Bayesian decision theory in its classical formulation supposes that for any rational agent and for any possible state x of the world, there is a number P(x). which represents the agent’s judgment of the probability of x. Similarly, the theory assumes that for any possible outcome y of the agent’s actions, there is a number u(y) which represents the utility or value of y to the agent. Given these assumptions, the theory is able to define the expected value for the agent of any act. Bayesian decision theory then recommends that the agent should choose, from amongst the available acts, one which has maximal expected utility.

The assumption that an agent has a determinate personal probability function has often been attacked.

We describe a general problem solving mechanism that is especially suited for performing a particular form of abductive inference, or best-explanation finding. A problem solver embodying this mechanism synthesizes composite hypotheses. It does so by by combining hypothesis parts as a means to the satisfaction of explanatory goals. In this way it is able to arrive at complex, integrated conclusions which are not pre-stored.

The intent is to present a computationally-feasible, task-specific problem solver for a particular information processing task which is nevertheless of very great generality. The task is that of synthesizing coherent composite explanatory hypotheses based upon a prestored, and possibly vast collection of hypothesis-generating “concepts”. The authors’ claim is nothing less than to have shown, in a new sense, and surpassing all other work in this area, how it is computationally possible for an agent to come to “know”, based upon the evidence of the case.

The units of selection problem is an issue within evolutionary theory (or the philosophy thereof) and concerns the question of what units or objects are acted upon by natural selection -- for example, whether these are genes, organisms or groups of organisms. One of the central theses of Elliot Sober’s recent book, The Nature of Selection, is that the philosophical problem of what it means for something to be a unit of selection is to be understood by applying the correct account of what it is in general for a factor to be a cause in a population.

The account of population-causation that Sober utilizes for this purpose comes from the recent literature on probabilistic causality. This account requires that for C to cause E in a population, a unanimity condition must be satisfied: C must raise the probability of E in each causal context.

The conflation of two fundamentally distinct issues has generated serious confusion in the philosophical and biological literature concerning the units of selection. The questions of how a unit of selection is defined, theoretically, is rarely distinguished from the question of how to determine the empirical accuracy of claims--either specific or general--concerning which unit(s) are undergoing selection processes. In this paper, I begin by refining a definition of the unit of selection, first presented by William Wimsatt, that is grounded in the structure of natural selection models. I then explore the implications of this definition for empirical evaluation of group selection models. I examine an objection to Wimsatt’s definition raised by Elliott Sober, and I conclude by indicating how the proposed refined definition can clarify which sorts of empirical evidence could be brought to bear on group selection controversies.

Recent attempts to clarify the fitness in evolutionary theory as a propensity (Brandon 1978; Brandon and Beatty 1984; Burian 1983; Mills and Beatty 1979; Sober 1984a, 1984b) or as a primitive theoretical term (Rosenberg 1983, 1985; Williams 1970, Williams and Rosenberg 1985) all miss the mark in clarifying the empirical content and explanatory power of natural selection theory.

I shall argue that the crucial distinction missing in these accounts is between the sense of fitness common in population genetics as actual relative rate of increase of genotypes and fitness in the more ordinary sense--and Darwin’s--of adaptedness of organisms. The relation between these senses of ‘fitness’ is that fitness as actual reproduction success depends on, is a function of, variables representing adaptive capacities and environmental properties.

“The question for us,” as Ronald Giere writes in Understanding Scientific Reasoning, “is whether analogies play any role in the JUSTIFICATION of [a] new theory.” Giere’s answer is an emphatic “No.” (Giere 1984, pp. 79-80). Although most philosophers of science would probably qualify Giere’s unmitigated rejection of analogical justification, few attribute much significance to analogical arguments in science. And when philosophers do grudgingly acknowledge an analogical argument, they are hesitant to analyze it.

Take, for example, Charles Darwin’s argument for natural selection. It is difficult to deny that the analogy between artificial and natural selection played an important justificatory role. After all, artificial selectibn was the topic of the first chapter of Darwin’s Origin of Species and was referred to in countless arguments throughout the text.2

An apparently peculiar form of explanation is found in evolutionary biology (and other historical fields); it is called a genetic explanation by Beckner (1959) (and, in a more general discussion, by Hempel (1965)), a narrative explanation by Goudge (1961), and a Darwinian history by Kitcher (1985). Kitcher, assuming that the Darwinian history has some kind of logically respectable structure, is primarily concerned with arguing that it is the cornerstone of Darwin’s historical methodology; Beckner and Goudge, on the other hand, assuming that the genetic (or narrative) explanation is an important part of evolutionary biology, are primarily concerned with the difficulty of fitting this form of explanation into a logically respectable structure. Adopting the narrative explanation terminology, and assuming that the narrative explanation is central to Darwin’s historical methodology and to the methodology of contemporary evolutionary biology, I will in this paper delineate its logically respectable structure.