Objective Consequentialism Compatible with the Principle that “Ought” Implies “Can”?

Some philosophers hold that objective consequentialism is false because it is incompatible with the principle that “ought” implies “can”. Roughly speaking, objective consequentialism is the doctrine that you always ought to do what will in fact have the best consequences. According to the principle that “ought” implies “can”, you have a moral obligation to do something only if you can do that thing. Frances Howard-Snyder has used an innovative thought experiment to argue that sometimes you cannot do what will in fact have the best consequences because you do not know what will in fact have the best consequences. Erik Carlson has raised two objections against Howard-Snyder’s argument. This paper examines Howard-Snyder’s argument as well as Carlson’s objections, arguing that Carlson’s objections do not go through but Howard-Snyder’s argument fails nonetheless. Moreover, this paper attempts to show that objective consequentialism and other objectivist moral theories are compatible with the principle that “ought” implies “can”. Finally, this paper analyses a special kind of inability: ignorance-induced inability.     

如何让电脑真正懂汉语?——一种以许慎的“六书”理论为母型的汉语处理模型

哲学家塞尔曾指出,任何一台处理汉语输入的计算机都不能够像真人那样理解汉语,因为计算机无法获取汉语表达式的语义内容。他的这个论证,既预设了语义学和句法学之间的二分,而且也预设了一种外在论式的语义学观点。上述两个预设在本文中都会受到挑战。首先,本文将采纳一种内在论的(并且是准整体论)的语义学进路:根据此进路,意义完全可以在不直接牵涉到和外部对象的关系的情况下而被带入语义网。此外,本文所试图构建的语义学模型也不是句法驱动的——换言之,在该模型中,并没有任何语义中立的原则可以告知系统,复合语义是如何从"原子"语义中生成的。毋宁说,"原子"语义之间的亲和性,便已然在复合语义的构成中扮演了枢纽性角色。需要指出的是,关于语义融合的语义学难题,绝非当代语言学家和语言哲学家的专利。实际上,它早在东汉学者许慎的"六书"论中就已得到触及。根据许论,只要一个汉字的构件的语义得到了其它汉语表达式的恰当注解,该汉字本身的语义也便可得到确认。在本文中,许慎的"六书"论将以一种可计算的方式而得到系统化的重构,而该重构的技术基础则是王培先生发明的"纳思系统"(非公理化推理系统)。在这种重构基础上,现代汉语中语词的语义复合方式,也可以得到一种新颖的理解。     

Wittgensteinian content‐externalism

Content‐externalism is the view that a subject's relations to a context can play a role in individuating the content of her mental states. According to social content‐externalists, relations to a socio‐linguistic context can play a fundamental individuating role. Åsa Wikforss has suggested that “social externalism depends on the assumption that individuals have an incomplete grasp of their own concepts”. In this paper, I show that this isn't so. I develop and defend an argument for social content‐externalism which does not depend on this assumption. The argument is animated by strands of thought in the later work of Wittgenstein. In addition to demonstrating that social externalists are not necessarily committed to thinking that a subject can have thoughts involving concepts which she incompletely understands, this argument is important insofar as it (a) supports a form of content‐externalism with extended scope, (b) avoids the controversy surrounding the claim that subjects can think with concepts which they incompletely understand, and (c) situates Wittgenstein's later work with respect to contemporary debates about content‐externalism.     

The curious case of the Chinese gym

Searle has recently used two adaptations of his Chinese room argument in an attack on connectionism. I show that these new forms of the argument are fallacious. First I give an exposition of and rebuttal to the original Chinese room argument, and then a brief introduction to the essentials of connectionism.     

Against alief

A physicalist holds, in part, that what properties are instantiated depends on what physical properties are instantiated; a physicalist thinks that mental properties, for example, are instantiated in virtue of the instantiation of physical “realizer” properties. One issue that arises in this context concerns the relationship between the “causal powers” of instances of physical properties and instances of dependent properties, properties that are instantiated in virtue of the instantiation of physical properties. After explaining the significance of this issue, I evaluate two core lines of thought that have been advanced in favor of Subset Inheritance, the view that instances of dependent properties typically have some, but not all, of the powers of physical realizers, and do not have any powers that are not also powers of physical realizers. The first argument that I address turns on our intuitive reactions to certain cases; the second appeals to the phenomenon of multiple realization. I argue that neither line of thought succeeds, and thus that insofar as we grant that an instance of a dependent property inherits some of the powers of its physical realizer, defenders of subset inheritance have not provided a compelling reason to think that it will not inherit all of the powers of its physical realizer.     

