The logic of causal methods in social science

Two kinds of causal inference rules which are widely used by social scientists are investigated. Two conceptions of causation also widely used are explicated — the INUS and probabilistic conceptions of causation. It is shown that the causal inference rules which link correlation, a kind of partial correlation, and a conception of causation areinvalid. It is concluded anew methodology is required for causal inference.     

Correlation,partial correlation,and causation

Philosophers and scientists have maintained that causation, correlation, and “partial correlation” are essentially related. These views give rise to various rules of causal inference. This essay considers the claims of several philosophers and social scientists for causal systems with dichotomous variables. In section 2 important commonalities and differences are explicated among four major conceptions of correlation. In section 3 it is argued that whether correlation can serve as a measure of A's causal influence on B depends upon the conception of causation being used and upon certain background assumptions. In section 4 five major kinds of “partial correlation” are explicated, and some of the important relations are established among two conceptions of “partial correlation”, the conception of “screening off”, the conception of “partitioning”, and the measures of causal influence which have been suggested by advocates of path analysis or structural equation methods. In section 5 it is argued that whether any of these five conceptions of “partial correlation” can serve as a measure of causal influence depends upon the conception of causation being used and upon certain background assumptions. The important conclusion is that each of the approaches (considered here) to causal inference for causal systems with dichotomous variables stands in need of important qualifications and revisions if they are to be justified.     

Causal reasoning and backtracking

I argue that one central aspect of the epistemology of causation, the use of causes as evidence for their effects, is largely independent of the metaphysics of causation. In particular, I use the formalism of Bayesian causal graphs to factor the incremental evidential impact of a cause for its effect into a direct cause-to-effect component and a backtracking component. While the “backtracking” evidence that causes provide about earlier events often obscures things, once we our restrict attention to the cause-to-effect component it is true to say promoting (inhibiting) causes raise (lower) the probabilities of their effects. This factoring assumes the same form whether causation is given an interventionist, counterfactual or probabilistic interpretation. Whether we think about causation in terms of interventions and causal graphs, counterfactuals and imaging functions, or probability raising against the background of causally homogenous partitions, if we describe the essential features of a situation correctly then the incremental evidence that a cause provides for its effect in virtue of being its cause will be the same.     

No End in Sight: Causal Loops in Philosophy, Physics and Fiction

There have been many objections to the possibility oftime travel. But all the truly interesting ones concern the possibility of reversecausation. What is objectionable about reverse causation? I diagnose that the trulyinteresting objections are to a further possibility: that of causal loops. I raisedoubts about whether there must be causal loops if reverse causation obtains; but devote themajority of the paper to describing, and dispelling concerns about, various kinds ofcausal loop. In short, I argue that they are neither logically nor physically impossible.The only possibly objectionable feature that all causal loops share is that coincidenceis required to explain them. Just how coincidental a loop will be varies: some arereally quite ordinary, and some are incredibly unlikely. I end by speculating thatthe tendency amongst physicists to avoid discussion of causal loops involving intentionalaction may have been unfortunate, since intentional action is an excellent way tonon-mysteriously bring about what otherwise would have been an unlikely coincidence. Hencecausal loops may be more likely in a world with beings like us, than in one without.     

Computation and Causation

The computer's effect on our understanding of causation has been enormous. By the mid-1980s, philosophical and social-scientific work on the topic had left us with (1) no reasonable reductive account of causation and (2) a class of statistical causal models tied to linear regression. At this time, computer scientists were attacking the problem of equipping robots with models of the external that included probabilistic portrayals of uncertainty. To solve the problem of efficiently storing such knowledge, they introduced Bayes Networks and directed graphs. By attaching a causal interpretation to Bayes Networks, the philosophy of causation changed dramatically. We are now able to be extremely general about how causal structure connects to data, and systematic about when causal structures are empirically indistinguishable. In this essay I try to motivate and describe this synthesis.     

