DOJ/LEGAL INFO/Expert Systems Expert (Professor Carole Hafner)

[James Love <love@tap.org>]

Distributed to TAP-INFO, a free Internet Distribution List
(subscription requests to listproc@tap.org) 

TAXPAYER ASSETS PROJECT - INFORMATION POLICY NOTE
CROWN JEWELS CAMPAIGN - Juris, Legal Information
September 23, 1994

-    Academic Researcher Urges Department of Justice (DOJ) to
     make legal information more widely available.  Explains how
     limited access to federal case law has hampered research
     into "expert systems" for analysis of legal information.

The following is a letter from Professor Carole Hafner
(hafner@ccs.neu.edu, 617/373-5116) of Northeastern University to
Attorney General Janet Reno.  Professor Hafner is one of the
nation's best known academic researchers who study the use of
"expert systems" (artificial intellectual) to analyze and
research legal information.  Her letter discusses the
difficulties faced by academic researchers who cannot obtain
access to the large bodies of case law which are required to test
software.  Professor Hafner urges the Attorney General to take
pro-active steps to make federal case law more widely available,
in order to spur innovation and the development of new computer
assisted legal research products.  James Love, TAP
(jamie@tap.org)

-----------------------------------------------------------------
The Honorable Janet Reno
Attorney General of the United States
U.S. Department of Justice
10th Street and Constitution Avenue, N.W.
Washington, D.C.   20530

Dear Madam Attorney General:

I am writing this letter regarding the Department of Justice's
Request for Proposal for computer-assisted legal research.  As a
computer scientist whose goal is academic research, I would like
to add my voice to those asking the Justice Department to use its
market power to promote open public access to electronic legal
documents. My particular purpose in writing is to describe the
artificial barriers that currently preclude almost all serious 
research and innovation in computer-based legal research systems. 
The creation of a public archive of Federal case law and other
authoritative legal texts would remove these barriers. 

As one of the so-called ``pioneers'' in designing knowledge-based
systems for legal research, my own work has been frustrated for
nearly 20 years by the lack of available legal text to perform
realistic experiments. In the early 1970's, as a  graduate
student at the University of Michigan, I realized that 
computerization of legal indexing schemes (such as West key
numbers) would greatly increase their usefulness compared to the
same information on the printed page.  Researchers could move
from topic to related topic instantly, combining relevant topics
to form a precise statement of their research needs. The need to
laboriously type complicated queries would disappear. Many
structural limitations of traditional indices would not exist in
an automated system -- for example, a narrow topic could belong
to several broad categories instead of just one.

For my Ph.D. dissertation, published in 1978, I demonstrated the
feasibility of these ideas using a small sample database of 200
cases and 200 statutes from negotiable instruments law. I showed
how the computer could perform a crude sort of ``legal
reasoning'' to find relevant documents using a semantic network
of legal topics.

Unfortunately, no one (including myself) has ever been able to
test this approach in a practical setting, or develop products
for the legal market based on these or other knowledge-based
system principles, because the text to carry out experiments has
been in the hands of two large companies and one large government
agency (the Department of Justice), all of whom have refused to
share it.

The large companies that currently own the electronic text of
Federal law, and use it to sell on-line research services, have
adopted a consistent policy of refusing requests from researchers
for samples of text. Unlike many other technology companies, they
do not cooperate with University researchers. They have little
motivation to innovate, since they both use the same 1960's
technology and it would be expensive to change. Even today, 
legal indices such as the West key number system do not exist in 
computer-usable form, although this could be readily
accomplished.

In 1991, I requested from the Justice Department some samples
of federal case law from JURIS, to be included in a large
text database being developed by the Linguistic Data Consortium
at the University of Pennsylvania. The LDC is a NSF and ARPA
funded project to collect a large corpus of diverse kinds of text
and speech data that will be shared among all scientists for
computational linguistics research.

Although I knew that JURIS data was not normally available
outside the government, I thought perhaps I could convince the
DOJ to help us with this ``big science'' project.  Although
members of the JURIS staff were eager to help and interested in
the results of the research that I and others would perform on
the text, I was eventually told that DOJ attorneys had over-ruled
the decision of the JURIS group. (I did not know at the time 
that a contract with West prohibited sharing cases from JURIS,
even the portion in the public domain, with anyone outside the
Federal government.)

It is my professional judgement as a computer scientist
that the tools available for computer-assisted legal research
today are at least 10 years behind where they should be. The
legal profession has been denied the flood of consumer-oriented
products that a free market could have produced, and researchers
who believe that computerization of legal information offers a
fascinating challenge have been frustrated at every turn.

I was privileged to attend a recent meeting at the National
Archives organized by Bruce McConnell of OMB, at which these
questions were discussed, as well as an earlier meeting at the
DOJ organized by Mr. Colgate.  At both meetings, I was impressed
by the sincere desire of those representing the Administration to
have a better system come about for distributing electronic legal
documents. I do not think these representatives were swayed by
the publishers' statements that things are fine and the
marketplace is working. I do think they were understandably
confused about how they could make something happen within the
missions of their various agencies and departments.

I hope the Administration will see the wisdom of taking a
pro-active role in creating an archive of legal documents,
available to all at a reasonable cost. As with other paradigms
for disseminating the law,once the Federal government develops a
workable approach, state governments are likely to adopt similar
measures. Creating and maintaining an electronic legal archive
need not be very expensive, but it will require direction from a
policy level to define this as one of the government's
objectives. The current procurement of legal research services
presents a unique opportunity for the Justice Department to work
with other agencies to make this happen.  

An archive of legal documents, available to all at a reasonable
cost, would open the door to research, innovation and
competition. A public domain system of citations would ensure
that those innovations which prove superior could be utilized by
the legal profession.  With open competition, the incredible
number and variety of low-cost consumer-oriented computer 
products that we see appearing every month would include products
specially designed for legal research. The legal profession needs
these products, and the computer science profession wants to
create them.

Respectfully yours,

Carole D. Hafner
Associate Professor

cc: Alice Rivlin, Office of Management and Budget
cc: Sen. Edward Kennedy
cc: Sen. Carl Levin
cc: Rep. Bernie Sanders
cc: Rep. Marty Meehan
cc: Bruce McConnell, OMB
cc: Stephen Colgate, DOJ
cc: Rebecca Finch, LDC

---------------------------------------------------------------------
TAP-INFO is an Internet Distribution List provided by the Taxpayer
Assets Project (TAP).  TAP was founded by Ralph Nader to monitor the
management of government property, including information systems and
data, government funded R&D, spectrum allocation and other government
assets.  TAP-INFO reports on TAP activities relating to federal
information policy.  tap-info is archived at ftp.cpsr.org;
gopher.cpsr.org and wais.cpsr.org

Subscription requests to tap-info to listproc@tap.org with
the message:  subscribe tap-info your name
---------------------------------------------------------------------
Taxpayer Assets Project; P.O. Box 19367, Washington, DC  20036
v. 202/387-8030; f. 202/234-5176; internet:  tap@tap.org
---------------------------------------------------------------------