Radical Scepticism & Envatted Brains: what Putnam proves

PHILIPP BARTH, University of Cambridge

In this paper I am considering various interpretations of Putnam’s well-known argument to the effect that we are not brains in a vat. Starting off with an interpretation that remains very close to the original text and which I reject on the grounds of its ineffectiveness in the context of the sceptical debate I go on to discuss what I call the “aggressive” interpretation. The latter comes close to Wright’s famous reconstrual but has to be rejected, too, because its reliance on semantic self-knowledge renders it epistemically circular in the context of the debate. Following this I develop a variation of the “aggressive” interpretation that does away with any such knowledge and remains wholly neutral as regards what language it is that we are in fact using. I argue that such a variation succeeds in “quieting” the sceptic though it falls short of proving her wrong. But, I contend, having “quieted” the sceptic is a significant result.

Belief, Assertion and Moore’s Paradox

TIMOTHY CHAN, University of Oxford

This paper argues that two received accounts of belief and assertion cannot both be correct, because they entail mutually contradictory claims about Moore’s Paradox. The two accounts in question are, first, the Action Theory of Belief (ATB), the functionalist view that belief must be manifested in dispositions to act, and second, the Belief Account of Assertion (BAA), the view that an asserter must present himself as believing what he asserts. It is generally accepted also that Moorean assertions are absurd, and that (BAA) explains why they are. I shall argue that (ATB) implies that some Moorean assertions are, in some fairly ordinary contexts, well justified. Thus (BAA) and (ATB) are mutually inconsistent.

Coherence and Rationality: The Challenge to Davidson’s View of Akrasia

ALEXANDRE ERLER, University of East Anglia

In this paper I look at the question whether it is always rational for an agent to act in keeping with her best judgement,?and at the recent critique of the common view on the matter. I focus on Donald Davidson’s theory of action as representative of the “intellectualist position” regarding rationality, which is the view under attack. I begin by giving an overview of that theory and of the objections that have recently been levelled at it by authors such as Robert Audi, Nomy Arpaly, or Allison McIntyre. I distinguish between an objective and a subjective sense of “rational”, and stress that it is the latter one which is relevant to our discussion. I then endorse the seemingly paradoxical thesis that it can sometimes be more rational to act against our best judgement than to follow it – as illustrated by the behaviour of Mark Twain’s character Huckleberry Finn, when he fails to turn in his friend Jim, a slave, to the men who are looking for him, even though he had resolved to do so. At the same time, I deny that such a thesis can be maintained if we assume Davidson’s particular conception of rationality. I argue, however, that there are good reasons to take a broader view of what it is to be a “rational” agent, and that such a view does indeed legitimate the point made by Davidson’s critics. Finally, I consider whether this point implies, as some of these critics have argued, that the requirements of continence can sometimes be defeated by other considerations, so that it might even be more rational for an agent to choose to act akratically in certain circumstances. I argue that such a claim rests on a misunderstanding of what Davidson called the “principle of continence”, which he regarded as an inescapable rule for practical reasoning; and that given the kind of agents we are, the policy that this claim entails does not really appear to be a “rational” one.

Knowledge of Contextualism in Context

MICHAEL FENTON, University of Warwick

Contextualist theories in epistemology suggest a novel response to the puzzle which arises from the Argument from Ignorance. ?1 of this paper presents the sceptical puzzle as an inconsistent triad and distinguishes contextualist responses from those of Moore and the sceptic. Two versions of contextualism are described, and it is claimed that only one of these provides an adequate response to the sceptical puzzle. ?2 discusses an argument against contextualism presented by Timothy Williamson, and more recently, in a slightly different form, by Crispin Wright. The conclusions of both arguments are unacceptable for the contextualist – the first concludes that contextualism entails commitment to asserting a sentence which is Moore-paradoxical; the second, that contextualism entails commitment to asserting a contradiction. I will suggest a way for a contextualist to respond to these arguments, and then outline some implications of this response.

What is Practical Reasoning?

