Saturday, July 15, 2006

Causality, Responsibility, and Blame

I have found no references in LEXIS or WESTLAW to the interesting paper Responsibility and Blame: A Structural-Model Approach (Nov. 6, 2003) by Hana Chockler & Joseph Y. Halpern.

Granted, the mathematical notation and symbolic logic are very difficult for amateurs like me. But this can't be the entire explanation for the seeming indifference of the legal academy: many law teachers now have advanced degrees. Furthermore, Chockler and Halpern use ordinary English and nifty examples to describe the gist of their thesis. So why the deafening silence in the legal academy? Consider: Halpern is one smart cookie. (The same is very probably true of Chockler!)

Perhaps the explanation is simple lack of serendipity: although the paper by Chockler and Halpern resides on Halpern's web site and has been in circulation for several years, it apparently has not yet been published in a peer review journal.

P.S. This post does not constitute an endorsement (or repudiation) of Halpern's theory of causality, explanation, responsibility, and blame. I am insufficiently familiar with Halpern's views either to endorse or reject them. But I would like to hear what thoughtful legal scholars think of Halpern's views of questions that seem to have a bearing on law.

Thursday, July 13, 2006

Lowering the Bar

In People v. Ibarra, 2001 WL 1330296 (Cal.App. 2001) (not certified for publication) the prosecutor had an interesting view of the meaning of the government's burden of persuasion in criminal trials. Using a bar graph, she asserted that
proof beyond a reasonable doubt requires at least a 60 percent probability of guilt

proof beyond a shadow of a doubt puts the bar at 70 percent

proof beyond beyond all doubt sets the bar at 80 percent, and

proof to an absolute certainty raises the bar to 90 percent.

Perhaps proof to an utterly conclusive and absolutely certain certainty with no suspicion of a hint of a shadow of a doubt would require that the probability of guilt be shown to be 100%?

But given the prosecutor's view of matters it hardly matters which formulation of the prosecution's burden of proof would require that the trier believe that the chances of a defendant's guilt are 100% -- because on the prosecutor's view of the reasonable doubt standard a trier of of fact can find a defendant guilty of a crime such as murder even though the trier thinks there is a 40% chance that the defendant is innocent. (The prosecutor quite evidently -- dare I say certainly? -- thinks that it is important not to set the bar too high.)

The prosecutor's view of the meaning of reasonable doubt is somewhat at variance with the view of the medieval philosopher Maimonides, who declared "it is better and more desirable to free a thousand sinners, than ever to kill one innocent." United States v. Fatico, 458 F. Supp. 388, 410 (E.D.N.Y., 1978) (quoting N.L. Rabinowitz, Probability and Statistical Inference in Ancient and Medieval Jewish Literature 111 (1973). Compare Justice Harlan's concurring opinion in In re Winship, 397 U.S. 358, 368 (1970), where he said that the reasonable doubt standard is "bottomed on a fundamental value determination of our society that it is far worse to convict an innocent man than to let a guilty man go free." Id. at 380.

Tuesday, July 11, 2006

Conference: Graphic and Visual Representations of Evidence and Inference in Legal Settings

Cardozo School of Law will host a conference on

Graphic and Visual Representations of Evidence and Inference in Legal Settings


Dates: January 28-29, 2007.

Venue: Cardozo Law School, Yeshiva University, 55 Fifth Avenue (5th Ave. & 12th St.), New York (Manhattan), New York


 

Description of conference:

One of the largest problems faced by crime investigators, litigators, paralegals, judges, triers of fact, and other actors interested in disputes about factual questions in legal settings is the sheer mass of available evidence in many cases. It is often difficult to remember, retrieve, and interpret evidential information, so that patterns, relations, and inconsistencies often go unnoticed. Tools that support the storage, retrieval, and interpretation of masses of evidence could therefore be of great use.

Psychological studies have shown that people's ability to remember, retrieve, and interpret information is greatly enhanced if they organize information in a way that is meaningful to them. Scholars of the law of evidence have long suggested that graphical representations of evidential arguments and inferences could support humans in making sense of masses of evidence. As early as 1913, John Henry Wigmore claimed that his charting method promoted rational thinking about legal evidence. While Wigmore had only pencil and paper to draw his cumbersome graphs, today the computer could make his ideas practically feasible for everyone. Software could be used to draw graphical representations of arguments and inferences about masses of evidence. Moreover, such software could be combined with existing database, document management, and search technology so that collections of evidentiary documents could be stored and retrieved in terms of the user's thinking about a case. Such software would also facilitate transfer of case files to others by increasing the transparency of the files, so that subsequent investigators, prosecutors, and fact finders could gain a quicker and better understanding of the case.

Such software is currently being investigated for use in various domains. Argument visualization software has been designed, for instance, to support the teaching of scientific reasoning or critical thinking skills (Belvedere, Reasonable, Araucaria, Convince Me), to support intelligence analysis, and to facilitate individual or collaborative problem solving (Questmap, SEAS). Moreover, current artificial intelligence research offers precise accounts of evidential reasoning and thus provides a clear semantics of graphical notations as well as ways to compute with them.

