Medium-Strength Modality in Science

Alethic modality is typically understood in terms of necessity, truth in all possible cases, and possibility, truth in at least one possible case. But a great deal of our scientific knowledge falls in between necessity and contingency. We build models which apply to many, but by no means all, possible cases under consideration. Further, alethic modal (henceforth just â modalâ ) claims have typically been understood as applying to very broad domains, like conceptual, logical, or physical necessity. But many more local domains could likewise be considered modal. This thesis is a contribution towards characterizing this middle ground of medium-strength modality, which is graded and local. To show that such domain-relative and graded modality can be given a precise formulation, I develop a measure theoretic model of generality. This is meant mainly as a possibility proof, that we can at least sometimes make graded and domain-relative modal talk formally precise. Using this formal model as a backdrop, I argue that generality and contingency should be understood as modal concepts. Something is general to the extent that it applies to a large measure of a domain of possibilities, and contingent to the extent that it applies to a small measure. Chapter 3 ties this notion of generality to explanatory depth. This way of understanding modality also helps to clarify the relationship between laws and models (chapter 4). Laws, I claim, define domains of possibilities by drawing their limits. Since generality and contingency are best understood as domain-relative concepts, laws set the standard for what counts as general or contingent in a given context. Models operate to fill in the structure and content of domains of possibilities. Finally, medium-strength modal concepts help to clarify the difference between mechanistic and self-organized systems, which corresponds to the traditional distinction drawn between merely resultant and emergent properties. I argue that a system is mechanistic if its system-level regularities are the result of parts-level regularities (general features of its parts). Conversely, a system is self-organized if it displays system-level regularities, despite the contingent behavior of its parts. This analysis shows how we can be non-reductive physicalists.Ph.D