Ontology, Complexity, and Compositionality
metaphysical curves. That explanatory carving itself, however, confers metaphysical- ity: the answer to the question, “What are the high-level laws?” is “Whatever plays the law role in high-level explanation.” Likewise, the answer to the question, “What are the high-level entities ?” is “Whatever plays the entity role (or roles) in high-level explanation.”
I have given a sketch of the resulting metaphysics of special science laws, deterministic and statistical, elsewhere (Strevens 2008a, §§7.6 and 10.5). Here, I hope to do the same for a certain specific class of physical probabilities. Although these probabilities are not natural from a naive or a priori perspective, they play an important explanatory role in science. In virtue of their power to illuminate the world, I propose, they deserve a place in the world.
I offer no argument for this proposal. Rather, I hope that the picture of the high- level world implied by the view is sufficiently attractive to command respect and even allegiance on its own merits.
My method is straightforward. The inventories of entities compiled by the various sciences for explanatory purposes, I simply take for granted as genuine ontologies. What the sciences say exists—provided that it has a genuine theoretical function—does exist. You can see for yourself whether or not you like the consequences.
-  A certain approach to the metaphysics ofscience, or to scientific metaphysics,is demonstrated in this piece about compositional theories of complex systems. Inthese opening remarks I describe the approach; my aim in the rest of the paper is notto argue for so much as to illustrate what I have described. My specialty is the metaphysics of the higher level—the way that the fundamental constituents of the material world and the fundamental laws come togetherto compose or otherwise determine the existence of high-1 evel causal relations,probability distributions, laws, entities, properties, and other manifestations ofnon-fundamental being.
-  take the basic physical stuff as given, then—whether it is provided by a Humean(Lewis 1994), a realist (Armstrong 1997), a scientific essentialist (Ellis 2001), or someother metaphysical Bauplan. My question is what else exists above and beyond thefundamental level, in virtue of the fundamental level, and why. Do high-level lawsexist in virtue of their entailment or probabilification by the fundamental-level laws(Loewer 2008)? Or in virtue of their being systematizers par excellence of singularhigh-level facts (Cohen and Callender 2009)? Or in virtue of a special mix of physical necessity and singular fact that gives them the ability to support a wide range ofcounterfactuals (Strevens 2008b)? These are my kinds of problems. The key to solving such problems is, I hold, scientific explanation: the high-levellaws, probabilities, and entities are endowed with ontological significance by theirrole in making sense of the world. Science carves them out of the fundamental-levelsubstrate to take advantage of their explanatory power, rather than to admire their