JULIAN FINK, University of Oxford

This paper argues that practical reasoning is a mental process which leads a person from a set of existent mental states to an intention. In Section 1, I defend this view against two other proposals according to which practical reasoning either concludes in an action itself or in a normative belief. Section 2 discusses the correctness of practical reasoning and explains how the correctness of instrumental reasoning can be explained by the logical relations that hold between the contents of the mental states. In Section 3, I explore the correctness of normative practical reasoning. I conclude with the sceptical view that correct practical reasoning cannot require us to intend to do what we believe we ought to do.

Killing John to Save Mary: A Defence of the Moral Distinction between Killing and Letting Die

HELEN FROWE, University of Reading

In this paper I outline the attractions of rejecting a morally significant distinction between killing and letting die. In particular, I explain why this position is attractive to those who argue for the permissibility of active euthanasia. By equating killing with letting die, arguments which support the less controversial practice of passive euthanasia will also support active euthanasia. It is argued in this paper that attempts to deny the distinction between killing and letting die fail. However, I give a brief indication of how one might reconcile a belief in the killing / letting die distinction with a belief in the permissibility of active euthanasia.

Firstly, I consider Michael Tooley’s claim that the initiation of a causal process is morally equivalent to refraining from interfering in that process. Tooley denies that he is committed to the additional, implausible claim that refraining from preventing somebody else from initiating a causal process is equivalent to initiating that process oneself. I argue that Tooley is indeed implicitly committed to this claim. From this, it would follow that I am just as responsible for preventing you from initiating causal processes as I am for refraining from initiating those processes myself. So, the duty not to kill will include just as stringent a duty to prevent other people from killing. This is not an attractive position, and nor is it a position which Tooley wishes to endorse.

I then examine Tooley’s claim that while external factors can mean that killing appears worse than letting die, the killing / letting die distinction is itself morally neutral. Tooley hypothesises a machine in which Mary will die unless one presses a button to kill John. He claims that it makes no moral difference whether one pushes the button or not, since one child will die either way. I show that Tooley’s view sanctions the killing of innocent bystanders as a means to saving one’s own life. Such killings are, I suggest, paradigm cases of the morally impermissible.

I then discuss James Rachels’s attempts to show that in fact our intuitions do not support a moral distinction between killing and letting die. I concede that in the examples Rachels offers it is very difficult to draw an intuitive line between cases of killing and letting die. However, I argue that this does not mean that there is no line to be drawn. Rachels’s famous Drowning Child example relies upon a representation of events which in itself blurs the distinction between killing and letting die, and more generally, between action and inaction. By eliminating these distorting factors, one can re-establish the priority of the duty not to kill over the duty to save.

Conceptual ‘How-Possible’ Questions

ANIL GOMES, Oxford University

Quassim Cassam has drawn our attention to the structure of ‘how-possible’ questions: questions which ask how an event or achievement was possible. Cassam’s focus is on epistemological how-possible questions: how is knowledge of a certain sort possible? In this paper I suggest that this structure can be usefully applied to what I call conceptual ‘how-possible’ questions: questions about how a certain sort of thought is possible. I do this with reference to two debates – one in the philosophy of perception, one in the philosophy of other minds – both of which, I suggest, can be usefully viewed as responses to specific conceptual ‘how-possible’ questions. More generally, there is a style of response to conceptual ‘how-possible’ questions which focuses on the role of experience in providing us with our concepts – the perception and other mind debates can be viewed in this light.

The (topo)logic of vagueness

BRIAN HILL, IHPST-CNRS, University Paris 1

Zeno’s ‘dichtomy’ paradox argues that a runner in a race will never finish: before reaching the finish, he must get to the half-way point p1; but once he has got to p1, he must still reach the point half-way between p1 and the finish, say p2; but once he has reached p2 . . . So, one might conclude, he never really finishes the race. The sorites paradox argues that a man with a full head of hair may count as bald: for a man with 0 hairs counts as bald; and if a man with 0 hairs counts as bald, a man with 1 hair counts as bald; and if a man with 1 hair counts as bald, a man with 2 hairs counts as bald; and . . . ; and if a man with 4999 hair counts as bald, a man with 5000 hairs counts as bald, so a man with 5000 counts as bald.

