I argue that diversity and pluralism are valuable not just for science but for philosophy of science. Given the partiality and perspectivism of representation, pluralism preserving integration can increase accuracy. Perspectivism is often supported by appeal to visual representation. I draw further insights from multimodel sensory integration for understanding experiment-based predictions of protein structure. The epistemic lessons learned from the scientific case also apply to philosophy of science itself. Finally, I suggest that a critical, nuanced philosophical view of legitimate sources of pluralism in science has an important role to play in public discourse.

Model-data symbiosis is the view that there is an interdependent and mutually beneficial relationship between data and models, whereby models are data-laden and data are model-laden. In this article I elaborate and defend the second, more controversial, component of the symbiosis view and construct a taxonomy of the different ways in which theoretical and simulation models are used in the production of data sets. Each is defined and briefly illustrated with an example from the geosciences. I argue that model-filtered data are typically more accurate and reliable than so-called raw data and, hence, beneficially serve the epistemic aims of science.

I introduce a distinction between general and local model-data symbiosis and offer three examples of local symbiosis in the fields of meteorology and climate science. Local model-data symbiosis refers to a beneficial, two-way interdependence between a particular model and data set or between closely related models and data sets. For each example presented here, I show how the symbiotic relationship works—what the interdependence consists in and how it is supposed to be beneficial—and I consider whether, alongside the benefits, there is some risk of problematic circularity.

A number of philosophers draw a close analogy between scientific modeling and fiction, often appealing to Kendall Walton’s make-believe view. I assess the models-fictions analogy from a cognitive angle, suggesting that from this perspective it appears relatively weak. More specifically, I argue that, on the one hand, the appeal to Walton is appropriate inasmuch as his view fits well with how modelers employ the imagination. On the other hand, what makes Walton’s view attractive as an account of the cognitive aspects of modeling makes it less attractive as an account of fiction.

Most attempts to justify the use of idealized models to explain appeal to the accuracy of the model with respect to difference-making causes. In this article, I argue for an alternative way to justify using idealized models to explain that appeals to universality classes. In support of this view, I show that scientific modelers seeking to explain stable limiting behaviors often explicitly appeal to universality classes in order to justify their use of idealized models to explain.

A common strategy for simplifying complex systems involves partitioning them into subsystems whose behaviors are roughly independent of one another at shorter timescales. Dynamic causal models clarify how doing so reveals a system’s nonequilibrium causal relationships. Here I use these models to elucidate the idealizations and abstractions involved in representing a system at a timescale. The models reveal that key features of causal representations—such as which variables are exogenous—may vary with the timescale at which a system is considered. This has implications for debates regarding which systems can be represented causally.

This article defends the notion of downward causation, relating it to a notion of conditional independence.

Reflection and martingale principles are central to models of rational learning. They can be justified in a variety of ways. In what follows we study martingale and reflection principles in the context of measure theory. We give special attention to two approaches for justifying these principles that have not been studied in that context before: diachronic coherence and the value of information. Together with an extant argument based on expected accuracy, these arguments lend support to the thesis that reflection and martingale principles govern rational learning.

We examine contemporary game-theoretic accounts of ownership as a convention. New results from dynamic networks complicate matters, suggesting that if ownership is conventional, it should not be as prevalent as it seems to be. In fact, such models reveal a tendency toward antiownership norms. The value of resources may be crucial: low stakes lead to conventional ownership, but ownership norms rarely evolve; high stakes lead to a predominance of ownership at the cost of its conventionality. We argue that conventional ownership norms can originate in nonconventional ways and discuss some philosophical implications.

All natural languages exhibit a distinction between content words (nouns, verbs, etc.) and function words (determiners, auxiliaries, tenses, etc.). Yet surprisingly little has been said about the emergence of this universal architectural feature of human language. This article argues that the existence of this distinction requires the presence of nontrivial compositionality and identifies assumptions that have previously been made in the literature that provably guarantee only trivial composition. It then presents a signaling game with variable contexts and shows how the distinction can emerge via reinforcement learning.

