Modal logic has been applied in many different areas, as reasoning about time, knowledge and belief, necessity and possibility, to mention only some examples. In the present paper, an attempt is made to use modal logic to account for the semantics of theoretical sentences in scientific language. Theoretical sentences have been studied extensively since the work of Ramsey and Carnap. The present attempt at a modal analysis is motivated by there being several intended interpretations of the theoretical terms once these terms are introduced through the axioms of a theory.

What is a theoretical term? This question can be answered in at least two different ways. First, a theoretical term is simply a non-observational term. Second, a theoretical term is one whose meaning depends on the axioms of a scientific theory. According to the first explanation, a theoretical term cannot be applied using just unaided perception, without drawing inferences. This explanation defines the notion of theoreticity merely as the absence of observability. The second explanation, by contrast, has the virtue of giving a positive characterization of the notion of theoreticity. Both explanations stand in the need of further elaboration. If we characterize theoretical terms by non-observability, we need to explain what an observational term is. There is no consensus in the literature as to whether and, if so, to what extent it is feasible to draw the theory-observation distinction. On the one hand, critics of the theory-observation distinction have often attacked only weak proposals of how to draw the distinction in question. On the other hand, the extreme skepticism by Thomas S. Kuhn, Paul K. Feyerabend, and Norwood R. Hanson concerning the distinction is increasingly losing consensus among contemporary philosophers of science. This is evidenced, for example, by attempts at exploiting the formal semantics of theoretical terms in one version of structural realism. If we explain the notion of a theoretical term by way of semantic dependency upon a scientific theory, we need to give an account of this semantic relation. How does a theory determine the meaning of a theoretical term? What, if any, are the differences between theoretical terms and defined terms? How can we distinguish, in a sensible way, between the synthetic assertions of a scientific theory about the world and meaning postulates determining the meaning of theoretical terms? Various formal semantics of theoretical terms have been devised in order to answer these questions. Notably, the idea that the meaning of a theoretical term is determined by a scientific theory, or a set of such theories, has already been expressed by Pierre Duhem and Henrie Poincaré. The theory-observation distinction can be applied to syntactic and semantic entities. Thus, we can speak of theoretical terms and theoretical concepts. Moreover, we can speak of theoretical entities, in the sense of specific objects that are the referents of theoretical concepts. Philosophical research on theoreticity concerns syntactic aspects inasmuch as semantic aspects of theoreticity.

We aim to devise a Ramsey test analysis of actual causation. Our method is to define a strengthened Ramsey test for causal models. Unlike the accounts of Halpern and Pearl ([2005]) and Halpern ([2015]), the resulting analysis deals satisfactorily with both overdetermination and conjunctive scenarios. 1Introduction2An Extension of Causal Model Semantics 2.1Halpern and Pearl’s causal model semantics2.2Agnostic models3A Strengthened Ramsey Test for Causal Models 3.1The Ramsey test and causal models3.2A strengthened Ramsey test for causal models3.3A Ramsey test definition of actual causation3.4Minimality4Applying the Definition of Actual Causation 4.1Overdetermination and conjunctive scenarios4.2Preemption4.3Conjunctive and disjunctive scenarios combined5Comparison to the Halpern–Pearl Definitions 5.1Conjunctive scenarios5.2Preemption5.3Symmetric overdetermination5.4Summary5.5A note on causal models6Conclusion

We put forth an analysis of causation. The analysis centers on the notion of a causal model that provides only partial information as to which events occur, but complete information about the dependences between the events. The basic idea is this: an event causes another just in case there is a causal model that is uninformative on both events and in which the first event makes a difference as to the occurrence of the other. We show that our analysis captures more causal scenarios than the other counterfactual accounts to date.

No abstract