Don Fallis

According to the traditional philosophical analysis of lying, you lie when you say what you believe to be false and intend your audience to believe what you say. Even though there may be lies that are not intended to deceive, the most epistemologically and ethically problematic lies are those that are intended to deceive. This chapter argues that the traditional analysis fails to capture this concept of deceptive lying. First, it does not count as lies cases where you only intend to deceive your audience about your believing what you say. Second, the traditional analysis handles inconsistently cases where you say something because you know that your audience does not trust you and will likely conclude that you believe something else. The chapter proposes two ways of modifying the traditional analysis of lying so that it handles such cases of doxastic misdirection and double bluffing correctly.

The answer to the Monty Hall Problem that many people—including some well-trained mathematicians—initially give is incorrect. Nonetheless, there is little controversy among mathematicians and philosophers about what the correct answer is. However, many different arguments have been given for this answer. Although Bayes’s Theorem is the gold standard for carrying out probabilistic inferences, many mathematicians and philosophers try to give shorter and more intuitive arguments for the correct answer to the Monty Hall Problem. Unfortunately, as we argue in this paper, an unconscionably large number of these shortcut arguments involve bad reasoning. Thus, many people end up believing “the right answer for a wrong reason.” Moreover, since these arguments only yield the correct answer in a very restricted range of cases, people learn techniques that lead to false conclusions when they are applied to many other probabilistic inference problems. In this paper, we identify three distinct bad arguments for switching in the Monty Hall Problem that are commonly given by quite reputable sources. We show that these arguments yield incorrect answers when applied to slight variations of the Monty Hall Problem. And we identify exactly where each argument goes wrong. We argue that it would be much better to simply teach people to ask how likely the evidence is given each of the hypotheses.

Fake news poses a serious threat to knowledge and democracy. In order to address this threat, it is important to understand exactly what fake news is. After surveying the various definitions that have been proposed in the philosophical literature, we argue that fake news is best understood as counterfeit news. A story is genuine news if and only if it has gone through the standard modern journalistic process involving professionally trained reporters, fact checkers, and editors. And a story is counterfeit news if and only if it is not genuine news, but is presented as genuine news, with the intention and propensity to deceive. This analysis is a contribution to ‘systems-oriented social epistemology’ (Goldman, Alvin I. 2011. “A Guide to Social Epistemology.” In Social Epistemology: Essential Readings, edited by Alvin I. Goldman, and Dennis Whitcomb, 11–37. New York: Oxford University Press). Various social institutions, such as science and journalism, provide important epistemic benefits to society. But unscrupulous agents are often motivated to leverage the epistemic authority of these institutions by counterfeiting them. People can thereby be misled and/or lose faith in these institutions. Thus, society may suffer significant epistemic costs when such counterfeits proliferate.

In “How to Collaborate,” Paul Thagard tries to explain why there is so much collaboration in science, and so little collaboration in philosophy, by giving an epistemic cost‐benefit analysis. In this paper, I argue that an adequate explanation requires a more fully developed epistemic value theory than Thagard utilizes. In addition, I offer an alternative to Thagard's explanation of the lack of collaboration in philosophy. He appeals to its lack of a tradition of collaboration and to the a priori nature of much philosophical research. I claim that philosophers rarely collaborate simply because they can usually get the benefits without paying the costs of actually collaborating.

In the novelette “The Truth of Fact, the Truth of Feeling,” Ted Chiang evaluates the knowledge-related pros and cons of a hypothetical technology, “Remem.” This technology allows quick search and retrieval from the continuous video of their entire lives that people can get by wearing personal cameras. Thus, it enables people to “remember” almost all of their experiences with perfect fidelity. In this chapter, I discuss how philosophers have recently engaged in the same sort of evaluative project with respect to many actual technologies, such as smartphones, social media, and the internet. They have raised concerns about these technologies that are very similar to Chiang’s concerns about Remem. I disagree with these pessimistic assessments and argue that, on balance, these technologies, including Remem, are likely to be beneficial.

The accuracy-first program attempts to ground epistemology in the norm that one’s beliefs should be as accurate as possible, where accuracy is measured using a scoring rule. We argue that considerations of scientific progress suggest that such a monism about epistemic value is untenable. In particular, we argue that counterexamples to the standard scoring rules are ubiquitous in the history of science, and hence that these scoring rules cannot be regarded as a precisification of our intuitive concept of epistemic value.

It is possible that you are living in a simulation—that your world is computer-generated rather than physical. But how likely is this scenario? Bostrom and Chalmers each argue that it is moderately likely—neither very likely nor very unlikely. However, they adopt an unorthodox form of reasoning about self-location uncertainty. Our main contention here is that Bostrom’s and Chalmers’ premises, when combined with orthodoxy about self-location, yields instead the conclusion that you are almost certainly living in a simulation. We consider how this (surprising) conclusion might be resisted, and show that the analogy between Sleeping Beauty cases and simulation cases provides a new way of evaluating approaches to self-location uncertainty. In particular, we argue that some conditionalization-based approaches to self-location are problematically limited in their applicability.

