In this contribution, I comment on Raffaella Campaner’s defense of explanatory pluralism in psychiatry (in this volume). In her paper, Campaner focuses primarily on explanatory pluralism in contrast to explanatory reductionism. Furthermore, she distinguishes between pluralists who consider pluralism to be a temporary state on the one hand and pluralists who consider it to be a persisting state on the other hand. I suggest that it would be helpful to distinguish more than those two versions of pluralism – different understandings (...) of explanatory pluralism both within philosophy of science and psychiatry – namely moderate/temporary pluralism, anything goes pluralism, isolationist pluralism, integrative pluralism and interactive pluralism. Next, I discuss the pros and cons of these different understandings of explanatory pluralism. Finally, I raise the question of how to implement or operationalize explanatory pluralism in scientific practice; how to structure the “genuine dialogue” or shape “the pluralistic attitude” Campaner is referring to. As tentative answers, I explore a question-based framework for explanatory pluralism as well as social-epistemological procedures for interaction among competing approaches and explanations. (shrink)

This survey presents some of the main principles involved in discovering causal relations. They belong to a large array of possible assumptions and conditions about causal relations, whose various combinations limit the possibilities of acquiring causal knowledge in different ways. How much and in what detail the causal structure can be discovered from what kinds of data depends on the particular set of assumptions one is able to make. The assumptions considered here provide a starting point to explore further the (...) foundations of causal discovery procedures, and how they can be improved. (shrink)

The literature on causal discovery has focused on interventions that involve randomly assigning values to a single variable. But such a randomized intervention is not the only possibility, nor is it always optimal. In some cases it is impossible or it would be unethical to perform such an intervention. We provide an account of ‘hard' and ‘soft' interventions and discuss what they can contribute to causal discovery. We also describe how the choice of the optimal intervention(s) depends heavily on the (...) particular experimental setup and the assumptions that can be made. ‡The first author is funded by the Causal Learning Collaborative Initiative supported by the James S. McDonnell Foundation. Many aspects of this paper were inspired by discussions with members of the collaborative. †To contact the authors, please write to: Department of Philosophy, Carnegie Mellon University, Pittsburgh, PA 15213; e-mail: [email protected] and [email protected]. (shrink)

Simpson’s paradox (SP) is a statistical phenomenon where the association between two variables reverses, disappears, or emerges, after conditioning on a third variable. It has been proposed (by, e.g., Judea Pearl) that SP should be analyzed using the framework of graphical causal models (i.e., causal DAGs) in which SP is diagnosed as a symptom of confounding bias. This paper contends that this confounding-based analysis cannot fully capture SP: there are cases of SP that cannot be explained away in terms of (...) confounding. Previous works have argued that some cases of SP do not require causal analysis at all. Despite being a logically valid counterexample, we argue that this type of cases poses only a limited challenge to Pearl’s analysis of SP. In our view, a more powerful challenge to Pearl comes from cases of SP that do require causal analysis but can arise without confounding. We demonstrate with examples that accidental associations due to genetic drift, the use of inappropriate aggregate variables as causes, and interactions between units (i.e., inter-unit causation) can all give rise to SP of this type. The discussion is also extended to the amalgamation paradox (of which SP is a special form) which can occur due to the use of non-collapsible association measures, in the absence of confounding. (shrink)

Causation is a macroscopic phenomenon. The temporal asymmetry displayed by causation must somehow emerge along with other asymmetric macroscopic phenomena like entropy increase and the arrow of radiation. I shall approach this issue by considering ‘causal inference’ techniques that allow causal relations to be inferred from sets of observed correlations. I shall show that these techniques are best explained by a reduction of causation to structures of equations with probabilistically independent exogenous terms. This exogenous probabilistic independence imposes a recursive order (...) on these equations and a consequent distinction between dependent and independent variables that lines up with the temporal asymmetry of causation. (shrink)