Activity-Based Accounts of Mechanism and the Threat of Polygenic Effects

Accounts of ontic explanation have often been devised so as to provide an understanding of mechanism and of causation. Ontic accounts differ quite radically in their ontologies, and one of the latest additions to this tradition proposed by Peter Machamer, Lindley Darden and Carl Craver reintroduces the concept of activity. In this paper I ask whether this influential and activity-based account of mechanisms is viable as an ontic account. I focus on polygenic scenarios—scenarios in which the causal truths depend on more than one cause. The importance of polygenic causation was noticed early on by Mill (1893). It has since been shown to be a problem for both causal-law approaches to causation (Cartwright 1983) and accounts of causation cast in terms of capacities (Dupré 1993; Glennan 1997, pp. 605–626). However, whereas mechanistic accounts seem to be attractive precisely because they promise to handle complicated causal scenarios, polygenic causation needs to be examined more thoroughly in the emerging literature on activity-based mechanisms. The activity-based account proposed in Machamer et al. (2000, pp. 1–25) is problematic as an ontic account, I will argue. It seems necessary to ask, of any ontic account, how well it performs in causal situations where—at the explanandum level of mechanism—no activity occurs. In addition, it should be asked how well the activity-based account performs in situations where there are too few activities around to match the polygenic causal origin of the explanandum. The first situation presents an explanandum-problem and the second situation presents an explanans-problem—I will argue—both of which threaten activity-based frameworks.     

A probabilistic theory of second order causation

Larry Wright and others have advanced causal accounts of functional explanation, designed to alleviate fears about the legitimacy of such explanations. These analyses take functional explanations to describe second order causal relations. These second order relations are conceptually puzzling. I present an account of second order causation from within the framework of Eells' probabilistic theory of causation; the account makes use of the population-relativity of causation that is built into this theory.     

A Deliberative Approach to Causation

Fundamental physics makes no clear use of causal notions; it uses laws that operate in relevant respects in both temporal directions and that relate whole systems across times. But by relating causation to evidence, we can explain how causation fits in to a physical picture of the world and explain its temporal asymmetry. This paper takes up a deliberative approach to causation, according to which causal relations correspond to the evidential relations we need when we decide on one thing in order to achieve another. Tamsin's taking her umbrella is a cause of her staying dry, for example, if and only if her deciding to take her umbrella for the sake of staying dry is adequate grounds for believing she'll stay dry. This correspondence explains why causation matters: knowledge of causal structure helps us make decisions that are evidence of outcomes we seek. The account also explains why we can control the future and not the past, and why causes come before their effects. When agents properly deliberate, their decisions can never count as evidence for any outcomes they may seek in the past. From this it follows that causal relations don't run backwards. This deliberative asymmetry is itself traced back to asymmetries of evidence and entropy, providing a new way of deriving causal asymmetry from temporally symmetric laws.     

Is everyday causation deterministic or probabilistic?

One view of causation is deterministic: A causes B means that whenever A occurs, B occurs. An alternative view is that causation is probabilistic: the assertion means that given A, the probability of B is greater than some criterion, such as the probability of B given not-A. Evidence about the induction of causal relations cannot readily decide between these alternative accounts, and so we examined how people refute causal assertions. In four experiments most participants judged that a single counterexample of A and not-B refuted assertions of the form, A causes B. And, as a deterministic theory based on mental models predicted, participants were more likely to request multiple refutations for assertions of the form, A enables B. Similarly, refutations of the form not-A and B were more frequent for enabling than causal assertions. Causation in daily life seems to be a deterministic concept.     

Interventionist counterfactuals

A number of recent authors (Galles and Pearl, Found Sci 3 (1):151?C182, 1998; Hiddleston, No?s 39 (4):232?C257, 2005; Halpern, J Artif Intell Res 12:317?C337, 2000) advocate a causal modeling semantics for counterfactuals. But the precise logical significance of the causal modeling semantics remains murky. Particularly important, yet particularly under-explored, is its relationship to the similarity-based semantics for counterfactuals developed by Lewis (Counterfactuals. Harvard University Press, 1973b). The causal modeling semantics is both an account of the truth conditions of counterfactuals, and an account of which inferences involving counterfactuals are valid. As an account of truth conditions, it is incomplete. While Lewis??s similarity semantics lets us evaluate counterfactuals with arbitrarily complex antecedents and consequents, the causal modeling semantics makes it hard to ascertain the truth conditions of all but a highly restricted class of counterfactuals. I explain how to extend the causal modeling language to encompass a wider range of sentences, and provide a sound and complete axiomatization for the extended language. Extending the truth conditions for counterfactuals has serious consequences concerning valid inference. The extended language is unlike any logic of Lewis??s: modus ponens is invalid, and classical logical equivalents cannot be freely substituted in the antecedents of conditionals.     