There is a certain structural resemblance between these arguments. Each involves a long series of apparently legitimate steps, which, taken together, yield an unacceptable conclusion. Despite this similarity, philosophers have tended to treat the two paradoxes separately and in different ways. One commonly accepts Zeno’s infinite sequence as ‘compatible’ with the race finishing in a finite number of seconds, whereas one often reacts to the sorites paradox by proposing a semantic, epistemic or pragmatic theory of vagueness which denies the sorites argument. In this paper, we shall employ the apparent structural similarity between the two paradoxes to ‘transpose’ aspects of a common reply to Zeno’s paradox ‘onto’ the sorites case. More precisely, it shall be noted that ‘compatibility’ is not enough to defuse Zeno’s paradox, for it is also necessary to assert that the number of seconds (not the number of ‘stages’ of Zeno’s sequence) is the appropriate measure for whether the runner has finished the race or not. If one tries to apply this aspect of the reply to Zeno’s paradox in the sorites case, one ends up rejecting the argument on the grounds that the number of hairs is not appropriate for reasoning about baldness. This will have some interesting consequences for the conception of the problem which vagueness poses, and thus for what one should expect from a theory of vagueness. Questions regarding general relationships (such as compatibility and appropriateness) between terms (be they vague or precise) shall take precedent over the question of the truth of sentences featuring vague terms. As a final illustration of the different perspective offered by the analogy with Zeno’s paradox, Williamson’s anti-luminosity argument (2000) shall be discussed and found wanting.

Why Reasons Need Not Be Conceptual

MARIE LUNDSTEDT, Umea University, Sweden

In this talk I will be concerned with the issue of conceptual versus nonconceptual content of reasons. More specifically I will be focusing on whether perceptual experiences need to be conceptual in order to qualify as justifiers for beliefs. Two things will be brought out in this talk: first, I shall argue that conceptual content – in the sense of concepts being the constituents of experiential content – is a necessary condition for justification only if we are committed to a particular account of propositions and their constituents. I claim that an account of justification does not require that we decide in favour of any particular account of propositions. Secondly, I shall argue that conceptual content – in the sense of subjects having to possess the relevant concepts – is a necessary condition for justification only on the assumption that a subject needs to be capable of recognizing and stating her reasons as reasons. This requirement is unreasonable and might be a result of a level confusion regarding epistemic justification.

The Geometry of Visual Experience and Direct Realism

PHILLIP JOHN MEADOWS, University of Durham

In his article ‘Thomas Reid’s Geometry of Visibles’, James Van Cleve provides a reconstruction of Reid’s argument for a spherical geometry in the case of, at least, monocular visual experience. In my paper, I show that Van Cleve’s argument for a spherical geometry does not work for the monocular case. This is shown to be because a crucial assumption in the argument, which concerns the visible angles formed by tangents of great circles, is not adequately supported. Van Cleve has also highlighted, correctly, that his and Reid’s conclusion about the geometry of visual experience is in conflict with Direct Realism. In this paper I also show how the tension between Direct Realism and a spherical geometry for visual experience occurs at a very fundamental level of the argument, as Van Cleve conceives of it. I also provide reasons for thinking that even if the considerations Van Cleve offers were convincing in the monocular case, they do not even apply in the binocular case. A consequence of this is that the question of the truth of Direct Realism can remain open, even if Van Cleve’s considerations were convincing.

Against Williamson's Anti-Luminosity Argument

CHARLES PELLING, University of Reading/North Carolina

Following Timothy Williamson, let us say that a condition is luminous if and only if one is in a position to know that it obtains if in fact it does obtain. In his Knowledge and Its Limits, Williamson argues that there is no remotely substantial central core of mental states, S1, S2, S3, etc. which is such that the conditions of being in S1, S2, S3, etc. are luminous conditions. In this paper, I show that Williamson’s argument is inconclusive.

