COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002		As Principal Investigators of the DARPA  contract and NSF
C00009 ENDMK
C⊗;
	As Principal Investigators of the DARPA  contract and NSF
grants that support the robotics work in the Computer Science
Department, we have concluded that it is in the best interests
of the United States, Stanford and our work itself that Dr. Umnov not visit the
Computer Science Department
Robotics Laboratory and not be given copies of unpublished reports concerning
its work.  We hope that all members of the Robotics Laboratory will
co-operate in implementing this, although we have no sure way of requiring
such co-operation.  The reasons for this conclusion are as follows:

1. The robotics work has eventual military applications.  Classifying
the subject would probably harm our progress more than it would retard
the Soviets.  However, the rate of Soviet progress depends to an important
degree on information received from the West, and a lack of reports
and a lack of personal contact causes delays, even though all the information
will be published eventually anyway.

2. It has been argued that a mere demonstration visit to the Laboratory
and seeing the arms waved will not transfer much technology.  This is
possibly true, but the following considerations make refusing such a visit
desirable even if little transfer will occur.

Exactly how much technology is transferred will not be apparent
to the non roboticists who make decisions in the Soviet Union.  This is
especially evident in that it is to Umnov's personal interest to make his
superiors believe that sending him abroad is worthwhile.  He is not likely
to tell them that everything he learned was already in the journals even
if it turns out to be true.  They will
understand, however, Umnov's report on whether he visited the Laboratory
and whether he was given unpublished material.

For example, should the matter reach the level of Gurii Ivanovich Marchuk,
the head of the State Committee on Science and Technology (and a
computer scientist by background), he will get the message
that Umnov is being treated differently than he (Marchuk) was
during his visit to Stanford and differently than his former subordinate
Andrei Ershov was treated.

	This message has three aspects:

	a. It signals that American disapproval of recent Soviet aggression
in Afghanistan and Poland extends deeper into the population than a mere
tactical position of the Reagan Administration.

	b. It signals that they can no longer be sure that they immediately
get all unclassified information about American technology.  It suggests
caution about making assumptions of American weakness in other areas
as well.

	c. It signals that if they want access to American technological
developments they will have to be more forthcoming about their own.  In
this regard, the status of unpublished reports is particularly important.
It is Soviet law that unpublished scientific information may not be given
to foreigners except with special permission in each case.  With all the
personal hospitality that one receives in visiting the Soviet Union, this
law is rarely violated.  The State Department and the National Academy of
Sciences have bargained on our behalf for many years, and this bargaining
has resulted in greater American access to Soviet technology than countries
that don't bargain have achieved.  However, the effectiveness of the
bargaining has been greatly reduced by Soviet knowledge that American
scientists and technologists often don't co-operate with their government's efforts.

	There have been some hints that it violates some Stanford
policy for the Principal Investigators to try to restrict Umnov's access
to the Laboratory.  This hint appears to have some elements of bluff,
and part of the purpose of this letter is to call the bluff.
If it is against Stanford policy, the appropriate officials are requested
to state in writing what policy is being violated, to state in writing
what actions by the Robotics Laboratory are required by the policy, and
to state in writing what actions they propose to take in case of
non-compliance with their wishes.

John McCarthy, Professor of Computer Science

Tom Binford, Adjunct Professor of Computer Science