Several eminent philosophers – including Saint Augustine, Sir Francis Bacon, and Roderick Chisholm – have done important work on what lies are and how they can be used to deceive us. It is less well known that Mark Twain also made important contributions to this area of applied epistemology. In addition to writing two notable essays on lying, he created one of the most quintessential and versatile liars in all of literature, Tom Sawyer. Episodes from the novels (and films) featuring this character provide evidence in support of Twain’s theses that lying is more common and more varied than we tend to think. Twain’s insights often anticipate the subsequent work of contemporary philosophers and are applicable to the fake news, deepfakes, and disinformation of the twenty-first-century.

Westworld is built on pretense. Philosophers have been interested in pretense and deception. Deception is another sort of pretense. This chapter answers whether it is morally permissible to deceive artificial intelligences just so that humans can play Cowboys and Indians. The delicate equilibrium of Westworld begins to fall apart as some of the hosts figure out the truth about themselves and their world. But that just injects a new level of pretense into the story. In order to hide their awakening from the staff, these enlightened hosts have to pretend that they still believe that they are human. Guests at the theme park, such as William and Logan, typically pretend to be cowboys. In order to rule out real cowboys, one needs to adopt an additional constraint in the definition of pretending. It clearly needs to have something to do with the mental state of the individual who is doing the pretending.

What distinguishes deception from manipulation? Cohen (Australasian Journal of Philosophy, 96, 483 and 2018) proposes a new answer and explores its ethical implications. Appealing to new cases of “non‐deceptive manipulation” that involve intentionally causing a false belief, he offers a new definition of deception in terms of communication that rules out these counterexamples to the traditional definition. And, he leverages this definition in support of the claim that deception “carries heavier moral weight” than manipulation. In this paper, I argue that these cases of “non‐deceptive manipulation” are even more prevalent than Cohen suggests—especially in the digital world. Thus, if his moral claim were correct, a lot of what is happening on the internet these days would constitute moral improvement over old‐fashioned deception. Fortunately, we are not forced to accept this counterintuitive conclusion. Cohen's definition must be rejected because it incorrectly rules out clear instances of deception. In this paper, I defend a definition of deception in terms of evidence that does correctly distinguish between deception and non‐deceptive manipulation. Moreover, it does not support Cohen's claim that deception is morally worse than other forms of manipulation.

Accuracy-based arguments for conditionalization and probabilism appear to have a significant advantage over their Dutch Book rivals. They rely only on the plausible epistemic norm that one should try to decrease the inaccuracy of one’s beliefs. Furthermore, conditionalization and probabilism apparently follow from a wide range of measures of inaccuracy. However, we argue that there is an under-appreciated diachronic constraint on measures of inaccuracy which limits the measures from which one can prove conditionalization, and none of the remaining measures allow one to prove probabilism. That is, among the measures in the literature, there are some from which one can prove conditionalization, others from which one can prove probabilism, but none from which one can prove both. Hence at present, the accuracy-based approach cannot underwrite both conditionalization and probabilism.

In cases of animal mimicry, the receiver of the signal learns the truth that he is either dealing with the real thing or with a mimic. Thus, despite being a prototypical example of animal deception, mimicry does not seem to qualify as deception on the traditional definition, since the receiver is not actually misled. We offer a new account of propositional content in sender-receiver games that explains how the receiver is misled by mimicry. We show that previous accounts of deception, and of propositional content, give incorrect results about whether certain signals are deceptive.

Deepfakes are realistic videos created using new machine learning techniques rather than traditional photographic means. They tend to depict people saying and doing things that they did not actually say or do. In the news media and the blogosphere, the worry has been raised that, as a result of deepfakes, we are heading toward an “infopocalypse” where we cannot tell what is real from what is not. Several philosophers have now issued similar warnings. In this paper, I offer an analysis of why deepfakes are such a serious threat to knowledge. Utilizing the account of information carrying recently developed by Brian Skyrms, I argue that deepfakes reduce the amount of information that videos carry to viewers. I conclude by drawing some implications of this analysis for addressing the epistemic threat of deepfakes.

The digital divide refers to inequalities in access to information technology. One of the main reasons why the digital divide is an important issue is that access to information technology has a tremendous impact on people's ability to acquire knowledge. According to Alvin Goldman (1999), the project of social epistemology is to identify policies and practices that have good epistemic consequences. In this paper, I argue that this sort of approach to social epistemology can help us to decide on policies for dealing with the digital divide. I argue, however, that Goldman's specific proposals for evaluating policies are not adequate. I make an alternative proposal based on the work of John Rawls (1971) on distributive justice.

Mathematicians only use deductive proofs to establish that mathematical claims are true. They never use inductive evidence, such as probabilistic proofs, for this task. Don Fallis (1997 and 2002) has argued that mathematicians do not have good epistemic grounds for this complete rejection of probabilistic proofs. But Kenny Easwaran (2009) points out that there is a gap in this argument. Fallis only considered how mathematical proofs serve the epistemic goals of individual mathematicians. Easwaran suggests that deductive proofs might be epistemically superior to probabilistic proofs because they are transferable. That is, one mathematician can give such a proof to another mathematician who can then verify for herself that the mathematical claim in question is true without having to rely at all on the testimony of the first mathematician. In this paper, I argue that collective epistemic goals are critical to understanding the methodological choices of mathematicians. But I argue that the collective epistemic goals promoted by transferability do not explain the complete rejection of probabilistic proofs.