Williamson argues that the condition of feeling cold is not luminous. To do this, he appeals to a thought-experiment in which one moves gradually from feeling cold to not feeling cold. I argue that since Williamson intends his argument against the luminosity of feeling cold to generalize to conditions such as that of something's appearing to one to be a particular shade of colour, he faces a dilemma. Either his argument will have to appeal to a directly parallel type of thought-experiment, in which case the argument is vulnerable to a pair of objections that have to do with temporal considerations; alternatively, his argument could appeal to a slightly different type of thought-experiment, in which case – though the objections from temporal considerations then disappear – it is no longer clear that the imagined experiment is one that might actually occur.

Perception and Context

IAN PHILLIPS, UCL & All Souls College, Oxford

How we describe a stretch of someone’s mental life inherently depends on our folk-psychological interests. Two speakers in different contexts, with different interests, may describe a subject’s mental life using exactly the same form of words, with all those words having their same, ordinary meanings and yet only one of them speak truly. I contend that such context-sensitivity is a deep fact about our perceptual vocabulary. I focus on the richness of our characterizations of perceptual experience, specifically the visibility or otherwise of natural kind properties – my argument can be easily adapted to many other types of property (e.g., artefactual and certain relational properties).

I begin by describing a way in which philosophers have traditionally sought to draw a distinction amongst the properties of the objects of perception. With any such distinction in play it makes sense to ask questions such as, Do natural kind properties, strictly speaking, characterize the phenomenology of experience or do they enter merely in interpreting our experience? Or, Are natural kind properties visible? The answer to these questions depends crucially on when it is considered appropriate to postulate visual illusion. A related distinction due to Peacocke divides concepts into the observational and the non-observational. This distinction also rests on a certain conception of when it is appropriate to posit what Peacocke speaks of as (purely) perceptual error.

A tension arises whenever one asks of certain ordinary experiences whether a given kind property is visible or whether a given concept is observational. On the one hand, the way we naïvely describe our experience commits us to the visibility of natural kind properties. On the other hand, such a view commits one to an implausible account of illusion or perceptual error in certain matching cases where the natural kind is not present.

Having brought out the tension, I claim that, short of an error theory, the only way to overcome it is to treat it as a symptom of a false assumption, namely that how rich an ascription of perceptual content it is correct to make is independent of the folk theoretical concerns of one’s context. With this assumption removed, the tension can be seen not as a reflection of an inconsistency in our ordinary ways of thinking about experience but rather as a failure to appreciate the context-sensitivity of any sensory/cognitive or observational/non-observational distinction. I end by showing how this contention can be seen as a local application of Travis’ radical contextualism, the claim that context-sensitivity is a ubiquitous and ineliminable truth about language.

Two Platitudes about Interpersonal Comparisons

MAURO ROSSI, LSE

One problem concerning interpersonal utility comparisons is whether or not they are empirically meaningful. The debate is often introduced by reference to daily life examples. This strategy raises three questions. First, what are the intuitions conveyed by these examples? Second, how do they contrast with theoretical reflection? Third, to what extent do daily life situations fit the theoretical picture of decision-making?

I distinguish two platitudes about interpersonal comparisons. The first is that, decision-making requires making interpersonal comparisons, when a choice affects other people; the second is that we need interpersonal comparisons to explain people’s behaviour. These intuitions support an argument according to which interpersonal utility comparisons are empirically meaningful. The literature explores two strategies to defend this conclusion. They both fail. Therefore, interpersonal utility comparisons are empirically meaningless. However, we can maintain our intuitions. Davidson claims that the attribution of mental states to an agent requires a projection of the interpreter’s standards of rationality and values. In turn, this projection implies mental states comparability in some relevant respects. I endorse Davidson’s argument and argue that comparability of preferences is necessary for the explanation of individual behaviour to be possible, although not with respect to the dimension of preference strength. On the other hand, actual decision-making can be explained without assuming the empirical meaningfulness of interpersonal utility comparisons. I explore four different ways in which interpersonal comparisons help adjudicate conflicting claims. Ultimately, they recommend a more careful reading of the relationship between actual decision-making and standard theoretical accounts.