Elliott Sober

Nelson Goodman suggests that a generalization of the form “all A’s are B” is confirmable by an observed instance only if the generalization is law-like. Jackson and Pargetter deny this and give examples of how accidental generalizations can be confirmed. A possible response from Goodman appears to make these accidental generalizations look law-like, but I show it’s defective. And Jackson and Pargetter's substitute nomological condition fares no better than Goodman’s. Because of the multiplicity of possible background assumptions, I doubt that there is a single “nomological condition” that each and every act of confirmation must obey.

Maximum Parsimony (MP) and Maximum Likelihood (ML) are two methods for evaluating which phlogenetic tree is best supported by data on the characteristics of leaf objects (which may be species, populations, or individual organisms). MP has been criticized for assuming that evolution proceeds parsimoniously -- that if a lineage begins in state i and ends in state j, the way it got from i to j is by the smallest number of changes. MP has been criticized for needing to assume some model or other of the evolutionary process even though biologists often do not know what that process really is. This paper critically evaluates both criticisms.

To evaluate Hume's thesis that causal claims are always empirical, I consider three kinds of causal statement: ?e1 caused e2 ?, ?e1 promoted e2 ?, and ?e1 would promote e2 ?. Restricting my attention to cases in which ?e1 occurred? and ?e2 occurred? are both empirical, I argue that Hume was right about the first two, but wrong about the third. Standard causal models of natural selection that have this third form are a priori mathematical truths. Some are obvious, others less so. Empirical work on natural selection takes the form of defending causal claims of the first two types. I provide biological examples that illustrate differences among these three kinds of causal claim

Michael Scriven’s (1959) example of identical twins (who are said to be equal in fitness but unequal in their reproductive success) has been used by many philosophers of biology to discuss how fitness should be defined, how selection should be distinguished from drift, and how the environment in which a selection process occurs should be conceptualized. Here it is argued that evolutionary theory has no commitment, one way or the other, as to whether the twins are equally fit. This is because the theory of natural selection is fundamentally about the fitnesses of traits, not the fitnesses of token individuals. A plausible philosophical thesis about supervenience entails that the twins are equally fit if they live in identical environments, but evolutionary biology is not committed to the thesis that the twins live in identical environments. Evolutionary theory is right to focus on traits, rather than on token individuals, because the fitnesses of token organisms (as opposed to their actual survivorship and degree of reproductive success) are almost always unknowable. This point has ramifications for the question of how Darwin’s theory of evolution and R. A. Fisher’s are conceptually different.

John Beatty (1995) and Alexander Rosenberg (1994) have argued against the claim that there are laws in biology. Beatty's main reason is that evolution is a process full of contingency, but he also takes the existence of relative significance controversies in biology and the popularity of pluralistic approaches to a variety of evolutionary questions to be evidence for biology's lawlessness. Rosenberg's main argument appeals to the idea that biological properties supervene on large numbers of physical properties, but he also develops case studies of biological controversies to defend his thesis that biology is best understood as an instrumental discipline. The present paper assesses their arguments.

I want to explore what happens to two philosophical issues when we assume that the mind, a functional device, is to be understood by the same sort of functional analysis that guides biological investigation of other organismic systems and characteristics. The first problem area concerns the concept of rationality, its connection with reliability and reproductive success, and the status of rationality hypotheses in attribution of beliefs. It has been argued that ascribing beliefs to someone requires the assumption that that person is rational. It also has been argued that evolutionary theory can account for the capacity we have for rational thought by appealing to the idea that rational methods for constructing beliefs are reliable and therefore represented a selected advantage. Both of these claims about rationality are critically examined. In the second part of the paper, I turn to some criticisms that have been advanced against "functionalist" solutions to the mind/body problem. These criticisms are defused when that doctrine is understood in terms of a Teleological (rather than a Turing Machine) notion of function. However, functionalism does not emerge totally unscathed. Voltaire's Dr. Pangloss saw everything in terms of its function and its perfect performance of its function. In the philosophy of mind, his influence lives on in the form of two ideas: (i) that we must possess rational methods for constructing beliefs, since our minds are the product of natural selection, and (ii) that psychological states and properties are individuated by their psychological functions. Contemporary biology provides the resources for avoiding Pangloss's wideeyed teleology in the study of adaptation. It is to be hoped that progress in psychology will loosen the grip of Panglossian assumptions in the study of mind.