We present here a hierarchical model for the evolution of compositional language. The model has the structure of a two-sender/one-receiver Lewis signaling game augmented with executive agents who may learn to influence the behavior of the basic senders and receiver. The model shows how functional agents might coevolve representational roles even as they evolve a reliable compositional language in the context of costly signaling. When successful, the evolved language captures both the compositional structure of properties in the world and the compositional structure of successful actions involving those properties.

The understanding that comes with scientific explanation is regarded as one of the central epistemic aims of science. In earlier work I have argued that scientists achieve understanding of phenomena by basing their explanations on intelligible theories, where intelligibility is a contextually determined value. In this article, I address the question of how the aim of understanding relates to other epistemic aims of science, such as prediction of empirical evidence and accurate description of phenomena. Moreover, I examine the associated values and analyze their role and interaction through a historical case study.

In this article, I first outline the view developed in my recent book on the role of idealization in scientific understanding. I discuss how this view leads to the recognition of a number of kinds of variability among scientific representations, including variability introduced by the many different aims of scientific projects. I then argue that the role of idealization in securing understanding distances understanding from truth but that this understanding nonetheless gives rise to scientific knowledge. This discussion will clarify how my view relates to three other recent books on understanding by Henk de Regt, Catherine Elgin, and Kareem Khalifa.

Several argue that truth cannot be science’s sole epistemic goal, for it would fail to do justice to several scientific practices that advance understanding. I challenge these arguments but only after making a small concession: science’s sole epistemic goal is not truth as such; rather, its goal is finding true answers to relevant questions. Using examples from the natural and social sciences, I then show that scientific understanding’s epistemically valuable features are either true answers to relevant questions or a means thereof.

Typically, the imagination in thought experiments has been taken to consist in mental images; we visualize the state of affairs described. A recent alternative from Fiora Salis and Roman Frigg maintains that it is only the propositional imagination that is necessary for the conduct of a thought experiment. I set out problems with these monistic accounts and develop a pluralist stance. Thought experiments appeal to a variety of our imaginative capacities, and we ought to focus on the function of particular thought experiments when considering what type of imaginative engagement they invite.

Imagination is important for many things in science: solving problems, interpreting data, designing studies, and much else. Philosophers of imagination typically account for the productive role played by imagination in science by focusing on how imagination is constrained, for example, by self-imposed rules to infer logically or model events accurately. But the constraints offered by these philosophers constrain either too much or not enough, and they can never account for uses of imagination that are needed to break today’s constraints in order to make progress tomorrow. Thus, epistemology of imagination needs to make room for an element of epistemological anarchy.

Using climate policy debates as a case study, I argue that a certain response to the argument from inductive risk, the hedging defense, runs afoul of a reasonable ethical principle: the no-passing-the-buck principle according to which scientists ought to offer their scientific judgment when such judgment is both possible and needed in public policy debates.

Wendy S. Parker and I have advanced an inductive-risk approach to the provision of climate information that relies on the contextual values of information users. This approach aims to improve the effectiveness of climate information in social decision making. The approach’s emphasis on user values, however, conflicts with the so-called democratic view: if scientists employ contextual values, they ought to employ democratically endorsed values to preserve political legitimacy. This article draws on deliberative democratic theory to resolve the conflict, demonstrating that whether user values are legitimate to employ depends on the way labor is divided across deliberative moments.

The article examines the idea that commercialized science is a central factor in the erosion of trust in science. I claim that commercial interests have a negative impact on the trustworthiness of science by leading the public to mistakenly trust or distrust scientific results. I identify two mechanisms. First, commercial interests can directly encourage the production of consensus or dissent that is industry friendly but not epistemically reliable. Second, commercial interests support a narrow understanding of trustworthiness that allows the industry to claim that their research is reliable when it is not. A more fine-grained analysis is needed.

Scientists are becoming increasingly aware of a “replicability crisis” in the behavioral, social, and biomedical sciences. Researchers have made progress identifying statistical and methodological causes of the crisis. However, the social structure of science is also to blame. In the fields affected by the crisis, nobody is explicitly responsible and rewarded for doing confirmation and replication work. This article makes the case for a social structural reform to address the problem. I argue that we need to establish a reward system that supports a dedicated group of confirmation researchers and formulate a proposal that would achieve this.