"The Termination Risks of Simulation Science" Erkenntnis 1-21. forthcoming.
Historically, the hypothesis that our world is a computer simulation has struck many as just another improbable-but-possible “skeptical hypothesis” about the nature of reality. Recently, however, the simulation hypothesis has received significant attention from philosophers, physicists, and the popular press. This is due to the discovery of an epistemic dependency: If we believe that our civilization will one day run many simulations concerning its ancestry, then we should believe that we are probably in an ancestor simulation right now. This essay examines a troubling but underexplored feature of the ancestor-simulation hypothesis: the termination risk posed by both ancestor-simulation technology and experimental probes into whether our world is an ancestor simulation. This essay evaluates the termination risk by using extrapolations from current computing practices and simulation technology. The conclusions, while provisional, have great implications for debates concerning the fundamental nature of reality and the safety of contemporary physics.
"Success-First Decision Theories" In Arif Ahmed (ed.), Newcomb's Problem, Cambridge University Press, 2018.
The standard formulation of Newcomb's problem compares evidential and causal conceptions of expected utility, with those maximizing evidential expected utility tending to end up far richer. Thus, in a world in which agents face Newcomb problems, the evidential decision theorist might ask the causal decision theorist: "if you're so smart, why ain’cha rich?” Ultimately, however, the expected riches of evidential decision theorists in Newcomb problems do not vindicate their theory, because their success does not generalize. Consider a theory that allows the agents who employ it to end up rich in worlds containing Newcomb problems and continues to outperform in other cases. This type of theory, which I call a “success-first” decision theory, is motivated by the desire to draw a tighter connection between rationality and success, rather than to support any particular account of expected utility. The primary aim of this paper is to provide a comprehensive justification of success-first decision theories as accounts of rational decision. I locate this justification in an experimental approach to decision theory supported by the aims of methodological naturalism.
“Value in Very Long Lives” Journal of Moral Philosophy 14 (4): 416-434. 2017.
As things currently stand, our deaths are unavoidable and our lifespans short. It might be thought that these qualities leave room for improvement. According to a prominent line of argument in philosophy, however, this thought is mistaken. Against the idea that a longer life would be better, it is claimed that negative psychological states, such as boredom, would be unavoidable if our lives were significantly longer. Against the idea that a deathless life would be better, it is claimed that such a life would be lacking in important sources of value, because death is a precondition for many of our valuing attitudes. I argue that these problems are avoided by very long lives that incorporate fading memory, limited ignorance of future events, and temporal scarcity. I conclude that very long lives are, in principle, desirable, and that death does not play an essential role in our valuing attitudes.
“Against Time Bias” Ethics 125 (4): 947-970. 2015. With Meghan Sullivan.
Most of us display a bias toward the near: we prefer pleasurable experiences to be in our near future and painful experiences to be in our distant future. We also display a bias toward the future: we prefer pleasurable experiences to be in our future and painful experiences to be in our past. While philosophers have tended to think that near bias is a rational defect, almost no one finds future bias objectionable. In this essay, we argue that this hybrid position is untenable. We conclude that those who reject near bias should instead endorse complete temporal neutrality.
“When Is A Belief True Because of Luck?” Philosophical Quarterly 63 (252): 465-475. 2013.
Many epistemologists are attracted to the claim that knowledge possession excludes luck. Virtue epistemologists attempt to clarify this idea by holding that knowledge requires apt belief: belief that is true because of an agent's epistemic virtues, and not because of luck. Thinking about aptness may have the potential to make progress on important questions in epistemology, but first we must possess an adequate account of when a belief is true because of luck. Existing treatments of aptness assume a simple and natural view of luck attribution, according to which the success of a performance is attributable to luck if one of the principal causes of the success is a lucky event. I show that this view is false, and should be replaced. This has major implications for virtue-theoretic accounts of knowledge, as well as the role of luck in epistemology more generally.