Several evolutionary biologists have used a parsimony argument to argue that the single gene is the unit of selection. Since all evolution by natural selection can be represented in terms of selection coefficients attaching to single genes, it is, they say, "more parsimonious" to think that all selection is selection for or against single genes. We examine the limitations of this genic point of view, and then relate our criticisms to a broader view of the role of causal concepts and the dangers of reification in science

This paper proposes a game-theoretic solution of the surprise examination problem. It is argued that the game of “matching pennies” provides a useful model for the interaction of a teacher who wants her exam to be surprising and students who want to avoid being surprised. A distinction is drawn between prudential and evidential versions of the problem. In both, the teacher should not assign a probability of zero to giving the exam on the last day. This representation of the problem provides a diagnosis of where the backwards induction argument, which “proves” that no surprise exam is possible, is mistaken.

Realists persuaded by indispensability arguments af- firm the existence of numbers, genes, and quarks. Van Fraassen's empiricism remains agnostic with respect to all three. The point of agreement is that the posits of mathematics and the posits of biology and physics stand orfall together. The mathematical Platonist can take heart from this consensus; even if the existence of num- bers is still problematic, it seems no more problematic than the existence of genes or quarks. If the two positions just described were the only ones possible, there could be no objection to this melding of numbers with genes and quarks. However, the position I call contrastive empiricism (Sober 1990a) stands opposed to both realism and to Van Fraassen's em- piricism. As it turns out, contrastive empiricism entails that coalesc- ing mathematics with empirical science is highly problematic. I believe that there is an important kernel of truth in abductive ar- guments for genes and quarks. But no counterpart argument exists for the case of numbers. Of course, the existence of this third way would be uninteresting, if contrastive empiricism were wholly implausible. However, I will argue that contrastive empiricism captures what makes sense in standard versions of realism and empiricism, while avoiding the excesses of each. This third way is a middle way; it cannot be dismissed out of hand.

Parsimony arguments are advanced in both science and philosophy. How are they related? This question is a test case for Naturalismp, which is the thesis that philosophical theories and scientific theories should be evaluated by the same criteria. In this paper, I describe the justifications that attach to two types of parsimony argument in science. In the first, parsimony is a surrogate for likelihood. In the second, parsimony is relevant to estimating how accurately a model will predict new data when fitted to old. I then consider how these two justifications apply to parsimony arguments in philosophy concerning theism and atheism, the mind/body problem, ethical realism, the question of whether mental properties are causally efficacious, and nominalism versus Platonism about numbers

Is all explanations causal explanation? Puzzles about barometer readings "explain" storms and shadow lengths "explaining" flagpole heights make it attractive to think so. Wesley Salmon (1984) has endorsed this causal thesis. One way to test this thesis is to assess the explanatory import of pseudo-processes. I do so by discussing the concept of heritability, which measures a pseudo-process, and one role it played in the theory of natural selection: explaining response to selection. This will show, not just that heritability has heuristic or predictive utility, but that it can be explanatory. Possible responses to this case are also discussed.