Korsgaard on Hypothetical Imperatives

I argue that rationalists need not adopt Kant’s method for determining what one has reason to do, where by “Kant’s method” I mean the view that normative guidance comes only from directives imposed on the agent by the agent’s own will. I focus on Kant’s argument for “imperatives of skill,” one sort of hypothetical imperative. I argue, against Korsgaard, that Kant’s argument is neither better nor significantly different than the sort of argument non-Kantian rationalists offer. I close by arguing that Korsgaard is wrong to think that her question “why should I care about performing the means to my ends?” is a serious worry.     

Experimental mathematics, computers and the a priori

In recent decades, experimental mathematics has emerged as a new branch of mathematics. This new branch is defined less by its subject matter, and more by its use of computer assisted reasoning. Experimental mathematics uses a variety of computer assisted approaches to verify or prove mathematical hypotheses. For example, there is “number crunching” such as searching for very large Mersenne primes, and showing that the Goldbach conjecture holds for all even numbers less than 2 × 1018. There are “verifications” of hypotheses which, while not definitive proofs, provide strong support for those hypotheses, and there are proofs involving an enormous amount of computer hours, which cannot be surveyed by any one mathematician in a lifetime. There have been several attempts to argue that one or another aspect of experimental mathematics shows that mathematics now accepts empirical or inductive methods, and hence shows mathematical apriorism to be false. Assessing this argument is complicated by the fact that there is no agreed definition of what precisely experimental mathematics is. However, I argue that on any plausible account of ’experiment’ these arguments do not succeed.     

Towards a literary account of mental health from James' Principles of Psychology

The field of mental health tends to treat its literary metaphors as literal realities with the concomitant loss of vague “feelings of tendency” in “unusual experiences”. I develop this argument through the prism of William James’ (1890) “The Principles of Psychology”. In the first part of the paper, I reflect upon the relevance of James' “The Psychologist's Fallacy” to a literary account of mental health. In the second part of the paper, I develop the argument that “connotations” and “feelings of tendency” are central to resolving some of the more difficult challenges of this fallacy. I proceed to do this in James' spirit of generating imaginative metaphors to understand experience. Curiously, however, mental health presents a strange paradox in William James’ (1890) Principles of Psychology. He constructs an elaborate conception of the “empirical self” and “stream of thought” but chooses not to use these to understand unusual experiences – largely relying instead on the concept of a “secondary self.” In this article, I attempt to make more use of James' central division between the “stream of thought” and the “empirical self” to understand unusual experiences. I suggest that they can be usefully understood using the loose metaphor of a “binary star” where the “secondary self” can be seen as an “accretion disk” around one of the stars. Understood as literary rather the literal, this metaphor is quite different to more unitary models of self-breakdown in mental health, particularly in its separation of “self” from “the stream of thought” and I suggest it has the potential to start a re-imagination of the academic discourse around mental health.     

The “Positive Argument” for Constructive Empiricism and Inference to the Best Explanation

In this paper, I argue that the “positive argument” for Constructive Empiricism (CE), according to which CE “makes better sense of science, and of scientific activity, than realism does” (van Fraassen in The scientific image, Clarendon Press, Oxford, 1980, 73), is an Inference to the Best Explanation (IBE). But constructive empiricists are critical of IBE, and thus they have to be critical of their own “positive argument” for CE. If my argument is sound, then constructive empiricists are in the awkward position of having to reject their own “positive argument” for CE by their own lights.     