Humans seem to be near biased about tradeoffs concerning their own pleasures and pains. Philosophers and economists assume that a similar bias applies to tradeoffs concerning the welfare of future generations. This assumption is unmotivated and leads to absurdities. I argue that intergenerational tradeoffs mostly concern factors unrelated to time. Social discounting, in particular, is easy to confuse with time discounting because they are both hyperbolic and usually correlated. Nevertheless, these types of discounting can be separated by thought experiments. I conclude that models focusing on social distance are superior for both normative and descriptive inquiries about intergenerational tradeoffs.
The inference that life does not matter because it will end is an ongoing source of despair for many reflective people, and literature, popular culture, and philosophy demonstrate its widespread influence. Given its power and prominence, alleviating the despair would be both an intellectual achievement and a contribution to welfare. I evaluate existing philosophical models of the despair, which diagnosis it as involving an unreasonable desire for permanence or positive “final outcome.” I conclude that these diagnoses are sometimes unsatisfactory, and especially so for lifelong atheists. I then argue that a temporal argument for nihilism often generates the despair and that this argument applies to all naturalistic evaluations of mattering, meaning, and value. The temporal argument for nihilism, however, involves an inconsistent shift between temporally neutral and temporally biased evaluations of mattering/meaning/value. Resolving the inconsistency, one way or the other, would remove the source of despair.
“Hedonic and Non-Hedonic Bias Toward the Future” (with Andrew Latham, Kristie Miller, and James Norton)
It has widely been assumed, by philosophers, that our first-person preferences regarding pleasurable and painful experiences exhibit a bias toward the future (hedonic future bias), and that our preferences regarding non-hedonic events exhibit no such bias (non-hedonic time neutrality). Further, it has been assumed that our third-person preferences regarding both hedonic and non-hedonic events are time neutral. Some have attempted to use this (presumed) differential pattern of future bias—different across kinds of events, and different across first- vs third-person preferences—to argue for the irrationality of hedonic future bias. This paper experimentally tests these descriptive hypotheses. We found that, as predicted, first-person preferences are hedonically future biased and non-hedonically time neutral. However, contrary to prediction, we found that third-person preferences are both hedonically and non-hedonically future biased. Hence, the presumed pattern of first- and third-person preferences cannot be used to argue for the irrationality of time bias, since no such pattern exists.
“Discounting the Past: Past/Future Symmetries in Temporal Preferences?” (with Alex Holcombe, Andrew Latham, Kristie Miller, and James Norton)
A person is near biased if, all else being equal, they prefer positive events to be near rather than distant and negative events to be distant rather than near. The most prominent form of the bias is future-directed. A future near-biased agent prefers positive events to be in their near future rather than their distant future and negative events to be in their distant future rather than their near future. To date, experimental work has focused on future-directed near bias. In this paper we empirically investigate past-directed near bias. As predicted, we found a population-level symmetry between past-directed and future-directed near bias across both hedonic and non-hedonic events. That is, we see the same pattern of preferences about future events as we see regarding past events. Further, we added to earlier results that show a difference in time bias between preferences for hedonic and non-hedonic events. Startlingly, however, despite the population-level pattern of past/future symmetry, we found no correlation between individual participants’ past- and future-directed preferences. This suggests that the mechanisms that underlie the pattern of preferences regarding future events are different from the mechanisms that underlie the pattern of preferences regarding past events.
“Act Consequentialism without Free Rides” (with Ben Levinstein)
Consequentialist theories determine rightness solely based on real or expected consequences. Although such theories are popular, they often have difficulty with generalizing intuitions, which, in their pre-theoretic form, require concern for the question “What if everybody did that?” When generalizing versions of consequentialism have been attempted, as with rule consequentialism, the results are messy. We claim that the conceptual apparatus currently employed in generalizing consequentialism is not adequate to the task. Just as decision theory is crucial to modern consequentialism for handling uncertainty, so too is it crucial for handling generalizability. Here, we present a relatively new decision theory, functional decision theory, that will allow us to sketch a theory of generalized act consequentialism. We argue that this theory is superior to rule consequentialism both in modeling the actual reasoning of generalizers and in delivering correct results.
