This chapter introduces a distinction between two sorts of scientific narrative, modelled on Ted Porter’s discussion of thick and thin description. In thin narratives, sequences of processes and experimental interventions are presented in a highly conventionalized form, their notation often assembled from a stock of familiar elements. Thick narratives, by contrast, offer a greater degree of context and contingency and may be attentive to social, environmental and other considerations. The distinction is discussed with examples from chemistry; I suggest that chemical reaction schemes, written to describe organic syntheses, are examples of thin narratives. But some chemists, as well as historians, geographers and sociologists who study chemistry, have expressed reservations about what such accounts leave out, and seek to develop modes for narrating chemical processes, experiments and impacts which can provide a thicker account.

This chapter examines the role of three kinds of narratives in producing knowledge about the rupture process of the Tohoku earthquake of 2011. I show that each of the three kinds of narratives appears in one of three stages on the way from data recorded of the earthquake to a reconstruction of the rupture process. In the first stage, rupture narratives are produced by computational tools called source models. In the second stage, a set of details that is taken accurately to represent features of the actual rupture process is distilled out of these conflicting rupture narratives through the use of a ‘research narrative’. In the third stage, these distilled details are strung together into an integrating narrative. This integrating narrative is used as a research tool for formulating questions, the pursuit of which has led to the production of further evidence about the rupture process.

One of the primary goals of archaeology is to construct narratives of past human societies through the material evidence of their activities. Such narratives address how people led their lives and how they viewed and interacted with their world at different times in the past. However, the way archaeologists look at time is becoming increasingly disparate, fragmented and sometimes contradictory. While we now have more exact ways of dating past remains and deposits, and more sophisticated ways of examining how past humans may have engaged with their physical and social environments, there is some internal confusion as to the relative merits of alternative interpretations and evidence. In the research drive to determine a greater precision of dating and chronology, the effect that increased dating effort has on the accuracy of archaeological narratives has rarely been discussed. This chapter discusses the problems and opportunities for archaeological narratives in approaches to time.

Mathematical proofs and narratives may seem to be opposites. Indeed, deductive arguments have been highlighted as clear examples of non-narrative sequences by narrative theorists. I claim that there are important similarities between mathematical proofs and narrative texts. Narrative texts are read in a quite distinct way, and I argue that mathematical proofs are often read like narrative texts by research mathematicians. In this way, narratives play an important role in mathematical knowledge-making. My argument draws on recent empirical data on how mathematicians read proofs. Furthermore, my examination of mathematical proofs and narratives provides an account of what it means for research mathematicians to understand mathematical proofs.

Throughout the nineteenth century, shipwrecks during tropical cyclones in the Indian Ocean resulted in extended legal battles in the Marine Court of Enquiry in Calcutta. This chapter explores how cyclones became an object of scientific curiosity at the intersection of the imperial legal world and marine insurance. It explores the court records, consisting of legal depositions about the wrecks by mariners and insurance agents, ships’ logs with barometric readings, and diaries kept by the captain and pilots, which formed a significant archive for the colonial scientist Henry Piddington (1797–1858), made famous for coining the term ‘cyclone’. Piddington narrativized storm observations by condensing accounts from multiple sources and created a ‘storm card’ to finally develop a theory of tropical cyclones. His storm narratives and the accompanying visualization through the storm card shaped the very object – the cyclone – as a scientific category of investigation, transforming storm memories into a narrative science of forecasting.

Storytelling can be understood as a performative social event that instantiates a specific relationship between storyteller and audience. This relationship supports inferences of narrative causation in hearers, both locally (episode x caused episode y) and globally (repeated patterns of causation at a more abstract level). This applies to passages of performative speech in a narrative event that are non-narrative, such as description or digression. Scientific writing is often conceived as non-performative and impersonal, with causation expressed explicitly. However, I suggest in this chapter that discourse of this kind can make the task of configuring global patterns of causation more difficult. Performative narrative discourse, on the other hand, offers support for readers in the task of remodelling existing theoretical causal structures through reconceptualization. I illustrate this argument through an analysis of narrative and non-narrative performative discourse in the field of cognitive psychology.

This chapter explores narratives that informed two influential attempts to automate and consolidate mathematics in large computing systems during the second half of the twentieth century – the QED system and the MACSYMA system. These narratives were both political (aligning the automation of mathematics with certain cultural values) and epistemic (each laid out a vision of what mathematics entailed such that it could and should be automated). These narratives united political and epistemic considerations especially with regards to representation: how will mathematical objects and procedures be translated into computer languages and operations and encoded in memory? How much freedom or conformity will be required of those who use and build these systems? MACSYMA and QED represented opposite approaches to these questions: preserving pluralism with a heterogeneous modular design vs requiring that all mathematics be translated into one shared root logic. The narratives explored here shaped, explained and justified the representational choices made in each system and aligned them with specific political and epistemic projects.

To give a Darwinian explanation of the traits of a species, it is not enough to show that the traits are appropriate for the environments inhabited. One must also show that the traits in question are more appropriate than the (presumed) ancestral traits from which they are derived. But one must go further still. Even if there is no question that the derived traits are more appropriate, one must still specify the sequence of modifications leading from the ancestral to the derived traits, each step of which is fitness-enhancing. How better – indeed, how else – than by a narrative? I illustrate these points through the evolution of flatfish eyes. This is part of an ongoing project concerning what narratives are good for, what narratives do better than non-narrative arguments: in short, why we need narratives.