In the legal domain, fact investigators and litigators increasingly use software that supports the storage and retrieval of information in terms of conceptual and relational networks (Holmes 2, Analyst's Notebook). However, as yet, such tools offer little or no support for the structuring of human thinking about thus stored information. This software allows users to store evidentiary data in terms of events, objects, actors, and the relations among these things, but it does not allow users to represent how such data support or undermine hypotheses about what has happened.

This interdisciplinary conference brings together scholars and practitioners from such fields as law, philosophy, computer science, artificial intelligence, cognitive psychology, and linguistics who are interested in the graphic visualization of legal evidentiary inference and its support by software tools. The following issues will be addressed:

  • Current and new graphical means to visualize factual inference and proof
  • Semantics of such graphical notations: what are the underlying theories of evidential reasoning? (jurisprudential, philosophical, psychological, rhetorical, logical, or mathematical)
  • Which software tools for graphical representations of factual inference and proof are currently available or being developed?
  • What are the potential contexts for the use of such software and what are the potential benefits of such software? (crime investigation, litigation, trial, law teaching, etc.)
  • To what extent can graphic representation of evidential arguments support the automatic evaluation of hypotheses?
  • How can current insights about human-computer interaction be exploited to increase the usefulness of such software? (e.g., how can visual complexity created by the size of the available mass of evidence be managed?)
  • Are empirical results available on usability and effects of use of charting methods (whether manual or digital) in legal or other contexts?
  • What are the practical constraints faced by crime investigators or legal professionals who want to use such software?
  • Conference officials:
    Peter Tillers (Cardozo Law School): Conference chair e-mail address: peter@tillers.net
    Henry Prakken (Universiteit Utrecht / University of Groningen): Program chair e-mail address: henry@cs.uu.nl
    Thomas D. Cobb (University of Washington, Seattle): Deputy program chair e-mail address: tomcobb@u.washington.edu

    Panelists:

  • Thomas D. Cobb (University of Washington School of Law)
  • Philip Dawid (U. College London)
  • Neal Feigenson (Quinnipiac University School of Law)
  • Branden Fitelson (U. of California at Berkeley)
  • Tim van Gelder (U. of Melbourne)
  • Thomas F. Gordon (FOKUS [Frauenhofer Institut fuer Offene Kommunikationssysteme]); web log
  • John Josephson (Ohio State University)
  • Marc Lauritsen (CEO, Capstone Practice Systems)
  • Richard Lempert (U. of Michigan Law School)
  • Ronald P. Loui (Washington University, St. Louis; Computer Science)
  • John D. Lowrance (Program Director, Artificial Intelligence Center, SRI International)
  • Jennifer Mnookin (UCLA School of Law)
  • Dale Nance (Case School of Law)
  • Priit Parmakson (Tallinn University, Tallinn, Estonia)
  • John L. Pollock (U. of Arizona)
  • Henry Prakken (Utrecht University & U. Groningen)
  • Chris Reed (U. of Dundee)
  • Burkhard Schafer (U. of Edinburgh, Law School)
  • David Schum (George Mason U.)
  • Richard Sherwin (New York Law School)
  • Samuel Solomon (CEO of DOAR, Inc.)
  • Peter Tillers (Cardozo Law School, Yeshiva University)
  • William Twining (U. College London, Law Faculty & U. of Miami, School of Law)
  • Bart Verheij (U. Groningen, Dept. of Artificial Intelligence; ALICE Institute)
  • Vern Walker (Hofstra U. School of Law)
  • Douglas Walton (U. of Winnipeg)

    Program details will be provided later.



    Drafts and abstracts of some or all papers will be made available online at http://tillers.net/conference.html shortly before the conference begins. Final versions of the papers will be published in Law, Probability and Risk in 2007 and 2008.


    The public is warmly encouraged to attend the conference. Advance registration is not required and there is no registration fee. However, there will be a charge for any lunches or dinners that attendees elect to take at the site of the conference, at Cardozo Law School. If you are not a panelist and would like to join us for any lunches or dinners at Cardozo, please RSVP to the address shown below by January 15, 2007, and indicate which meals you wish to purchase, enclosing payment of $35 for each. Please make checks payable to Cardozo School of Law. Send payment to

    Alisa Norr
    Legal Secretary
    Friedman Kaplan Seiler & Adelman LLP
    1633 Broadway, 46th floor
    New York, NY 10019
    United States


    Hotel and flight arrangements may be made through the conference travel agency:

    Morris Park Travel Bureau
    1745 Wiliamsbridge Road
    Bronx, NY 10461
    1-718-792-9850 or toll free 1-877-526-8844
    Fax: 1-718-863-7121
    Email: MPTVL@AOL.COM
  •