C Chapter Contents

Chapter 1 is for readers with little or no background in philosophy or for those who would benefit by a short refresher in basic topics such as fallibilism, objec­tivity, deduction, induction, etc.

Chapter 2 describes what a modern scientific hypothesis is, distinguishes hy­pothesis from prediction, and explains their relationships. We need a rich and precise language for discussing the hypothesis, and Chapter 2 also discusses science and truth and emphasizes the flexible, recursive nature of the Scientific Method. I will define associated concepts, such as direct and indirect evidence, using both scientific and nonscientific examples. What makes for a good hypo­thesis? I cover non-obvious predictions (riskiness), parsimony, scientific signifi­cance, specificity, and constraint. Not all hypotheses are explicit, and there is the crucial, although underappreciated, influence of background assumptions that are nonetheless implicit hypotheses.

If science had a philosophical patron saint, it would be Karl Popper, an ex­tremely influential, though often maligned, thinker. Chapter 2 goes into the programs of Popper and John Platt in greater depth than you find in science courses. Almost everyone has heard about Popper's doctrine of falsification, although his critics distort it so much that you may wonder why you heard about it. I'll try to clear things up. And I'll go over other elements of Popper's philos­ophy such as “tested-and-not-falsified” hypotheses, corroboration and how it compares with confirmation, and other issues that scientists are rarely taught. Platt's practical experimental program of strong inference augments Popper's phi­losophy in a couple of significant ways, and I'll review it briefly as well.

Philosophers of science would no doubt disapprove of my take on Popper's philosophy if they happened to hear about it. “It's so, I don't know, so 1970s,” they'd say. “We've come so far since then.” I plead guilty to a friendly reading of Popper, although I'll point out some issues that he glosses over. On the other hand, I'll also criticize his critics and argue that, while the philosophers have un­questionably moved away from Popper, they disagree on where they are now, which gives scientists little to go on.

Chapter 3 is somewhat more advanced than Chapters 1 and 2. It is intended for readers who want to dig a little into the nuances of Popper's and Platt's programs or who have questions about them. I provide answers to several common objections that philosophers raise, for example the problem of holism and negative data.

Disputes can arise when critics point to various non-hypothesis-based modes of science and declare that, therefore, hypotheses are unnecessary. The conclu­sion doesn't follow. Chapter 4 distinguishes among several major “kinds” of sci­ence, including “confirmatory versus exploratory,” “natural versus social,” Big Data versus Little Data, and makes the case that they all fit comfortably within the realm of modern science.

Besides the scientific hypothesis scientists also use the statistical hypothesis, which, despite its name, has little in common with the scientific hypothesis. Chapter 5 compares the two classes of hypothesis by highlighting their concep­tual similarities and differences, such as how they incorporate empirical versus numerical content and how they relate to the real world. While familiarity with basic statistics terminology is helpful, expertise is not required because I focus on fundamental issues in general terms. In addition to exposing hidden com­plexities of the typical null hypothesis statistics methods, we'll look at how pos­sible replacements or supplements to p- valued testing, including confidence intervals, effect sizes, and errors, affect the way in which we conceive of the statis­tical hypothesis.

The statistical discussion continues in Chapter 6. Whereas conventional frequentist statistical ideas are the focus of Chapter 5, we also need to be acquainted with Bayesian statistics, so I'll cover the basics of Bayesianism in this chapter. I believe that relatively few bioscientists learn about Bayesian statistics, but Bayesian methods have an increasingly prominent role in science, and the Bayesian philosophy of science is very different from the one we're used to. For example, Bayesians and traditionalists disagree on what probability means, as well as on the purpose of scientific research. I'll compare them and illustrate how the Bayesian approach can be incorporated into a standard hypothesis testing procedure.

In Chapter 7, I'll dive into the controversy surrounding the Reproducibility Crisis, which centers on the fraught question: Is science reliable? The acid test of scientific reliability has been reproducibility—when scientists in their own labora­tories can duplicate the results reported by scientists in different laboratories. But what does “reproducibility” really mean, and do reported difficulties in reproduci­bility constitute a crisis, a problem, or just alarmist talk? Rather than a comprehen­sive review, this chapter goes through evidence for and against the notion that a crisis exists, and it highlights roles that hypothesis-based science can play in easing the stresses that do exist. The Reproducibility Project: Psychology and the statistical critiques of John Ioannidis and colleagues are at the center of the chapter.

Chapter 8 makes the case in favor of the scientific hypothesis from two dis­tinctly different points of view: statistical and cognitive. To demonstrate the statistical advantages of the hypothesis, I pick up the discussion of the Reproducibility Crisis from where we left it in Chapter 7. I explore reasoning that implies that hypothesis-based research in general will be more reliable than, for example, open-ended gene searches.

Pronouncements about how scientists perceive and use the hypothesis are rarely accompanied by actual evidence so, in Chapter 9, I report the results of a SurveyMonkey poll of several hundred members of scientific societies that I conducted to find out how scientists themselves think about the hypothesis and related matters. The poll covered topics from the extent of the respondents' formal instruction (70% had be better. He is not alone: typical working scientists do not have much use for philosophy, making their way more or less satisfactorily while ignoring it. Indeed, many philosophical issues are so arcane that we can safely skip them. Why not skip all of them? There are several reasons. Philosophers have thought a great deal about topics that scientists, whether we like it or not, must confront because these topics directly affect how we think about and do science. Critics of hypothesis-based science, including the scientist-critics whom we’ll meet in Chapter 10, ground their arguments firmly in philosophical concepts. Finally, the scientific hypothesis, the primary topic of this book, is enmeshed in a web of philosophical associations acquired long before science emerged from philos­ophy. We’ll need to be familiar with these associations to understand the hypo­thesis and the arguments surrounding it.

Despite their current lack of interest in philosophy, scientists themselves used to take part in pragmatic philosophical discussion. The Nobel Prize-winning Spanish neuroanatomist, Santiago Ramon y Cajal, begins his witty and still in­formative Advice for a Young Investigator,¹ written in 1898, with this: “I shall assume that the reader’s general education and background in philosophy are sufficient to understand that the major sources of knowledge include, observa­tion, experiment and reasoning by induction and deduction.”

Because modern undergraduate and graduate science curricula are jam- packed with science courses and other requirements, it is doubtful that all scientists nowadays have the background that Ramon y Cajal had in mind. We’ll also have to confront the estrangement that exists between science and philos­ophy; except on occasion, they don’t talk. Like my colleague, most scientists ei­ther brush philosophy aside as irrelevant or are openly hostile to it. Nonetheless, we do need to know a few key concepts, as even die-hard, anti-philosophy scientists agree.²

1.