Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Michael Weisberg


Philosophers and scientists have sought to draw methodological distinctions among different kinds of experiments, and between experimentation and other scientific methodologies. This dissertation focuses on two such cases: hypothesis-testing versus exploratory experiments, and experiment versus simulation. I draw on examples from experimental evolution--evolving organisms in a controlled laboratory setting to study evolution via natural selection in real time--to challenge the way we think about these distinctions. In the case of hypothesis-testing versus exploratory experiments, philosophers have distinguished these categories in terms of the role of theory in experiment. I discuss examples from experimental evolution which occupy the poorly characterized middle ground between the two categories. I argue that we should take more seriously the point that multiple theoretical backgrounds can come into play at multiple points in an experiment, and propose some new contributions toward clarifying the conceptual space of experimental inquiry. In the case of experiment versus simulation, people have attempted to clearly delineate cases of science into these two categories, and base judgments about their epistemic value on these categorizations. I discuss and reject two arguments for the epistemic superiority of experiments over simulations: (1) Experiments put scientists in a better position to make valid inferences about the natural world; (2) Experiments are a superior source of surprises or novel insights. Both of these claims are false as generalizations across science. Focusing on the experiment/simulation distinction as a basis for in-principle judgments about epistemic value focuses us on the wrong issues. This leaves us with a question: What should we focus on instead? I offer preliminary considerations for a framework for evaluating inferences from objects of study to targets of inquiry in the world, which departs from the problematic custom of basing such evaluations on questions like "Was it an experiment or a simulation?" This framework is based on the ideas of capturing relevant similarities while appropriately accounting for what researchers already know and what they are trying to learn by asking the scientific question at hand.