Causal Reasoning and Clinical Practice: Challenges from Molecular Biology

Not only has the philosophical debate on causation been gaining ground in the last few decades, but it has also increasingly addressed the sciences. The biomedical sciences are among the most prominent fields that have been considered, with a number of works tackling the understanding of the notion of cause, the assessment of genuinely causal relations and the use of causal knowledge in applied contexts. Far from denying the merits of the debate on causation and the major theories it comprises, this paper is meant as a stimulus for theorists of causation in the philosophy of biomedicine, with a focus on clinical matters. Without aiming at putting forward an original theory of causation and starting from the narration of two actual but paradigmatic cases at the joints between biomedical research and clinical practice, we want to point out that some pathological situations addressed by molecular medicine actually prove resistant to (at least) some of our major epistemological accounts of causal explanation. Given this scenario, which is very frequent in our hospitals, our analysis aims to provide a stimulus for the debate among theorists of causation in biomedicine interested in real cases in science in practice. We believe that this might in turn encourage some more general rethinking of the complex intertwinement of science, philosophy of science and ethics, as well as of the role of philosophy of science for clinical medicine itself.

     

Can determinable properties earn their keep?

Sydney Shoemaker’s ‘Subset Account’ offers a new take on determinable properties and the realization relation as well as a defense of non-reductive physicalism from the problem of mental causation. At the heart of this account are the claims that (1) mental properties are determinable properties and (2) the causal powers that individuate a determinable property are a proper subset of the causal powers that individuate the determinates of that property. The second claim, however, has led to the accusation that the effects caused by the instantiation of a determinable property will also be caused by the instantiation of the determinates of that property—so instead of solving the problem of mental causation, the Subset Account ends up guaranteeing that the effects of mental properties (and all other types of determinable property) will be causally overdetermined! In this paper, I explore this objection. I argue that both sides in this debate have failed to engage the question at the heart of the objection: Given that both a determinable property and its determinates have the power to cause some effect (E), does it follow that both will actually cause E when the relevant conditions obtain? To make genuine progress towards answering this question, we need to take a serious look at the metaphysics of causation. With the debate properly reframed and issues about the metaphysics of causation front and center, I explore the question of whether the Subset Account is doomed to result in problematic causal overdetermination.     

Causal reasoning and the diagnostic process

Background: Causal reasoning as a way to make a diagnosis seems convincing. Modern medicine depends on the search for causes of disease and it seems fair to assert that such knowledge is employed in diagnosis. Causal reasoning as it has been presented neglects to some extent the conception of multifactorial disease causes. Goal: The purpose of this paper is to analyze aspects of causation relevant for discussing causal reasoning in a diagnostic context. Procedures: The analysis will discuss different conceptions of causal reasoning in medical diagnosis, discriminating primarily between narrow causal diagnosis and more thorough causal explanation. The theory of causes as non-redundant factors in effective causal complexes is used as an analytical background. Causal explanations are performed according to different causal models. Such models of diagnosis are assumptions concerning structure and mechanisms, which cannot be directly or immediately observed. Conceptions and results of causal search strategies differ, according to the focus of the searcher. Causal reasoning is also seen in diagnosis in a more extensive meaning: the pin-pointing of factors responsible for the condition of the patient at any time during the course of disease. Conclusion: Causal reasoning and diagnosis go well in hand, especially if both concepts are widened. The theory of causes as non-redundant components in effective causal complexes, modulated by what is referred to as the stop problem and causal fields, is valuable for explaining the many aspects of causal reasoning in medical diagnosis.     

Seeing versus doing: two modes of accessing causal knowledge

The ability to derive predictions for the outcomes of potential actions from observational data is one of the hallmarks of true causal reasoning. We present four learning experiments with deterministic and probabilistic data showing that people indeed make different predictions from causal models, whose parameters were learned in a purely observational learning phase, depending on whether learners believe that an event within the model has been merely observed ("seeing") or was actively manipulated ("doing"). The predictions reflect sensitivity both to the structure of the causal models and to the size of their parameters. This competency is remarkable because the predictions for potential interventions were very different from the patterns that had actually been observed. Whereas associative and probabilistic theories fail, recent developments of causal Bayes net theories provide tools for modeling this competency.     

