Martin Zach

Conceptual analysis as a method of inquiry has long enjoyed popularity in analytic philosophy, including the philosophy of science. In this article I offer a perspective on the ways in which the method of conceptual analysis has been used, and distinguish two broad kinds, namely philosophical and empirical conceptual analysis. In so doing I outline a historical trend in which non-naturalized approaches to conceptual analysis are being replaced by a variety of naturalized approaches. I outline the basic characteristics of these approaches with illustrative examples, arguing that recent developments in the philosophy of science show that in order to achieve a more adequate understanding of scientific endeavour we need to prioritize the naturalized accounts of the method.

Abstraction and idealization are the two notions that are most often discussed in the context of assumptions employed in the process of model building. These notions are also routinely used in philosophical debates such as that on the mechanistic account of explanation. Indeed, an objection to the mechanistic account has recently been formulated precisely on these grounds: mechanists cannot account for the common practice of idealizing difference-making factors in models in molecular biology. In this paper I revisit the debate and I argue that the objection does not stand up to scrutiny. This is because it is riddled with a number of conceptual inconsistencies. By attempting to resolve the tensions, I also draw several general lessons regarding the difficulties of applying abstraction and idealization in scientific practice. Finally, I argue that more care is needed only when speaking of abstraction and idealization in a context in which these concepts play an important role in an argument, such as that on mechanistic explanation.

According to a widely held view, scientific modelling consists in entertaining a set of model descriptions that specify a model. Rather than studying the phenomenon of interest directly, scientists investigate the phenomenon indirectly via a model in the hope of learning about some of the phenomenon’s features. I call this view the description-driven modelling (DDM) account. I argue that although an accurate description of much of scientific research, the DDM account is found wanting as regards the mechanistic modelling found in many branches of biology. By analysing research practices in cancer immunology concerning the development of mechanistic models of the process of cancer metastasis, this paper presents and argues for a complementary account of scientific modelling, herein called the experimentation-driven modelling (EDM) account. In EDM, scientists investigate a set of experimental systems and then integrate the results obtained from experiments into a mechanistic model. While EDM shares some key features with DDM, the two are epistemically very different approaches to research.

In this paper we address the issue of how to think about immunity. Many immunological writings suggest a straightforward option: the view that the immune system is primarily a system of defense, which naturally invites the talk of strong immunity and strong immune response. Despite their undisputable positive role in immunology, such metaphors can also pose a risk of establishing a narrow perspective, omitting from consideration phenomena that do not neatly fit those powerful metaphors. Building on this analysis, we argue two things. First, we argue that the immune system is involved not only in defense. Second, by disentangling various possible meanings of ‘strength’ and ‘weakness’ in immunology, we also argue that such a construal of immunity generally contributes to the distortion of the overall picture of what the immune system is, what it does, and why it sometimes fails. Instead, we propose to understand the nature of the immune system in terms of contextuality, regulation, and trade-offs. We suggest that our approach provides lessons for a general understanding of the organizing principles of the immune system in health and disease. For all this to work, we discuss a wide range of immunological phenomena.

