Elamur Research⌕
☰
home / papers / 2009 / Analysis of the Talmudic Argumentum A Fortiori Inference Rule (Kal Vachomer) using Matrix Abduction

Analysis of the Talmudic Argumentum A Fortiori Inference Rule (Kal Vachomer) using Matrix Abduction

Paper · 2009 · M. Abraham, Dov M. Gabbay and Uri J. Schild

Summary

The paper asks how one can justify filling a missing value when only a small table of yes/no facts is available. This is the kind of non-deductive step the Talmud takes with Kal-Vachomer (argument from the lesser to the greater) and with analogy. The authors formalise the task as Matrix Abduction. Rows are agents or situations, columns are features, and every entry except one is 0 or 1. The algorithm then decides whether the blank should become 0, become 1, or stay undecided. The method only applies when all columns point in a shared "direction", such as more damage or better performance. Each candidate filling is tested by treating columns as vectors, comparing the partial orders they induce, and preferring the ordering that is structurally better behaved. The framework is shown on online shopping tables, a bull-damage law case, sentencing for traffic offences, and a full analysis of the Kal-Vachomer passage in Kidushin 5a–5b. Appendices extend it to argumentation networks, voting and judgement-aggregation paradoxes, finite monadic models and access control. The result is a general, algorithmic account of reasoning by analogy and a fortiori.

Key points

  • The problem is a 0/1 matrix with exactly one unknown cell. There are three possible outcomes: 0, 1, or undecided.
  • Applicability depends on a "direction" condition: 1 must mean "more" along one agreed axis for every column. A matrix without direction, such as a bare binary relation, gives meaningless answers.
  • In the bull example, foot damage in a public place pays 0, horn damage there pays ½, and foot damage in a private place pays in full. Setting the unknown to x ≥ ½ makes the columns linearly ordered, while x = 0 makes them incomparable, so the paper argues for the former.
  • Shopping illustrations use four LCD screens and four cameras from a real price-comparison site, each with one missing feature.
  • Appendix D shows that a theorem prover searching for a minimal model of a formula built from the matrix yields a set realisation of it. Monadic models with m predicates reduce to at most 2^m elements.

Interesting ideas

  • I1. Before inferring anything, check that the data are directional, meaning every feature is encoded so that 1 pushes the same way; otherwise refuse to infer.
  • I2. Fill a gap by trying each candidate value and keeping the one that yields the simplest, most ordered structure among the rows and columns, with "undecided" as a legitimate answer.
  • I3. An a fortiori argument can be read both by rows and by columns, such as severity of place versus severity of action, and agreement between the two readings strengthens the conclusion.
  • I4. Choosing a binary cut-off for numeric data, such as wind above 150 mph, is a modelling decision that must follow the shared direction, not mere notation.
  • I5. Minimal-model search with a theorem prover can recover a compact set labelling that realises a whole matrix.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (world): Compiled knowledge tables, such as verb classes or value scales, could carry an explicit polarity per column, so that gap-filling is allowed only when polarity is consistent.
  • I2 (world): A missing entity state or holding could be guessed by choosing the value that keeps the comparison order over entities coherent, and the reason could be logged as an explainable step.
  • I3 (answer): Comparative questions in fables or word problems ("if even X, then surely Y") could be checked from both axes, with an answer given only when both readings concur.
  • I4 (world): When measures are turned into boolean features, record the threshold and its direction so that later inferences stay traceable.
  • I5 (learn): Self-induced repair rules could be compressed by asking a prover for the smallest set of labels that explains an observed feature matrix of parser errors.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Bar-Ilan UniversityRamat Gan · Israel150 works · 25,428 citations🎓King's College LondonLondon · United Kingdom50 works · 1,852 citations

 Fields, topics and concepts

📚 Shelves 3The field of science this work belongs to (OpenAlex fields).Computer Science · 19,503 worksArtificial Intelligence · 18,805 worksComputational Theory and Mathematics · 496 works

🧭 Topics 1Research subjects and languages the work is about.Hebrew & Aramaic · 502 works

Cite

Cite as M. Abraham, Dov M. Gabbay and Uri J. Schild. 2009. Analysis of the Talmudic Argumentum A Fortiori Inference Rule (Kal Vachomer) using Matrix Abduction. Studia Logica.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1007/s11225-009-9202-5

@inproceedings{abraham2009analysis,
  title = {{Analysis of the Talmudic Argumentum A Fortiori Inference Rule (Kal Vachomer) using Matrix Abduction}},
  author = {M. Abraham, Dov M. Gabbay and Uri J. Schild},
  booktitle = {Studia Logica},
  year = {2009},
  doi = {10.1007/s11225-009-9202-5},
  url = {https://doi.org/10.1007/s11225-009-9202-5}
}

