Online research in philosophy

In this article we evaluate the BankXX program from several perspectives. BankXX is a case-based legal argument program that retrieves cases and other legal knowledge pertinent to a legal argument through a combination of heuristic search and knowledge-based indexing. The program is described in detail in a companion article in Artificial Intelligence and Law 4: 1--71, 1996. Three perspectives are used to evaluate BankXX:(1) classical information retrieval measures of precision and recall applied against a hand-coded baseline; (2) knowledge-representation and case-based reasoning, where the baseline is provided by the functionality of a well-known case-based argument program, HYPO (Ashley, 1990); and (3) search, in which the performance of BankXX run with various parameter settings, for instance, resource limits, is compared. In this article we report on an extensive series of experiments performed to evaluate the program. We also describe two additional experiments concerning(1) the program's search behavior; and (2) the use of a modified form of precision and recall based on case similarity. Finally we offer some general conclusions that might be drawn from these particular experiments.

Frank Jackson and Philip Pettit have defended a non-reductive account of causal relevance known as the ‘program explanation account’. Allegedly, irreducible mental properties can be causally relevant in virtue of figuring in non-redundant program explanations which convey information not conveyed by explanations in terms of the physical properties that actually do the ‘causal work’. I argue that none of the possible ways to spell out the intuitively plausible idea of a program explanation serves its purpose, viz., defends non-reductive physicalism against Jaegwon Kim’s Causal Exclusion Argument according to which non-reductive physicalism is committed to epiphenomenalism because irreducible mental properties are ‘screened off’ from causal relevance by their physical realizers. Jackson and Pettit’s most promising explication of a program explanation appeals to the idea of invariance of effect under variation of realization , but I show that invariance of effect under variation of realization is neither necessary nor sufficient for causal relevance.

Causal modeling methods such as path analysis, used in the social and natural sciences, are also highly relevant to philosophical problems of probabilistic causation and statistical explanation. We show how these methods can be effectively used (1) to improve and extend Salmon's S-R basis for statistical explanation, and (2) to repair Cartwright's resolution of Simpson's paradox, clarifying the relationship between statistical and causal claims.

As part of a series of contributions on patterns in the periodic table, the relationships among the transition metals are examined here in a systematic manner. It is concluded that the traditional method of categorizing transition elements by group or by period is not as valid as by using combinations thereof. From chemical similarities, it is proposed that the transition metals be considered as the [V–Cr–Mn] triad; the [Fe–Co–Ni–Cu] tetrad; the [Ti–Zr–Hf–Nb–Ta] pentad; the [Mo–W–Tc–Re] tetrad; and the [Ru–Os–Rh–Ir–Pd–Pt–Au] heptad. Silver does not fit neatly in anywhere and is better linked with thallium.

This paper is an empirical critique of causal accounts of scientific explanation. Drawing on explanations which rely on nonlinear dynamical modeling, I argue that the requirement of causal relevance is both too strong and too weak to be constitutive of scientific explanation. In addition, causal accounts obscure how the process of mathematical modeling produces explanatory information. I advance three arguments for the inadequacy of causal accounts. First, I argue that explanatorily relevant information is not always information about causes, even in cases where the explanandum has an identifiable causal history. Second, I argue that treating theoretical explanations as reductions from general causal laws does not accurately describe the types of "top-down" explanations typical of dynamical modeling. Finally, I argue that causal/mechanical accounts of explanation are intrinsically vulnerable to the irrelevance problem.

Recent philosophical studies of probabilistic causation and statistical explanation have opened up the possibility of unifying philosophical approaches with causal modeling as practiced in the social and biological sciences. This unification rests upon the statistical tools employed, the principle of common cause, the irreducibility of causation to statistics, and the idea of causal process as a suitable framework for understanding causal relationships. These four areas of contact are discussed with emphasis on the relevant aspects of causal modeling.

In 1982, when computers were just becoming widely available, I was a graduate student beginning my work with Clark Glymour on a PhD thesis entitled: “Causality in the Social Sciences.” Dazed and confused by the vast philosophical literature on causation, I found relative solace in the clarity of Structural Equation Models (SEMs), a form of statistical model used commonly by practicing sociologists, political scientists, etc., to model causal hypotheses with which associations among measured variables might be explained. The statistical literature around SEMs was vast as well, but Clark had extracted from it a particular kind of evidential constraint first studied by Charles Spearman at the beginning of the 20th century, the “vanishing tetrad difference.”1 As it turned out, certain kinds of causal structures entailed these constraints, and others did not. Spearman used this lever to argue for the existence of a single, general intelligence factor, the infamous g (Spearman, 1904).

The statistical evidence for the detrimental effect of exposure to low levels of lead on the cognitive capacities of children has been debated for several decades. In this paper I describe how two techniques from artificial intelligence and statistics help make the statistical evidence for the accepted epidemiological conclusion seem decisive. The first is a variable-selection routine in TETRAD III for finding causes, and the second a Bayesian estimation of the parameter reflecting the causal influence of Actual Lead Exposure, a latent variable, on the measured IQ score of middle class suburban children.

The statistical community has brought logical rigor and mathematical precision to the problem of using data to make inferences about a model’s parameter values. The TETRAD project, and related work in computer science and statistics, aims to apply those standards to the problem of using data and background knowledge to make inferences about a model’s specification. We begin by drawing the analogy between parameter estimation and model specification search. We then describe how the specification of a structural equation model entails familiar constraints on the covariance matrix for all admissible values of its parameters; we survey results on the equivalence of structural equation models, and we discuss search strategies for model specification. We end by presenting several algorithms that are implemented in the TETRAD II program.

We applied TETRAD II, a causal discovery program developed in Carnegie Mellon University’s Department of Philosophy, to a database containing information on 204 U.S. colleges, collected by the US News and World Report magazine for the purpose of college ranking. Our analysis focuses on possible causes of low freshmen retention in U.S. colleges. TETRAD II finds a set of causal structures that are compatible with the data.