A central goal of the scientific endeavor is to explain phenomena. Scientists often attempt to explain a phenomenon by way of representing it in some manner—such as with mathematical equations, models, or theory—which allows for an explanation of the phenomenon under investigation. However, in developing scientific representations, scientists typically deploy simplifications and idealizations. As a result, scientific representations provide only partial, and often distorted, accounts of the phenomenon in question. Philosophers of science have analyzed the nature and function of how scientists construct representations, deploy idealizations, and provide explanations. As such, our aim in this special issue is to bring these three pillars of research into closer contact with the contributions to it focusing on three main themes. The first set of papers, Alan Baker (2021) and Marc Lange (2021), address mathematical explanations in science. Baker (2021), a proponent of mathematical Platonism, examines its capacity to evade the critique that the so-called Enhanced Indispensability Argument is circular. Lange (2021) examines distinctively mathematical explanations, arguing that neither Platonism nor representationalism are successful paths, and instead argues in favor of Aristotelian realism. A second theme emerging from the papers in this special issue is the impact that various conceptualizations of idealization have on our abilities to offer scientific explanations, to pro- duce an analysis of what explanations are or should be, and to understand scientific representation. Peter Tan (2021) suggests amending inferentialist accounts of scientific representation to account for inconsistent idealizations. Michael Strevens (2021) advocates a view wherein the introduction of idealizations into a model is legitimate so long as it pertains to non-difference-making factors, arguing for a logical reading of the notion of difference-making. Natalia Carrillo and Tarja Knuuttila (2022) offer an alternative account to the idealization-as-distortion view, emphasizing instead the holistic nature of idealization. Finally, contributions by Carrillo and Knuuttila (2022), Terzian (2021), Valente (2021), and Rodriguez (2021) illustrate how issues regarding idealization, representation, and explanation are applied to specific contexts and across various sciences. Carrillo and Knuuttila (2022) examine conceptions of idealization in the context of models of the nerve impulse. Giulia Terzian (2021) extends the discussion of idealizations to the context of generative linguistics. Giovanni Valente (2021) examines how idealizations and evaluations of accurate representation impact capacities to explain phenomena in the context of statistical thermodynamics. Quentin Rodriguez (2021) examines the role of idealizations and analogies in various strategies to explain critical phenomena. In what follows, we offer a brief overview of important philosophical issues connected to representation, idealization, and explanation in science. We then provide short summaries of the eight papers in this special issue.

Agent-based models (ABMs) are one of the main sources of evidence for decisions regarding mitigation and suppression measures against the spread of SARS-CoV-2. These models have not been previously included in the hierarchy of evidence put forth by the evidence-based medicine movement, which prioritizes those research methods that deliver results less susceptible to the risk of confounding. We point out the need to assess the quality of evidence delivered by ABMs and ask the question of what is the risk that assumptions entertained in ABMs do not include all the key factors and make model predictions susceptible to the problem of confounding.

Many scientific disciplines rely on the construction and use of models: biomedical sciences are no exception. This PhD thesis addresses several aspects of the practice of scientific modeling. First, I discuss the nature of modeling as such, proposing a novel, complementary account of scientific modeling which I term the experimentation-driven modeling account and which drives the construction of mechanistic models in many fields of biological and biomedical research, such as cancer immunology. Second, I scrutinize an objection to the mechanistic account of explanation according to which the account fails to accommodate the common practice of idealizing difference-making factors. I argue that this objection ultimately fails because it is riddled with a number of conceptual inconsistencies. Third, I analyze the roles of similarity judgments in some fields of cancer research which employ a variety of mouse models to learn about the disease mechanisms, arguing that by appreciating the epistemic complexities it is possible to shed new light on more general philosophical debates regarding scientific representation. Fourth, mechanisms can also be studied using more theoretical apparatus in the form of simulations. I investigate an example of an agent-based model used to model the outbreak of SARS-CoV-2...

Many microbes responsible for infectious diseases are known to run an asymptomatic course in a significant portion of the population. By highlighting the conceptual complexities of host-microbe interactions, this paper elucidates the fact that while many infections remain asymptomatic, this does not necessarily mean that such infections are of no concern for health. The paper builds on the so-called damage-response framework and considers several developments required to gain a more comprehensive perspective on infections and their relationship to diseases. Irrespective of their (short-term) clinical manifestation, infections leave an imprint with consequences for health. Finally, these considerations regarding host-microbe interactions must be incorporated into policy decisions and public understanding of health if we hope to handle future pandemics such as Covid-19 better.

Scientific knowledge relies heavily on models, shaped by simplifying assumptions, with common categories being abstraction and idealization. This article aims to expose conceptual challenges inherent in conventional interpretations of these concepts, particularly in their practical application to scientific modeling. The primary hurdle emerges in applying these categories to real­world instances of scientific modeling, which we illustrate with examples of non­causal explanations. Key issues revolve around (i) the ambiguous distinction between abstraction and idealization and (ii) the application of the simplifying assumption of abstraction. Our hypothesis posits that non­causal explanations face unintelligibility due to an unclear understanding of the role of simplifying assumptions in them. To test this, we analyze selected examples, ranging from (toy­)examples to real­world instances, scrutinizing the alignment with the standard notions of abstraction and idealization. Throughout, we investigate the influence of simplifying assumptions on these explanations, assessing their adherence or deviation from conventional concepts.