This paper surveys and critically assesses existing theories of evidence with respect to four desiderata. A good theory of evidence should be both a theory of evidential support (i.e., be informative about what kinds of facts speak in favour of a hypothesis), and of warrant (i.e., be informative about how strongly a given set of facts speaks in favour of the hypothesis), it should apply to the non-ideal cases in which scientists typically find themselves, and it should be ‘descriptively adequate’, (...) i.e., able to adequately represent typical episodes of evidentiary reasoning. The theories surveyed here—Bayesianism, hypotheticodeductivism,satisfaction theories, error statistics as well as Achinstein’s and Cartwright’s theories—are all found wanting in important respects. I finally argue that a deficiency all these theories have in common is a neglect or underplaying of the epistemic context in which the episode of evidentiary reasoning takes place.Este artículo describe y valora críticamente diversas teorías de la evidencia en relación a cuatro desiderata. Una buena teoría de la evidencia debería ser tanto una teoría sobre el apoyo evidencial [evidential support] (ser informativa sobre qué tipos de hechos hablan a favor de la hipótesis) como sobre la justificación [warrant]; debería aplicarse en las situaciones no ideales en las que normalmente se encuentran los científicos; y debería ser ‘descriptivamente adecuada’, esto es, capaz de representar correctamente episodios típicos de razonamiento evidencial. Las teorías aquí revisadas—bayesianismo, hipotético-deductivismo, teorías de la satisfacibilidad, la estadística del error, así como las propuestas de Achinstein y Cartwright—se consideran deficientes en aspectos básicos. Argumentaré que un defecto común en todas ellas es que olvidan, o minusvaloran, el contexto epistémico en el que el episodio de razonamiento evidencial tiene lugar. (shrink)

Philosophers of science have often favoured reductive approaches to how-possibly explanation. This article identifies three alternative conceptions making how-possibly explanation an interesting phenomenon in its own right. The first variety approaches “how possibly X?” by showing that X is not epistemically impossible. This can sometimes be achieved by removing misunderstandings concerning the implications of one’s current belief system but involves characteristically a modification of this belief system so that acceptance of X does not result in contradiction. The second variety offers (...) a potential how-explanation of X. It is usually followed by a range of further potential how-explanations of the same phenomenon. In recent literature the factual claims implied by the second variety have been downplayed whereas the heuristic role of mapping the space of conceptual possibilities has been emphasized. I will focus especially on this truth-bracketing sense of potentiality when looking closer at the second variety in the paper. The third variety has attracted less interest. It presents a partial how-explanation of X. Typically it aims to establish the existence of a mechanism by which X could be and was generated. The third conception stands out as the natural alternative for the advocate of ontic how-possibly explanations. This article transfers Salmon’s view that explanation-concepts can be broadly divided into epistemic, modal, and ontic to the context of how-possibly explanations. Moreover, it is argued that each of the three above-mentioned varieties of how-possibly explanation occurs in science. To recognize this may be especially relevant for philosophers. We are often misled by the promises of various why-explanation accounts, and seem to have forgotten nearly everything about the diversity of how-possibly explanations. (shrink)

Two approaches to evidential reasoning compete in the biomedical and social sciences: the experimental and the pragmatist. Whereas experimentalism has received considerable philosophical analysis and support since the times of Bacon and Mill, pragmatism about evidence has been neither articulated nor defended. The overall aim is to fill this gap and develop a theory that articulates the latter. The main ideas of the theory will be illustrated and supported by a case study on the smoking/lung cancer controversy in the 1950s.

Modern empirical macroeconomic models, known as structural autoregressions (SVARs) are dynamic models that typically claim to represent a causal order among contemporaneously valued variables and to merely represent non-structural (reduced-form) co-occurence between lagged variables and contemporaneous variables. The strategy is held to meet the minimal requirements for identifying the residual errors in particular equations in the model with independent, though otherwise not directly observable, exogenous causes (“shocks”) that ultimately account for change in the model. In nonstationary models, such shocks accumulate (...) so that variables have discernible trends. Econometricians have conceived of variables that trend in sympathy with each other (so-called “cointegrated variables”) as sharing one or more of these unobserved trends as a common cause. It is possible for estimates of the values of both the otherwise unobservable individual shocks and the otherwise unobservable common trends to be backed-out of cointegrated systems of equations. The issue addressed in this paper is whether and in what circumstances these values can be regarded as observations of real entities rather than merely artifacts of the representation of variables in the model. The issue is related, on the one hand, to practical methodological problems in the use of SVARs for policy analysis—e.g., does it make sense to estimate of shocks or trends in one model and then use them as measures of variables in a conceptually distinct model? The issue is also related to debates in the philosophical analysis of causation—particularly, whether we are entitled, as assumed by the developers of Bayes-net approaches, to rely on the causal Markov condition (a generalization of Reichenbach’s common-cause condition) or whether cointegration generates a practical example of Nancy Cartwright’s “byproducts” objection to the causal Markov condition. (shrink)