A phylogeny that allows for lateral gene transfer (LGT) can be thought of as a strictly branching tree (all of whose branches are vertical) to which lateral branches have been added. Given that the goal of phylogenetics is to depict evolutionary history, we should look for the best supported phylogenetic network and not restrict ourselves to considering trees. However, the obvious extensions of popular tree-based methods such as maximum parsimony and maximum likelihood face a serious problem—if we judge networks by fit to data alone, networks that have lateral branches will always fit the data at least as well as any network that restricts itself to vertical branches. This is analogous to the well-studied problem of overfitting data in the curve-fitting problem. Analogous problems often have analogous solutions and we propose to treat network inference as a case of model selection and use the Akaike Information Criterion (AIC). Strictly tree-like networks are more parsimonious than those that postulate lateral as well as vertical branches. This leads to the conclusion that we should not always infer LGT events whenever it would improve our fit-to-data, but should do so only when the improved fit is larger than the penalty for adding extra lateral branches.

Several philosophers have argued recently that semantic properties do play a causal role. 1 It is our view that none of these arguments are satisfactory. Our aim is to reveal some of the deficiencies of these arguments, and to reassess the question in our own way. In section 1, we shall explain in more detail what is involved in the pretheoretical idea of a causally efficacious property and so provide a fuller sense of the issue. In section 2 we shall discuss Fodor's and Kim's arguments that the semantic properties of mental events are causally efficacious. Their tactic is to articulate sufficient conditions for the causal efficacy of a property, and then apply these conditions to the psychological properties under consideration. We shall show that both formulations are inadequate. However, each account can remedy some of the defects of the other. We shall argue that by putting the two together, and modifying this combined account in certain ways, we can arrive at something that might just pass as a general sufficient condition for causal efficacy. We believe that the question of causal efficacy of content is best approached as a question in the philosophy of science. If one wants to know whether a given species of property is efficacious, the most productive strategy is to find a science that investigates the properties and examine the role they are assigned. In keeping with this attitude we shall restrict our discussion to the notion of content deployed within cognitive science. In section 3 we examine the role of content in cognitive science and discuss whether this role suits content to the conditions on efficacy articulated in section two. Our conclusion will be conditional and hedged: on certain views of cognitive science, content will meet the sufficient condition for efficacy. If other views are correct, the question remains open.

One possible interpretation of the species concept is that specifies are natural kinds. Another species concept is that species are individuals whose parts are organisms. Philip Kitcher takes seriously both these ideas; he sees a role for the genealogical/historical conception and also for the one that is “purely qualitative”. I criticize his ideas here. I see the genealogical conception at work in biological discussion of species and it is presupposed by an active and inventive research program, but the natural kind concept is not. I also criticize Kitcher’s criticisms against Hull’s position that species are individuals, including Kitcher’s idea that species are sets of organisms, his explanations of why “all swans are white” isn’t law-like, his discussion of the example of multiple origination. Finally, I argue that given the current evolutionary developments, pluralism is the "null hypothesis" that we should attempt to refute.

As a species of egoism, Hedonism holds that our only ultimate pleasure is the self-directed desire for pleasure and the avoidance of pain. Bishop Butler is widely regarded as having refuted hedonism. I argue that Butler's argument failed to undermine Hedonism, because his premises concern what people want, while Hedonism concerns why people have the wants they do. Even if the desires for external things were a prerequisite for obtaining pleasure, nothing would follow about why people desire external things.

Evolutionary theory is awash with probabilities. For example, natural selection is said to occur when there is variation in fitness, and fitness is standardly decomposed into two components, viability and fertility, each of which is understood probabilistically. With respect to viability, a fertilized egg is said to have a certain chance of surviving to reproductive age; with respect to fertility, an adult is said to have an expected number of offspring.1 There is more to evolutionary theory than the theory of natural selection, and here too one finds probabilistic concepts aplenty. When there is no selection, the theory of neutral evolution says that a gene’s chance of eventually reaching fixation is 1/(2N), where N is the number of organisms in the generation of the diploid population to which the gene belongs. The evolutionary consequences of mutation are likewise conceptualized in terms of the probability per unit time a gene has of changing from one state to another. The examples just mentioned are all “forwarddirected” probabilities; they describe the probability of later events, conditional on earlier events. However, evolutionary theory also uses “backwards probabilities” that describe the probability of a cause conditional on its effects; for example, coalescence theory allows one to calculate the expected number of generations in the past that the genes in the present generation find their most recent common ancestor.

