[Paper Review] Team semantics for interventionist counterfactuals and causal dependence

Fausto Barbero, Gabriel Sandu|arXiv (Cornell University)|Dec 22, 2017

Semantic Web and Ontologies31 references3 citations

TL;DR

This paper introduces causal team semantics, a generalized team semantics framework that integrates interventionist counterfactual reasoning and causal dependence within structural equation models. By extending team semantics with intervention operators and introducing formal entries and weaker truth values (falsifiability and admissibility), it provides a unified logical foundation for deterministic and probabilistic causal discourse, enabling formal analysis of direct and total causation, invariance, and counterfactual dependence.

ABSTRACT

We introduce a generalization of team semantics which provides a framework for manipulationist theories of causation based on structural equation models, such as Woodward's and Pearl's; our causal teams incorporate (partial or total) information about functional dependencies that are invariant under interventions. We give a unified treatment of observational and causal aspects of causal models by isolating two operators on causal teams which correspond, respectively, to conditioning and to interventionist counterfactual implication. The evaluation of counterfactuals may involve the production of partially determined teams. We suggest a way of dealing with such cases by 1) the introduction of formal entries in causal teams, and 2) the introduction of weaker truth values (falsifiability and admissibility), for which we suggest some plausible semantical clauses. We introduce formal languages for both deterministic and probabilistic causal discourse, and study in some detail their inferential aspects. Finally, we apply our framework to the analysis of direct and total causation, and other notions of dependence and invariance.

Motivation & Objective

• To develop a logical framework that captures interventionist counterfactual reasoning in causal models, extending traditional team semantics.
• To address the limitations of standard team semantics in representing causal relationships, such as functional dependencies invariant under interventions.
• To unify observational and causal reasoning by introducing operators for conditioning and intervention on teams.
• To handle nonparametric models and partially determined teams through formal entries and weakened truth values (falsifiability and admissibility).
• To provide logical languages for deterministic and probabilistic causal discourse, enabling formal analysis of causation, dependence, and invariance.

Proposed method

• Generalize team semantics to causal teams by enriching assignment sets with functional dependencies and intervention capabilities.
• Introduce two key operators: one for conditioning (observational inference) and one for intervention (counterfactual reasoning), formalized via team transformations.
• Define interventions on causal teams as modifications of variable assignments under structural equations, preserving invariance under intervention.
• Introduce formal entries to represent partially determined values in nonparametric models, allowing for logical evaluation of counterfactuals.
• Adopt weaker truth values—falsifiability and admissibility—for handling cases where standard truth values fail in nonparametric settings.
• Construct logical languages (deterministic and probabilistic) with dependence atoms, counterfactual conditionals, and invariance formulas, using team-based semantics.

Experimental results

Research questions

• RQ1 How can team semantics be extended to model interventionist counterfactuals in structural equation models?
• RQ2 What logical operators are needed to distinguish observational conditioning from causal intervention in team-based frameworks?
• RQ3 How can nonparametric models with partially determined teams be handled within a logical semantics framework?
• RQ4 Can notions of direct and total causation, as well as invariance, be formally expressed and analyzed in this extended semantics?
• RQ5 What is the inferential and definitional power of the proposed logical languages for causal discourse?

Key findings

• The framework successfully integrates interventionist counterfactual reasoning into team semantics by introducing formal intervention operators that modify causal teams consistently with structural equation models.
• The introduction of formal entries and weaker truth values (falsifiability and admissibility) enables logical evaluation of counterfactuals in nonparametric models where standard truth values fail.
• The paper demonstrates that invariant dependencies—those preserved under all interventions—can be formally defined using a scheme of formulas involving counterfactual conditionals.
• The logical framework supports a unified treatment of deterministic and probabilistic causal discourse, with a language that distinguishes pre- and post-intervention states.
• The framework reveals that certain causal notions, such as direct and total causation, can be expressed and analyzed through logical formulas over causal teams, providing a formal basis for manipulationist theories.
• The study identifies open challenges in extending the framework to nonrecursive models, unobserved variables, and non-surgical interventions, suggesting directions for future research in expressivity and proof systems.