Comparing causes – an information-theoretic approach to specificity, proportionality and stability (original) (raw)
Related papers
Proportionality, Determinate Intervention Effects, and High-Level Causation
Erkenntnis, 2024
Stephen Yablo's notion of proportionality, despite controversies surrounding it, has played a significant role in philosophical discussions of mental causation and of high-level causation more generally. In particular, it is invoked in James Woodward's interventionist account of high-level causation and explanation, and is implicit in a novel approach to constructing variables for causal modeling in the machine learning literature, known as causal feature learning (CFL). In this article, we articulate an account of proportionality inspired by both Yablo's account of proportionality and the CFL account of variable construction. The resulting account has at least three merits. First, it illuminates an important feature of the notion of proportionality, when it is adapted to a probabilistic and interventionist framework. The feature is that at the center of the notion of proportionality lies the concept of "determinate intervention effects." Second, it makes manifest a virtue of (common types of) high-level causal/explanatory statements over low-level ones, when relevant intervention effects are determinate. Third, it overcomes a limitation of the CFL framework and thereby also addresses a challenge to interventionist accounts of high-level causation.
Quantifying proportionality and the limits of higher-level causation and explanation
2023
Supporters of the autonomy of higher-level causation (or explanation) often appeal to proportionality, arguing that higher-level causes are more proportional than their lower-level realizers. Recently, measures based on information theory and causal modelling have been proposed that allow one to shed new light on proportionality and the related notion of specificity. In this paper we apply ideas from this literature to the issue of higher vs. lower-level causation (and explanation). Surprisingly, proportionality turns out to be irrelevant for the question of whether higher-level causes (or explanations) can be autonomous; specificity is a much more informative notion for this purpose. Citation information: Gebharter, A., & Eronen, M. I. (2023). Quantifying proportionality and the limits of higher-level causation and explanation. British Journal for the Philosophy of Science, 74(3), 573-601. doi:10.1086/714818
Causal specificity, information flow, and causal independence (in prep.)
Causal specificity has been recently proposed to be measurable by the mutual information between interventions on a causal variable and observations of an effect variable. We show that this amounts to considering interventions as variables to be included in a modified causal graph. We compare this account with a related measure of causal effect stemming from computer sciences, that of information flow, which does not rely on reifying interventions. We argue that causal specificity measures potential control, while information flow measures actual explanatory power. We show how causal interactions can lead to situations where causation stricto sensu does not entail that the causal variable enable either control or explanation. This final point has significant implications for the popular 'manipulationist' account of causation.
Perspectives on Causal Specificity
2020
Causal specificity is a measure of how important a cause is relative to another. Waters (2007) has developed a theory of causation that deals with specificity. Weber (2006, 2017a, 2017b) has thoroughly criticized it. I defend Waters’s theory by showing that non-systematicity is unproblematic. I also argue that Weber’s desiderata for theories of causation are too restrictive and insensitive to developments in biological technology. I finally challenge the most fundamental assumption in the framework of causal specificity—that bijective functions are most specific— thus calling for its reassessment.
Causality, Confounding, and Control
2006
In a previous paper, Russo et al.(2006), causality is considered in the framework of structural models, ie statistical models characterized by parameters that are stable over a large class of interventions or of environmental changes and that take into account background and contextual knowledge. From this statistical viewpoint, causality is defined in terms of exogeneity in a structural model. This approach allows us to attain a concept of causality that is internal or relative to the structural model itself. Thus our knowledge of causal relations ...
Information Theoretic Causal Effect Quantification
Entropy, 2019
Modelling causal relationships has become popular across various disciplines. Most common frameworks for causality are the Pearlian causal directed acyclic graphs (DAGs) and the Neyman-Rubin potential outcome framework. In this paper, we propose an information theoretic framework for causal effect quantification. To this end, we formulate a two step causal deduction procedure in the Pearl and Rubin frameworks and introduce its equivalent which uses information theoretic terms only. The first step of the procedure consists of ensuring no confounding or finding an adjustment set with directed information. In the second step, the causal effect is quantified. We subsequently unify previous definitions of directed information present in the literature and clarify the confusion surrounding them. We also motivate using chain graphs for directed information in time series and extend our approach to chain graphs. The proposed approach serves as a translation between causality modelling and i...
Oxford University Press Oxford, 2011
This paper presents a general theory of causation based on the Structural Causal Model (SCM) described in (Pearl, 2000a). The theory subsumes and unifies current approaches to causation, including graphical, potential outcome, probabilistic, decision analytical, and structural equation models, and provides both a mathematical foundation and a friendly calculus for the analysis of causes and counterfactuals. In particular, the paper demonstrates how the theory engenders a coherent methodology for inferring (from a combination of data and assumptions) answers to three types of causal queries: (1) queries about the effects of potential interventions, (2) queries about probabilities of counterfactuals, and (3) queries about direct and indirect effects.
Top-Down Causation and Emergence, 2021
Two conceptual frameworks-in terms of phase space and in terms of structural equations-are sketched, in which downward causal influence of higher-level features on lower-level features is possible. The "Exclusion" principle, which is a crucial premise of the argument against the possibility of downward causation, is false in models constructed within both frameworks. Both frameworks can be supplemented with conceptual tools that make it possible to explain why downward causal influence is not only conceivable and compatible with the "Closure" principle, but also why it is often relevant to causally explain facts in terms of downward causal influence. It is briefly shown that 1) the analysis of downward causation in the two frameworks complements Bennett's (2003) analysis of overdetermination, 2) the analysis does not entail the failure of the "Closure" principle and 3) it does not require the postulate of synchronic downward causation.
Higher‐Level, Downward and Specific Causation
The Interventionist account of causation (Woodward 2003) seems to provide a rigorous framework for evaluating the possibility of downward causation. However, it has turned out 1) that only a modified version (Woodward 2014) of interventionism can be applied to situations of apparent downward causation and that 2) this model, though compatible with downward causation, makes it in principle impossible to find empirical support for downward causation (Baumgartner 2013). In this paper I show in which sense downward causation can be justified by using more fine-‐grained notions of causation, such as stable, proportional and specific causation (Woodward 2010). In particular, the intervention on a higher-‐level variable H(t) with respect to a lower-‐level variable P(t*) (where t* is later than t) may be more proportional compared to the parallel intervention on lower-‐level variable R(t) w.r.t. to P(t*), if R(t) is too determinate with respect to P(t*), i.e. if an intervention on R(t) is not necessary for manipulating P(t*).