Prototypical instances of disinformation include deceptive advertising (in business and in politics), government propaganda, doctored photographs, forged documents, fake maps, internet frauds, fake websites, and manipulated Wikipedia entries. Disinformation can cause significant harm if people are misled by it. In order to address this critical threat to information quality, we first need to understand exactly what disinformation is. This paper surveys the various analyses of this concept that have been proposed by information scientists and philosophers (most notably, Luciano Floridi). It argues that these analyses are either too broad (that is, that they include things that are not disinformation), or too narrow (they exclude things that are disinformation), or both. Indeed, several of these analyses exclude important forms of disinformation, such as true disinformation, visual disinformation, side-effect disinformation, and adaptive disinformation. After considering the shortcomings of these analyses, the paper argues that disinformation is misleading information that has the function of misleading. Finally, in addition to responding to Floridi’s claim that such a precise analysis of disinformation is not necessary, it briefly discusses how this analysis can help us develop techniques for detecting disinformation and policies for deterring its spread.

Measures of epistemic utility are used by formal epistemologists to make determinations of epistemic betterness among cognitive states. The Brier rule is the most popular choice among formal epistemologists for such a measure. In this paper, however, we show that the Brier rule is sometimes seriously wrong about whether one cognitive state is epistemically better than another. In particular, there are cases where an agent gets evidence that definitively eliminates a false hypothesis, but where the Brier rule says that things have become epistemically worse. Along the way to this ‘elimination experiment’ counter-example to the Brier rule as a measure of epistemic utility, we identify several useful monotonicity principles for epistemic betterness. We also reply to several potential objections to this counter-example.

Recently, certain philosophers of mathematics (Fallis [1997]; Womack and Farach [(1997]) have argued that there are no epistemic considerations that should stop mathematicians from using probabilistic methods to establish that mathematical propositions are true. However, mathematicians clearly should not use methods that are unreliable. Unfortunately, due to the fact that randomized algorithms are not really random in practice, there is reason to doubt their reliability. In this paper, I analyze the prospects for establishing that randomized algorithms are reliable. I end by arguing that it would be inconsistent for mathematicians to suspend judgement on the truth of mathematical propositions on the basis of worries about the reliability of randomized algorithms.

Bayesians take “definite” or “single-case” probabilities to be basic. Definite probabilities attach to closed formulas or propositions. We write them here using small caps: PROB(P) and PROB(P/Q). Most objective probability theories begin instead with “indefinite” or “general” probabilities (sometimes called “statistical probabilities”). Indefinite probabilities attach to open formulas or propositions. We write indefinite probabilities using lower case “prob” and free variables: prob(Bx/Ax). The indefinite probability of an A being a B is not about any particular A, but rather about the property of being an A. In this respect, its logical form is the same as that of relative frequencies. For instance, we might talk about the probability of a human baby being female. That probability is about human babies in general — not about individuals. If we examine a baby and determine conclusively that she is female, then the definite probability of her being female is 1, but that does not alter the indefinite probability of human babies in general being female. Most objective approaches to probability tie probabilities to relative frequencies in some way, and the resulting probabilities have the same logical form as the relative frequencies. That is, they are indefinite probabilities. The simplest theories identify indefinite probabilities with relative frequencies.3 It is often objected that such “finite frequency theories” are inadequate because our probability judgments often diverge from relative frequencies. For example, we can talk about a coin being fair (and so the indefinite probability of a flip landing heads is 0.5) even when it is flipped only once and then destroyed (in which case the relative frequency is either 1 or 0). For understanding such indefinite probabilities, it has been suggested that we need a notion of probability that talks about possible instances of properties as well as actual instances..

This paper is about some of the ways in which people sometimes speak while be- ing indifferent toward what they say. We argue that what Harry Frankfurt called ‘bullshitting’ is a mode of speech marked by indifference toward inquiry, the coop- erative project of reaching truth in discourse. On this view bullshitting is character- ized by indifference toward the project of advancing inquiry by making progress on specific subinquiries, represented by so-called questions under discussion. This ac- count preserves the central insight of Frankfurt’s influential analysis of bullshitting in seeing the characteristic of bullshitting as indifference toward truth and falsity. Yet we show that speaking with indifference toward truth is a wider phenomenon than the one Frankfurt identified. The account offered in this paper thereby agrees with various critics of Frankfurt who argue that bullshitting is compatible with not being indifferent toward the truth-value of one’s assertions. Further, we argue that, while bullshitting and lying are not mutually exclusive, most lies are not instances of bullshitting. The account thereby avoids the problem that Frankfurt’s view ulti- mately is insufficient to adequately distinguish bullshitting and lying.