Although the justification for using cladistic parsimony to infer phylogenetic trees has been extensively discussed, much less attention has been paid to the use of cladistic parsimony to reconstruct the character states of the ancestral species postulated by an inferred phylogenetic tree. These two problems differ in terms of both their inputs and their outputs, as shown in the following table. In the former, one begins with data on the character states of extant species and tries to find the best supported phylogenetic tree. In the latter, one begins with the tree that is most firmly supported by characteristics C1, C2,... Cn-1; one then takes some new characteristic Cn and records the character states of Cn that attach to the tree’s interior nodes, which represent common ancestors. In both cases, cladistic parsimony solves the problem by finding the hypothesis that requires the fewest changes in character state that are needed to explain the observations.

That some propositions are testable, while others are not, was a fundamental idea in the philosophical program known as logical empiricism. That program is now widely thought to be defunct. Quine’s (1953) “Two Dogmas of Empiricism” and Hempel’s (1950) “Problems and Changes in the Empiricist Criterion of Meaning” are among its most notable epitaphs. Yet, as we know from Mark Twain’s comment on an obituary that he once had the pleasure of reading about himself, the report of a death can be an exaggeration. The research program that began in Vienna and Berlin continues, even though many of the specific formulations that came out of those circles are flawed and need to be replaced.

I discuss two versions of the doomsday argument. According to ``Gott's Line'',the fact that the human race has existed for 200,000 years licences the predictionthat it will last between 5100 and 7.8 million more years. According to ``Leslie'sWedge'', the fact that I currently exist is evidence that increases the plausibilityof the hypothesis that the human race will come to an end sooner rather than later.Both arguments rest on substantive assumptions about the sampling process thatunderlies our observations. These sampling assumptions have testable consequences,and so the sampling assumptions themselves must be regarded as empirical claims.The result of testing some of these consequences is that both doomsday argumentsare empirically disconfirmed.

Instrumentalism is usually understood as a semantic thesis: scientific theories are neither true nor false, but are merely instruments for making predictions. Scientific realists are on firm ground when they reject this semantic claim. This paper focuses on epistemological rather than semantic instrumentalism. This form of instrumentalism claims that theories are to be judged by their ability to make accurate predictions, and that predictive accuracy is the only consideration that matters in the end. I consider how instrumentalism is related to a quite different proposal concerning how theories should be evaluated—scientific realism. Instrumentalism allows for the fact that a false model can get one closer to the truth than a true one.

The probability that the fitter of two alleles will increase in frequency in a population goes up as the product of N (the effective population size) and s (the selection coefficient) increases. Discovering the distribution of values for this product across different alleles in different populations is a very important biological task. However, biologists often use the product Ns to define a different concept; they say that drift “dominates” selection or that drift is “stronger than” selection when Ns is much smaller than some threshold quantity (e.g., ½) and that the reverse is true when Ns is much larger than that threshold. We argue that the question of whether drift dominates selection for a single allele in a single population makes no sense. Selection and drift are causes of evolution, but there is no fact of the matter as to which cause is stronger in the evolution of any given allele

How should the concept of evidence be understood? And how does the concept of evidence apply to the controversy about creationism as well as to work in evolutionary biology about natural selection and common ancestry? In this rich and wide-ranging book, Elliott Sober investigates general questions about probability and evidence and shows how the answers he develops to those questions apply to the specifics of evolutionary biology. Drawing on a set of fascinating examples, he analyzes whether claims about intelligent design are untestable; whether they are discredited by the fact that many adaptations are imperfect; how evidence bears on whether present species trace back to common ancestors; how hypotheses about natural selection can be tested, and many other issues. His book will interest all readers who want to understand philosophical questions about evidence and evolution, as they arise both in Darwin's work and in contemporary biological research.