Michael Strevens

High-level explanation would be impossible without a certain kind of independence of high-level behavior from low-level behavior. This partial autonomy has been characterized in various ways by scientists and philosophers; the present chapter advances a particular characterization deemed ‘semi-detachment’. The main business of the chapter is to pose and suggest an answer to the question why semi-detachment should be so widespread, enabling the existence of explanatory scientific disciplines—evolutionary biology, sociology, economics—that operate almost entirely separately from the disciplines that study their lower-level substrates (microbiology and psychology, for example). The approach explored here turns on low-level details' propensity to cancel or balance out in much the same way that the outcomes produced by a simple chance setup such as a coin toss balance out to produce stable frequencies. Such ‘canceling out’ considerations, it is argued, provide at least a partial account of why high-level explanations in the sciences should be, not merely possible, but plentiful.

Explanations are often probabilistic or statistical in nature. This chapter explores the contexts in which such explanations occur and provides accounts of their structure. The inductive-statistical model and models invoking intrinsic probabilistic causation are described and criticized, and the uses of statistical explanation in science are discussed. The subsumption model of statistical explanation and some of its demands and results are explored along with proposals that causation is essential even to statistical explanation and that statistical explanations are founded on irreducible stochastic dispositions. Indeterminism is outlined, and the chapter suggests that a notion of “robustness” is a requirement of a probabilistic explanation.

Research programs regularly compete to achieve the same goal, such as the discovery of the structure of DNA or the construction of a TEA laser. The more the competing programs share information, the faster the goal is likely to be reached, to society’s benefit. But the “priority rule”—the scientific norm according to which the first program to reach the goal in question must receive all the credit for the achievement—provides a powerful disincentive for programs to share information. How, then, is the clash between social and individual interest resolved in scientific practice? This chapter investigates what Robert Merton called science’s “communist” norm, which mandates universal sharing of knowledge, and uses mathematical models of discovery to argue that a communist regime may be on the whole advantageous and fair to all parties, and so might be implemented by a social contract that all scientists would be willing to sign.

Causal explanations in the high-level sciences typically black-box the low-level details of the causal mechanisms that they invoke to account for their explananda: economists’ black-box psychological processes, psychologists’ black-box neural processes, and so on. Are these black-boxing explanatory models complete explanations of the phenomena in question, or are they just sketches of or templates for the whole explanatory story? This chapter poses a focused version of the question in the context of convergent evolution, the existence of which appears to show that underlying mechanisms are completely irrelevant to the explanation of high-level biological features, including perhaps thought and language—in which case a black-boxing model would be a complete explanation of such features rather than a mere sketch. Arguments for and against such a model’s explanatory completeness are considered; the chapter comes down tentatively against.

In some scientific explanations, mathematical facts seem to be the primary bearers of enlightenment. Is this a case, in science, of “explanation beyond causation”? Might these explanations be causal only in part, or only in an auxiliary way, or not at all? To answer this question, this chapter examines some well-known examples of explanations that seem to operate largely or wholly through mathematical proof. It is concluded that the mathematical and the causal components of these explanations are complementary rather than rivalrous: the function of the mathematics is to help the explanations’ consumers better grasp relevant aspects of the causal structure that does the explaining, and above all, to better grasp how the structure causally makes a difference to the phenomena to be explained. The explanations are revealed, then, to be causal through and through.

Many varieties of understanding subsist in a thinker’s having the right kind of mental connection to a certain body of fact (or putative fact), a connection often called “grasp”. The use of a single term suggests a single connection that does the job in every kind of understanding. Then again, “grasp” might be an umbrella term covering a diverse plurality of understanding-granting mind-world relations. This paper argues for the former, unified view of grasp in two ways. First, it advances a broad, ability-based construal of grasp, along with a test for lack of grasp, that suggests that a certain specific connection plays an essential role in many varieties of understanding. Second, the paper considers a number of challenges to the thesis of unity that arise in a range of different kinds of understanding (scientific, moral, objectual, humanistic), and seeks to disarm them.

To understand why a phenomenon occurs, it is not enough to possess a correct explanation of the phenomenon: you must grasp the explanation. In this formulation, “grasp” is a placeholder, standing for the psychological or epistemic relation that connects a mind to the explanatory facts in such a way as to produce understanding. This paper proposes and defends an account of the “grasping” relation according to which grasp of a property (to take one example of the sort of entity that turns up in explanations) is a matter of recognitional ability: roughly, a property is grasped to the extent to which the would-be understander is capable of recognizing instances of the property.

