Kuhn left many readers with the impression that his version of paradigm choice was not a process governed by reason. That impression was correct. This chapter examines the development of Kuhn’s thinking on paradigm choice during and after Structure, focusing on a philosophical challenge here called the “puzzle of promise”: how can paradigm choice be conceived of as rational if it is, as Kuhn claims, made in defiance of the evidence from problem solving? The chapter argues that later developments of Kuhn’s views elaborate a platform for solving this puzzle, but that he himself always clung to the view that paradigm choice is an arational process.

Humanistic inquiry is not just about timeless questions and human experience. Viewed historically, it is equally about working within the constraints of a world of ideas shaped by a small set of exemplars. In this chapter I look at concrete instances of the use of canons in the history of the humanities. Different cultures designate different works as canonical. The point is not that everyone interested in the questions posed by the humanities should be reading the same works, but rather that humanistic inquiry in each community (however defined) must designate certain works as canonical in order to reap the scholarly benefits of a shared world.

In this chapter, I argue that interest (or interestingness) can serve as a reliable guide to the detection of important research topics, but only under specific conditions. My argument begins by reflecting on a specific cultural difference between the natural sciences and the humanities that exemplifies the disparate roles that interestingness plays in the two domains. I explain how this specific difference is a symptom of a much more far-reaching contrast between them, and this contrast tracks differences in the tendency of something of genuine scholarly significance to emerge from the pursuit of research questions that one finds interesting. I argue that disciplines whose members do not agree on a large corpus of substantive claims — in short, disciplines in which consensus is a rarity, if it exists at all — do not function as intellectual communities capable of cultivating a causal link between something’s being interesting and something’s being important.

The pronounced uniformity of learning in the natural sciences is a symptom of the relatively high propensity of natural science disciplines to strive for and achieve consensus on matters of scholarly concern. The culture of consensus that pervades the natural sciences has significant effects on the prospects for scientific knowledge. Indeed, owing to the socially infused conception of scientific knowledge that has become increasingly influential in the last few decades, many scholars have begun to embrace the idea that consensus is an essential component of scientific knowledge — that is, that what it means for knowledge to be scientific is, in part, for it to be an object of consensus among relevant members of the scientific community.

If some of our knowledge cannot be articulated, how does it make itself manifest? It will not surprise anyone who has followed the argument of this book up to now that there are things that we can do with knowledge besides talking about it. Millikan, as we saw, used his knowledge of experimentation and of professional discourse to guide his exemplary investigations of the charge of the electron. Neither was something he made explicit; I doubt that he (or anyone) could have. No practitioner who looked at Millikan’s work found any basis for these accusations, because their training endowed them with a knowledge only available to practitioners. They all made effective use of this knowledge, despite not being able to articulate its content. That kind of knowledge manifests itself not in the form of beliefs, but rather in the scholar’s sense of how things seem.

We tend to think of scientific knowledge as the paradigm case of what knowledge should look like. But the dominant image of science does not reflect what actually goes on in the making of scientific knowledge. Real scientific knowledge relies heavily on human judgment. And it relies heavily on communities.

In this chapter, I want to take a close look at two distinct but related phenomena in the humanities, each of which marks a stark contrast with the natural sciences for reasons that are not altogether clear. The first involves a recent episode in which several deliberately nonsensical articles were published as part of an effort to expose a certain family of humanities disciplines as intellectually hollow. I will argue that the general phenomenon it illustrates does in fact represent a serious problem in the humanities. The second phenomena is the general absence of article retraction in the humanities. Retraction plays an important epistemic role in science. Now, why is the retraction rate in the humanities holding steady at around 0%? Is this cause for celebration, cause for concern, or an insignificant byproduct of the fundamental difference between scientific knowledge and humanistic knowledge?

In this chapter, I aim to motivate two important theses. The first is the idea that the humanities enable us obtain knowledge of ourselves by providing the means to read what lies within. Knowledge of ourselves is not scientific knowledge. It never will be. It is made up of the facts concerning what we implicitly take to be of value. Knowledge of ourselves is a special kind of “knowledge of what matters.” It is knowledge of what matters to us. Humanities disciplines foster the exploration of that source of value and its congruence with our desires in various ways. Ultimately, I argue, these various forms of exploration succeed by enabling us to achieve a vantage point from which to view our personal commitments and desires with some variety of detachment. In doing so, they provide the means to avoid a pernicious kind of self-contamination.

We learn about what matters from exemplars. But we learn more from exemplars than we can be taught, and we know more about them than we can say. The challenge of articulating that knowledge is what drives scholarly inquiry forward. Our epistemological mythology confuses those efforts to express that knowledge propositionally with the knowledge itself. Propositional representations of our knowledge are the currency of scholarly debate; there is no alternative. These representations attempt to distill clusters of salient features that we perceive an exemplar as exemplifying, features that seem to us to matter most for our understanding of the relevant phenomenon. The more precisely we can render these features, the more they stand out and demand our attention. As we endeavor to produce ever finer-grained articulations, we inadvertently exclude other features, features which themselves might eventually emerge as salient and which thus call for fundamental revisions in how we represent what we know.

The natural sciences produce knowledge. Not necessarily because they do experiments, or because they use precise measurement devices, or because they investigate reality, but because they have developed highly conservative epistemic cultures whose members are overwhelmingly concerned with what the community thinks. My purpose in this chapter is to support this claim as one component of more general conception of disciplinary knowledge, a species of knowledge of which both the natural sciences and the humanities have historically been able stewards. If we use the natural sciences as a model for what real knowledge looks like, the question, “Do the humanities create knowledge?” turns not so much on the degree to which the humanities employ the Scientific Method, but on the degree to which they partake of the social processes by which disciplinary knowledge is achieved.

Humanities scholarship is hard to classify as knowledge because it normally takes a form that is quite different from scientific research, which has become nearly synonymous with knowledge. Our popular picture of scientific knowledge is a based on a caricature — the Scientific Method — which misrepresents the nature of scientific inquiry in a way that makes it look fundamentally at odds with the tenor of the humanities. When we look at how scientific knowledge actually develops, however, we find that both the process by which this occurs and the structure of the scholarly communities that oversee it are surprisingly similar to those that have defined the humanities since antiquity. This book argues for a kind of knowledge — disciplinary knowledge — that is characterized by certain social practices surrounding the production of knowledge. Both the humanities and the natural sciences routinely create disciplinary knowledge.