This project will examine closely the interplay between the theoretical, experimental, and normative elements involved in the process of clarifying data for scientific modeling. Scientific models form the basis of scientific understanding and interaction with the world. Since the targets of scientific modeling activity are initially ill-defined phenomena, it is crucial to determine how phenomena are clarified and stabilized in a scientific inquiry. That process involves settling the distinction between genuine information and noise so as to identify intrinsic features and the relevant parameters that produce those features. It has received little philosophical attention in the literature, though it is a crucial stage of scientific activity. It determines how the phenomena are finally categorized, the methods and focus of interaction, and the directions for further investigation. In addition, it is sometimes the object of longstanding controversies and fundamental re-evaluation. The upshot is that this project will yield substantial understanding about an underappreciated yet important scientific process.
Intellectual Merit In this project, the main case study is Landau's model of the onset of turbulence, and in particular its use to study the wake that forms behind a cylinder that is perpendicular to a fluid flow; there is a long controversy over models of the wake of an obstacle in fluid flow. The modeling activity will be approached as situated within an epistemic network that contains theoretical and empirical informational resources as well as norms that guide and constrain the construction and the evaluation of models. How these informational resources and normative constraints interact in experimental investigation, as well as how experimentation and simulation activities are coordinated, are sometimes at the heart of scientific controversy. Here are some of the central issues to be addressed in the project: What counts as relevant parameter of the model of a phenomenon and as a measurement thereof; how activities of simulation and experimentation inform, guide, and constrain each other; and how the impact of a model on the dynamics of the network, as investigative instrument, contributes to its epistemic value.
Potential Broader Impacts This project approaches scientific models in a new way. Unlike in more usual approaches, focus will be on the earlier stages of inquiry before model and experimental arrangement have become fixed or stabilized. A better understanding of the process of modeling, and of the characterization of phenomena that guides development of research programs in science, will thus be achieved. In particular, the results will provide a new form of interpretation of scientific conflicts and controversies, by bringing to light the role of judgments of relevance in the construction and evaluation of scientific models. This approach participates in a larger effort to foster constructive interactions between science and philosophy of science, and thus to facilitate the participation of philosophy of science in science policy debates concerning scientific development. Additionally, it will make philosophical study of science more attractive to philosophy students, and more relevant to science students, by uncovering the interactive and creative dynamics of scientific research.
Recent studies of scientific modeling and experimentation have contributed significantly to the understanding of scientific practice. But focusing on the role of models as representations of phenomena, and on the role of experiment as a tribunal where such models are tested as to representational adequacy, as is generally the case, has significant limitations. This project undertakes to explore precisely how the construction of model is related to experimentation, focusing on how these two activities are intertwined in practice. Thus our research project is constructed around one main goal: a better understanding of the process of construction of models. The sub-goals were: 1) a better understanding of the role of experimentation in this construction; 2) a better understanding of the interaction between experimentation and simulation; 3) a better understanding of the criteria of evaluation of models. Goal 1) was most fully realized. On the basis of case studies in fluid mechanics and neurobiology, we were led to focus on the notions of relevance and relevant factors. We defined the problem of relevance as the problem of determining what factors are relevant to the modeling of the phenomenon of interest and should figure in a model of the phenomenon. We also introduced a crucial distinction between two notions of relevance: local and general. Generally relevant factors form the dimensions of the total space in which the phenomenon under study develops. But that in a given experiment, the effect of some causally efficient factors is neutralized does not imply that they are absolutely irrelevant to the characterization of the phenomenon. They may be deemed locally non-relevant and be kept for further exploration. Developed in the PI’s articles in 2011 and 2012, this subject was most fully treated in the joint PI and Co-PI work presented at an international conference in the Netherlands, with publication to appear in 2013. Significant progress was made by the PI with respect to goal 2), with detailed evaluation of subtle but far-reaching differences between experimentation and simulation despite salient similarities described in the literature. Results, both published and presented at conferences by the PI engaged with current research by Michael Weisberg and Eric Winsberg, in the context of currently lively debates in the field. Crucial to the topic of goal 3) was the PI’s work on the ‘generative constructive use’ of models as basis for the construction of new models and the generation of new target systems. This role was illustrated with case studies of the HKB model in coordination dynamics and, in greater detail, with the model of a wake in fluid dynamics. It is shown how this use makes an essential contribution to the scientific significance of these models. In addition to the published results this work included fruitful interaction of the PI with scientists engaged in fluid dynamics in France and Romania. Additionally this subject was pursued by the Co-PI with special attention to the role of modeling in measurement. With the organization of the Experimental Side of Modeling Workshops and sessions of the Bay Area Philosophy of Science Working Group, the subject of this project reached a wide audience of students and professionals in our area. Additionally, the identification of the problem of relevance and the distinction between local and general relevance of factors in experimentation has an impact on how it is possible to understand a pluralist stance about scientific models. The pluralist stance states that there might be different models of a given phenomenon and that these models might constitute a non-unifiable plurality of partial knowledges. We show that different models will form a unifiable plurality if they pertain to domains of local relevance that are part of the same larger domain of general relevance and if their domains of local relevance can be situated with respect to one another within the space of general relevance.
