Abstract

Ecological populations and communities are highly complex systems and our ability to understand these systems are limited. Thus, the mathematical models used to represent these systems are often highly idealized. In this dissertation. I examine in the context of theoretical ecology what mathematical models are, how these models are evaluated, and how models can explain the dynamics of these systems given their complexity and the idealizations introduced. ;In the first section, I argue that the semantic view of theories and models in its conservative and liberal forms is unacceptable. Models are neither related to the world through isomorphisms nor similarity as the semantic view requires. I argue that models are abstract, idealized structures which are related to systems by accurately representing them. ;In the second section, I consider how models are evaluated. First, I consider how different desiderata of model building can come into conflict. Richard Levins has argued that there is a necessary trade-off between the generality, realism, and precision of biological models. I argue that Levins is correct and respond to worries raised by Elliott Sober and Steven Orzack . Second, I examine Elisabeth Lloyd's account of confirmation of models in biology. I show that her account neglects how idealizations are discharged and I supply some of the details needed. Third, I argue that models are evaluated relative to various goals and can successfully achieve those goals even when they are representationally inaccurate or difficult to test. I provided an account of what the goals of model building in theoretical ecology are and show how they are achieved. ;In Section Three, I consider how idealized models can be explanatory. Many philosophers claim that a theory or model explains an event or regularity only if it is true . Yet, idealized models sometimes explain events and regularities. I argue that the notion of accurately representing phenomena is what matters and not truth. Hence, models are explanatory only if they accurately represent ecological systems in certain respects and in certain degrees