"The Real Problem with Prepunishment"
Postpunishment legal systems convict people of crimes that they have committed, while prepunishment legal systems convict people of crimes that they will or would commit. This paper debunks two assumptions in the philosophical literature: 1) that prepunishment exists only in science fiction, and 2) that prepunishment is unproblematic for deterrence theorists. First, I show that the differences between the actual world and hypothetical cases of prepunishment — even those presented in Philip K. Dick’s “Minority Report” — are nothing but smoke and mirrors. In fact, there is no morally relevant difference between what happens in “Minority Report” and what often happens in the real-world punishment of “attempt” offenses. Therefore, if prepunishment is morally wrong, as non-consequentialist moral philosophers have argued, then vast reforms to current legal theory are required. Second, I reveal severe problems with prepunishment that stem from purely consequentialist considerations. Most importantly for deterrence theorists, prepunishment systems have no deterrent power. These conclusions call for changes in our thinking about both the philosophical importance and moral status of prepunishment.
"The (Ir)rationality of Temporal Discounting: A Heuristics and Biases Approach" (with Christopher Suhler)
The discount function is a centerpiece of empirical and theoretical discussions of intertemporal choice. In philosophy, discussions of temporal discounting look not just at whether and to what extent we actually do discount the future, but whether and to what extent doing so is rational. Given its robustness and the ubiquity of discount functions in other species, there is good reason to think that the (human) discount function has a strong evolutionary basis.
In this paper, we use this evolutionary backdrop, along with the heuristics and biases framework from psychology and behavioral economics, to develop a novel perspective on the rationality of temporal discounting. Our analysis has two main steps. First, we posit that time-neutral calculation concerning probability-weighted goods is the standard for rational choice, and that the discount function evolved as a cognitive heuristic that approximates the probability of obtaining a good on the basis of its temporal location. Second, a suite of relatively recent changes in human societies have dramatically altered the relationship between temporal location and probability. As a result of these changes, the discount function now leads to severe and systematic deviations from the rational standard.
“Future Bias and Unequal Tradeoffs” (with Andrew Latham, Kristie Miller, and James Norton)
Philosophers working on time biases assume that people are hedonically biased toward the future. A hedonically future-biased agent prefers pleasurable experiences to be future instead of past and painful experiences to be past instead of future. Philosophers further predict that this bias is strong enough to apply to unequal tradeoffs: people often prefer less pleasurable future experiences to more pleasurable past ones, and more painful past experiences to less painful future ones. Time neutralists claim that such preferences are irrational; people should have equal regard for the quality of their experiences regardless of their temporal location. Supporters of future bias, in contrast, claim that such preferences are rationally permissible or required. A popular time-neutralist argument involves a third prediction, which concerns the scope of future bias for unequal tradeoffs: they predict that this phenomenon is restricted to first-person preferences about hedonic events, and that people are otherwise time neutral. Time neutralists claim that this pattern of responses supports the idea that hedonic first-person future bias is an evolutionary heuristic that sometimes leads to irrational attitudes. Given the perceived normative implications of these predictions, it is crucial that they be experimentally tested. However, there is little direct empirical research on future bias. In this paper, we test whether first-person preferences exhibit a hedonic future bias for unequal tradeoffs and observe whether the same bias is absent in the case of non-hedonic events and third-person preferences. We then connect this empirical work to the normative debate over future bias. If non-hedonic and third-person preferences differ from hedonic first-person preferences, then this would support the claims of time neutralists. If, however, hedonic and third-person preferences tend to match first-person hedonic preferences, then this would support proponents of hedonic future bias.
"Rationality and Success" (2013)