In this paper I assess the adequacy of no-conspiracy conditions employed in the usual derivations of the Bell inequality in the context of EPR correlations. First, I look at the EPR correlations from a purely phenomenological point of view and claim that common cause explanations of these cannot be ruled out. I argue that an appropriate common cause explanation requires that no-conspiracy conditions are re-interpreted as mere common cause-measurement independence conditions. In the right circumstances then, violations of measurement independence need (...) not entail any kind of conspiracy (nor backwards in time causation). To the contrary, if measurement operations in the EPR context are taken to be causally relevant in a specific way to the experiment outcomes, their explicit causal role provides the grounds for a common cause explanation of the corresponding correlations. (shrink)

Nancy Cartwright's recent criticisms of efforts and methods to obtain causal information from sample data using automated search are considered. In addition to reviewing that effort, I argue that almost all of her criticisms are false and rest on misreading, overgeneralization, or neglect of the relevant literature.

This paper reviews all major theories of evidence such as the Bayesian theory, hypothetico-deductivism, satisfaction theories, error-statistics, Achinstein's explanationist theory and Cartwright's argument theory. All these theories fail to take adequate account of the context in which a hypothesis is established and used. It is argued that the context of an inquiry determines important facts about what evidence is, and how much and what kind has to be collected to establish a hypothesis for a given purpose.

A probability space is common cause closed if it contains a Reichenbachian common cause of every correlation in it and common cause incomplete otherwise. It is shown that a probability space is common cause incomplete if and only if it contains more than one atom and that every space is common cause completable. The implications of these results for Reichenbach's Common Cause Principle are discussed, and it is argued that the principle is only falsifiable if conditions on the common cause (...) are imposed that go beyond the requirements formulated by Reichenbach in the definition of common cause. (shrink)

Reichenbach’s ‘principle of the common cause’ is a foundational assumption of some important recent contributions to quantitative social science methodology but no similar principle appears in econometrics. Angrist et al. has argued that the principle is necessary for instrumental variables methods in econometrics, and Angrist Krueger builds a framework using it that he proposes as a means of resolving an important methodological dispute among econometricians. Through analysis of instrumental variables methods and the issue of multicollinearity, we aim to show that (...) the relationship of the principle to econometric methods is more nuanced than implied by previous work but nevertheless may make a valuable contribution to the coherence and validity of existing methods. (shrink)

It is still a controversial issue whether Reichenbach’s Principle of the Common Cause (RPCC) is a sound method for causal inference. In fact, the status of the principle has been a subject of intense philosophical debate. An extensive literature has been thus generated both with arguments in favor and against the adequacy of the principle. A remarkable argument against the principle, first proposed by Elliott Sober (Sober, 1987, 2001), consists on a counterexample which involves corelations between bread prices in Britain (...) and sea levels in Venice. The aim of this paper is to put into perspective criticisms to RPCC of the kind of Sober’s in the light of recent formal results regarding the so-called extendability and common cause completability. (shrink)

A graphical model is a graph that represents a set of conditional independence relations among the vertices (random variables). The graph is often given a causal interpretation as well. I describe how graphical causal models can be used in an algorithm for constructing partial information about causal graphs from observational data that is reliable in the large sample limit, even when some of the variables in the causal graph are unmeasured. I also describe an algorithm for estimating from observational data (...) (in some cases) the total effect of a given variable on a second variable, and theoretical insights into fundamental limitations on the possibility of certain causal inferences by any algorithm whatsoever, and regardless of sample size. (shrink)

In a recent article, Elliot Sober responds to challenges to a counter-example that he posed some years earlier to the Principle of the Common Cause (PCC). I agree that Sober has indeed produced a genuine counter-example to the PCC, but argue against the methodological moral that Sober wishes to draw from it. Contrary to Sober, I argue that the possibility of exceptions to the PCC does not undermine its status as a central assumption for methods that endeavor to draw causal (...) conclusions from statistical data. 1 The PCC and the counter-example 2 Making non-stationary time series stand still 3 Sober's alternative. (shrink)

The role of measure theoretic atomicity in common cause closedness of general probability theories with non-distributive event structures is raised and investigated. It is shown that if a general probability space is non-atomic then it is common cause closed. Conditions are found that entail that a general probability space containing two atoms is not common cause closed but it is common cause closed if it contains only one atom. The results are discussed from the perspective of the Common Cause Principle.