THOUGHT,LANGUAGE, AND THE ARGUMENT FROM EXPLICITNESS

Abstract: This article deals with the relationship between language and thought, focusing on the question of whether language can be a vehicle of thought, as, for example, Peter Carruthers has claimed. We develop and examine a powerful argument—the “argument from explicitness”—against this cognitive role of language. The premises of the argument are just two: (1) the vehicle of thought has to be explicit, and (2) natural languages are not explicit. We explain what these simple premises mean and why we should believe they are true. Finally, we argue that even though the argument from explicitness shows that natural language cannot be a vehicle of thought, there is a cognitive function for language.     

The Owl and the Ostrich: Reply to Sami Pihlström on Ethical Unthinkabilities and Philosophical Seriousness

Sami Pihlström argues in his “Ethical Unthinkabilities and Philosophical Seriousness” that there are some philosophical views that are so dangerous that we should not discuss them. He advances this argument with special reference to my (anti‐natalist) view that being brought into existence is always a serious harm. In response I argue: (a) that there are major flaws in his argument for the conclusion that we should not think about (purportedly) unthinkable views; and (b) that my views about the harm of coming into existence are in any event neither unthinkable nor dangerous.     

Steep Cliff Arguments

In recent philosophical debates a number of arguments have been used which have so much in common that it is useful to study them as having a similar structure. Many arguments – Searle's Chinese Room, for example – make use of thought experiments in which we are told a story or given a narrative context such that we feel we are in comfortable surroundings. A new notion is then introduced which clashes with our ordinary habits and associations. As a result, we do not bother to investigate seriously the new notion any further. I call such an arrangement, which is perhaps a variation of the fallacy of presumption, a Steep Cliff argument. One remedy for the misdirection of a Steep Cliff argument is to tell a counterstory from the point of view of the rejected notion.     

Defending (a modified version of) the Zygote Argument

Think of the last thing someone did to you to seriously harm or offend you. And now imagine, so far as you can, becoming fully aware of the fact that his or her action was the causally inevitable result of a plan set into motion before he or she was ever even born, a plan that had no chance of failing. Should you continue to regard him or her as being morally responsible—blameworthy, in this case—for what he or she did? Many have thought that, intuitively, you should not. Recently, Alfred Mele has employed this line of thought to mount what many have taken to be a powerful argument for incompatibilism: the “Zygote Argument”. However, in interesting new papers, John Martin Fischer and Stephen Kearns have each independently argued that the Zygote Argument fails. As I see it, the criticisms of Fischer and Kearns reveal some important questions about how the argument is meant to be—or how it would best be—understood. Once we make a slight (but important) modification to the argument, however, I think we will be able to see that the criticisms of Fischer and Kearns do not detract from its substantial force.     

Metaphysics and the Future-Like-Ours Argument Against Abortion

Don Marquis’s “future-like-ours” argument against the moral permissibility of abortion is widely considered the strongest anti-abortion argument in the philosophical literature. In this paper, I address the issue of whether the argument relies upon controversial metaphysical premises. It is widely thought that future-like-ours argument indeed relies upon controversial metaphysics, in that it must reject the psychological theory of personal identity. I argue that that thought is mistaken—the future-like-ours argument does not depend upon the rejection of such a theory. I suggest, however, that given a widely-accepted view about contraception and abstinence, the argument is committed to contentious metaphysics after all, as it relies upon a highly controversial assumption about mereology. This commitment is not only relevant for those who are inclined to endorse the argument but reject the mereological view in question, but in addition entails dialectical and epistemological liabilities for the argument, which on some views will be fatal to the argument’s overall success.     

Qualia that It Is Right to Quine*

Dennett (1988) provides a much discussed argument for the nonexistence of qualia, as conceived by philosophers like Block, Chalmers, Loar and Searle. My goal in this paper is to vindicate Dennett's argument, construed in a certain way. The argument supports the claim that qualia are constitutively representational. Against Block and Chalmers, the argument rejects the detachment of phenomenal from information‐processing consciousness; and against Loar and Searle, it defends the claim that qualia are constitu‐tively representational in an externalist understanding of this. The core of the argument is contained in section 3. In the first part, I contrast a minimal conception of qualia, relative to which their existence is not under dispute, with the sort of view to which I will object. In the second part I set the stage by presenting the facts about (minimal) qualia on which a Dennett‐like argument can be based.