Research note
Abduction versus closure in causal theories

https://doi.org/10.1016/0004-3702(92)90073-7Get rights and content

Abstract

There are two distinct formalizations for reasoning from observations to explanations, as in diagnostic tasks. The consistency based approach treats the task as a deductive one, in which the explanation is deduced from a background theory and a minimal set of abnormalities. The abductive method, on the other hand, treats explanations as sentences that, when added to the background theory, derive the observations. We show that there is a close connection between these two formalizations in the context of simple causal theories: domain theories in which a set of sentences are singled out as the explanatorily relevant causes of observations. There are two main results, which show that (with certain caveats) the consistency based approach can emulate abductive reasoning by adding closure axioms to a causal theory; and that abductive techniques can be used in place of the consistency based method in the domain of logic based diagnosis. It is especially interesting that in the latter case, the abductive techniques generate only relevant explanations, while diagnoses may have irrelevant elements.

References (15)

  • V. Lifschitz et al.

    Miracles in formal theories of action

    Artif. Intell.

    (1989)
  • R. Reiter

    A theory of diagnosis from first principles

    Artif. Intell.

    (1987)
  • K. Clark

    Negation as failure

  • L. Console et al.

    Abductive reasoning through direct deduction from completed domain models

  • J. de Kleer et al.

    Characterizing diagnoses

  • J. de Kleer et al.

    Diagnosis with behavioral modes

  • H. Kautz

    A formal theory for plan recognition

There are more references available in the full text version of this article.

Cited by (62)

  • Optimizing group learning: An evolutionary computing approach

    2019, Artificial Intelligence
    Citation Excerpt :

    Douven and Wenmackers' [38] aim was to compare different update rules (rules for adapting probabilities in response to new evidence) within a social setting. Specifically, they compared Bayes' rule (Oaksford & Chater [98]) with an update rule intended to formalize the kind of explanatory reasoning that has lately been much in the limelight both in artificial intelligence (Konolige [72]; Boutilier & Becher [12]; Baral [8]; Lin & You [86]; Glass [52,53]; Teijeiro & Félix [121]) and in cognitive psychology (Koslowski et al. [75]; Bes et al. [10]; Williams & Lombrozo [131]; Legare & Lombrozo [83]; Lombrozo & Gwynne [88]; Douven & Schupbach [35,36]; Lombrozo [87]; Johnston et al. [69]; Douven & Mirabile [32]; Koslowski [74]). One way in which the present paper goes beyond Douven and Wenmackers' work is by considering a number of different formalizations of explanatory reasoning, particularly ones that were motivated by the aforementioned recent work in cognitive psychology.

  • Nonmonotonic reasoning

    2007, Handbook of the History of Logic
  • A causal approach to nonmonotonic reasoning

    2004, Artificial Intelligence
  • Abductive inference in defeasible reasoning: A model for research programmes

    2004, Journal of Applied Logic
    Citation Excerpt :

    Moreover, we provide criteria for representing the degree of success for selecting the most successful SRP among a group in a given context. There are good characterizations of abduction of surprising observations in monotonic theories [22,28]. In normal logic programs there is a tight relationship between SLDNF and the abduction of negative literals [20].

  • Preferences and explanations

    2003, Artificial Intelligence
View all citing articles on Scopus
View full text