References

26 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Barker, S., G. Boella, D. M. Gabbay, and V. Genovese, ‘A Meta-model of Access Control in a Fibred Security Language’, Studia Logica 92: 437–476, 2009. find
  2. Barringer, H., D. M. Gabbay, and J. Woods, ‘Temporal dynamics of argumentation networks’, in D. Hutter and W. Stephan (eds.), Mechanising Mathematical Reasoning, LNCS 2605, Springer, 2005, pp. 59–98. find
  3. Becker, M. Y., J. F. Mackay, and B. Dillaway, ‘Abductive Authorization Credential Gathering. http://research.microsoft.com/pubs/80508/becker2009ieee-policy-submission.pdf ↗
  4. Besnard, P., and A. B. Hunter, Elements of Argumentation, MIT Press, 2008, 300 pages. find
  5. Boella, G., D. M. Gabbay, V. Genovese, and L. van der Torre, ‘Fibred Security Language’, Studia Logica 92: 395–436, 2009. find
  6. Brams, S. J., D. M. Kilgour, and W. S. Zwicker, ‘The paradox of multiple elections’, Social Choice and Welfare 15: 211–236, 1998. find
  7. Caminada, M. W. A., and L. Amgoud, ‘On the evaluation of argumentation formalisms’, Artificial Intelligence 171 (5–6): 286–310, 2007. find
  8. Dung, Phan Minh, ‘An argumentation theoretic foundation for logic programming, Journal of Logic Programming 22 (2): 151–171, 1995. find
  9. Fisch, M., Rational Rabbis: Science and Talmudic Culture, Indiana University Press, Bloomignton, 1997. find
  10. Gabbay, D. M., Labelled Deductive Systems, OUP, 1996. find
  11. Gabbay, D. M., and J. Woods, The Reach of Abduction, Elsevier, 2005. find
  12. Gabbay, D. M., and J. Woods, ‘Resource origins of non-monotonicity’, Studia Logica 88: 85–112, 2008. find
  13. Grabenhorst, T. K., Das Argumentum A Fortiori, Peter Lang, 1990. find
  14. Hartmann, S., G. Pigozzi, and J. Sprenger, ‘Reliable methods of judgement aggregation’, Draft, 23 February 2009. find
  15. Hasan, A., Analogical reasoning in Islamic jurisprudence, 1986, Republished Adam Publishers, 2007, 486 pages. find
  16. Jacobs, L., Studies in Talmudic Logic and Methodology, London, Vallentine-Mitchell, 1061. Republished paperback, 2006. find
  17. Kamali, M. H., Principles of Islamic Jurisprudence, Islamic Text Society, 3rd Revised Eition, 2002, 550 pages. find
  18. Kornhauser, L., and L. Sager, ‘Unpacking the court’, Yale Law Journal 96: 82–117, 1986. find
  19. Kornhauser, L., and L. Sager, ‘The one and the many: adjudication in collegial courts’, California Law Review 81: 1–51, 1993. find
  20. Kunst, A., ‘An overlooked type of inference’, Bulletin of the School of Oriental and African Studies, X, part 4, 1942, pp. 976–991. find
  21. Neusner, J., ‘The making of the mind of Judaism; the formative age’, Brown Judaic Studies, vol. 133, Scholars press, Atlanta, 1987. find
  22. Pigozzi, G., and L. van der Torre, ‘Premise independence in judgement aggregation’, Dagstuhl Seminar 07531, 2007. find
  23. Prakken, H., and G. Sartor, ‘Argument based extended logic programming with defeasible priorities’, Journal of Applied Non-classical Logics 7: 25–75, 1997. find
  24. Schild, U. J., ‘Criminal Sentencing and Intelligent Decision Support’, Artificial Intelligence and Law 6: 2–4, 1998. find
  25. Schwarz, A., Der Hermeneutische Syllogismus in der Talmudischen, Ltitteratur, Karlsruhe, 1901. find
  26. Stebbing, V. L. S., A Modern Introduction to Logic, London, 1945. M. Abraham U. Schild Bar-Ilan University, Israel mdabraham@bezeqint.net schild@cs.biu.ac.il Dov M. Gabbay Department of Computer Science King’s College, London, UK and Bar-Ilan University, Israel dov.gabbay@dcs.kcl.ac.uk find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/abraham-2009-analysis-talmudic-argumentum-fortiori/