It is only in the last three centuries that the formidable knowledge-making machine we call modern science has transformed our way of life and our vision of the universe - two thousand years after the invention of law, philosophy, drama and mathematics. Why did we take so long to invent science? And how has it proved to be so powerful?The Knowledge Machine gives a radical answer, exploring how science calls on its practitioners to do something apparently irrational- strip away all previous knowledge - such as theological, metaphysical or political beliefs - and channel unprecedented energy into observation and experiment. In times of climate extremes, novel diseases and rapidly advancing technology, Strevens contends that we need more than ever to grasp the inner workings of our knowledge machine.

A paradigm-shifting work that revolutionizes our understanding of the origins and structure of science. Captivatingly written, interwoven with tantalizing illustrations and historical vignettes ranging from Newton's alchemy to quantum mechanics to the storm surge of Hurricane Sandy, Michael Strevens's wholly original investigation of science asks two fundamental questions: Why is science so powerful? And why did it take so long, two thousand years after the invention of philosophy and mathematics, for the human race to start using science to learn the secrets of nature? The Knowledge Machine's radical answer is that science calls on its practitioners to do something irrational: by willfully ignoring religion, theoretical beauty, and, especially, philosophy-essentially stripping away all previous knowledge-scientists embrace an unnaturally narrow method of inquiry, channeling unprecedented energy into observation and experimentation. Like Yuval Harari's Sapiens or Thomas Kuhn's 1962 classic, The Structure of Scientific Revolutions, The Knowledge Machine overturns much of what we thought we knew about the origins of the modern world.

Dynamic approaches to understanding probability in the non-fundamental sciences turn on certain properties of physical processes that are apt to produce “probabilistically patterned” outcomes. The dynamic properties on their own, however, seem not quite sufficient to explain the patterns; in addition, some sort of assumption about initial conditions must be made, an assumption that itself typically takes a probabilistic form. How should such a posit be understood? That is the problem of initial conditions. Reichenbach, in his doctoral dissertation, floated a Kantian solution to the problem. In this paper I provide a Reichenbachian alternative.

Every model leaves out or distorts some factors that are causally connected to its target phenomenon -- the phenomenon that it seeks to predict or explain. If we want to make predictions, and we want to base decisions on those predictions, what is it safe to omit or to simplify, and what ought a causal model to describe fully and correctly? A schematic answer: the factors that matter are those that make a difference to the target phenomenon. There are several ways to understand differencemaking. This paper advances a view as to which is the most relevant to the forecaster and the decision-maker. It turns out that the right notion of differencemaking for thinking about idealization in prediction is also the right notion for thinking about idealization in explanation; this suggests a carefully circumscribed version of Hempel's famous thesis that there is a symmetry between explanation and prediction.

What is going on under the hood in philosophical analysis, that familiar process that attempts to uncover the nature of such philosophically interesting kinds as knowledge, causation, and justice by the method of posit and counterexample? How, in particular, do intuitions tell us about philosophical reality? The standard, if unappealing, answer is that philosophical analysis is conceptual analysis—that what we learn about when we do philosophy is in the first instance facts about our own minds. Drawing on recent work on the psychology of concepts, this book proposes a new understanding of philosophical analysis, which I call inductive analysis. The thesis that philosophical analysis is inductive analysis explains how novel, substantive philosophical knowledge can be generated in the armchair. It also explains why attempts at philosophical analysis tend to fall short of providing a complete and uncontroversial definition, and supplies reasons not to lament this apparent shortcoming.

Robert Batterman and others have argued that certain idealizing explanations have an asymptotic form: they account for a state of affairs or behavior by showing that it emerges “in the limit”. Asymptotic idealizations are interesting in many ways, but is there anything special about them as idealizations? To understand their role in science, must we augment our philosophical theories of idealization? This paper uses simple examples of asymptotic idealization in population genetics to argue for an affirmative answer and proposes a general schema for asymptotic idealization, drawing on insights from Batterman’s treatment and from John Norton’s subsequent critique.

Maxwell's deduction of the probability distribution over the velocity of gas molecules—one of the most important passages in physics (Truesdell)—presents a riddle: a physical discovery of the first importance was made in a single inferential leap without any apparent recourse to empirical evidence. Tychomancy proposes that Maxwell's derivation was not made a priori; rather, he inferred his distribution from non-probabilistic facts about the dynamics of intermolecular collisions. Further, the inference is of the same sort as everyday reasoning about the physical probabilities attached to such canonical chance setups as tossed coins or rolled dice. The structure of this reasoning is investigated and some simple rules for inferring physical probabilities from symmetries and other causally relevant properties of physical systems are proposed. Not only physics but evolutionary biology and population ecology, the science of measurement error, and climate modeling have benefited enormously from the same kind of reasoning, the book goes on to argue. Inferences from dynamics to probability are so obvious to us, however, that their methodological importance has been largely overlooked.