A generalized probabilistic theory of causal relevance

I advance a new theory of causal relevance, according to which causal claims convey information about conditional probability functions. This theory is motivated by the problem of disjunctive factors, which haunts existing probabilistic theories of causation. After some introductory remarks, I present in Section 3 a sketch of Eells's (1991) probabilistic theory of causation, which provides the framework for much of the discussion. Section 4 explains how the problem of disjunctive factors arises within this framework. After rejecting three proposed solutions, I offer in Section 6 a new approach to causation that avoids the problem. Decision-theoretic considerations also support the new approach. Section 8 develops the consequences of the new theory for causal explanation. The resulting theory of causal explanation incorporates the new insights while respecting important work on scientific explanation by Salmon (1971), Railton (1981), and Humphreys (1989). My conclusions are enumerated in Section 9.I would like to thank Nuel Belnap, John Earman, Richard Gale, Paul Humphreys, Satish Iyengar, Wes Salmon, and two anonymous referees for comments and discussion. I am also indebted to the members of an audience at the Center for Philosophy of Science at the University of Pittsburgh, where some of the ideas contained in this paper were presented.     

How Probabilistic Causation Can Account for the Use of Mechanistic Evidence

In a recent article in this journal, Federica Russo and Jon Williamson argue that an analysis of causality in terms of probabilistic relationships does not do justice to the use of mechanistic evidence to support causal claims. I will present Ronald Giere’s theory of probabilistic causation, and show that it can account for the use of mechanistic evidence (both in the health sciences—on which Russo and Williamson focus—and elsewhere). I also review some other probabilistic theories of causation (of Suppes, Eells, and Humphreys) and show that they cannot account for the use of mechanistic evidence. I argue that these theories are also inferior to Giere’s theory in other respects.     

How Could There Be True Causal Claims Without There Being Special Causal Facts in the World?

Some philosophers of physics recently expressed their skepticism about causation (Norton 2003b, 2007). However, this is not new. The view that causation does not refer to any ontological category perhaps can be attributed to Hume, Kant and Russell. On the other hand, some philosophers (Wesley Salmon and Phil Dowe) view causation as a physical process and some others (Cartwright) view causation as making claims about capacities possessed by objects. The issue about the ontological status of causal claims involves issues concerning the ontological status of capacity, modality and dispositional claims. In this paper, my goal is to show that without engaging metaphysical debates about the ontological status of causal claims, it can be shown that we can objectively assign truth values to these statements. I argue that for causal claims to be objective we don't need to postulate the existence of special facts (specific to causal claims) in addition to ordinary physical facts described by physical theories. This, I think, is enough to justify the usefulness of this concept in certain branches (may be all) of science. Once this is achieved, there is no need to engage in unnecessary metaphysical debates. So, even if advanced physical theories don't mention this notion, causal reasoning can still be important in understanding the world not in the sense that science discovers special ontological category called causation but in the sense that we come to know certain facts about the world.     

Causal Relata: Tokens, Types, or Variables?

The literature on causation distinguishes between causal claims relating properties or types and causal claims relating individuals or tokens. Many authors maintain that corresponding to these two kinds of causal claims are two different kinds of causal relations. Whether to regard causal relations among variables as yet another variety of causation is also controversial. This essay maintains that causal relations obtain among tokens and that type causal claims are generalizations concerning causal relations among these tokens.     

World views and the concept of development: A reply to Lerner and Kauffman

I reply here to the argument of R. M. Lerner and M. B. Kauffman (1985, Developmental Review, 5, 309–333) that an adequate concept of human development is incompatible with a mechanist “world view” but rests instead on a principled integration of contextualist and organicist “world views.” I review how each of these metatheoretical positions is described by the philosopher who proposed them and conclude that the version of contextualism and organicism presented by Lerner and Kauffman is so diluted as to lose the essence of their original meaning. In consequence, the concept of development they propose, which includes the notion of integrative levels, causal variables that interact differently at different times in the course of ontogeny, and probabilistic outcomes is more compatible with the mechanistic metatheory they eschew than with the contextualist and organismic ones they ostensibly espouse.