-3

the merits.

agonizing choice the judge would decide for the government on Nevertheless, even in that case, the government failed in the world because the data was published elsewhere. It is hard to imagine that the data can actually be suppressed without an elaborate network of government observation of

pending publication. Such an enforcement arrangement and its probable concomitant of the emergence of an "insider's" information network

would be a very heavy and real First Amendment price.

Second, there is a vagueness problem. The exact contours of what information can and cannot be published under the above statutes are almost necessarily classified; the definition relies upon administrative determinations which are not generally made public. This forces persons to "speculate" as to whether their actions violate the statutes. Gorin v. United States, 312 U.S. 19, 26 (1940). It thus poses a barrier to criminal prosecution and encourages injunctive enforcement. It also risks a chilling effect on the distribution of information outside the contours the government seeks to protect.

freedom.

Third, the statutes violate a tradition of scientific

Scientists have generally

and almost certainly

correctly argued that the advancement of science depends on a tradition of free inquiry and questioning. As the Special Senate Committee Report on the Atomic Energy Act of 1946 stated:

-4

...vital objectives in a sense compete with or are
in direct conflict with one another. The common
defense and security requires control over

information which might help other nations to

build atomic weapons or power plants...and, at the

same time, sufficient freedom of interchange

between scientists to assure the Nation of con

tinued scientific progress.

4/

As noted above, it is conceivable that this tradition may have to be restrained, but the need is to restrain it as little as possible consistent with the social demands of a new

era of technology.

5/

Fourth, restraints should be applied only if they actually help.

Restraints on large research programs such as those supported by government or industry might, for example, effectively serve the goal of suppressing a technology. But those on what an individual does on his or her own, working with pure logic or public domain information are unlikely to be effective. Learned Hand described it as "fatuous" to forbid the transmission of what is already in the public domain and read the espionage acts as not prohibiting such transmission, U.S. v. Heine, 151 F. 2d 813, 816 (2d Cir. 1945). The same is true of information developed by individuals working on their own it is too likely that someone else, somewhere else in the world, will be able to develop the same information.

-5

Fifth, it is precisely with respect to those tech

nologies that might be suppressed that public debate is most essential. The decision to control a technology that can

alone

bring both risks and benefits is a decision in which the entire society should have a role. The "born classified" concept would place this decision in the hands of government officials in complete antithesis to our traditions. Plausibly, the debate can sometimes be divided, as through distinguishing weapon design information from information relating to the desirability of nuclear secrecy; such a distinction is not always possible and any arrangement that leaves categorization of the information in government hands is likely in practice to be biased against public debate. This fundamental political point is strongest when the technology affects the balance between the government and the rest of society, an issue raised in the recent controversy concerning the development of private 6/ cryptographic systems.

After reflection on the above factors, we conclude that with respect to the publication of privately generated technical information such as an individual has generated working on his or her own, the severe civil liberties problems associated with restraint through either injunctive or criminal law approaches substantially outweigh the very limited information control

tion

-6

benefits that might be argued to accrue. This conclusion is clearly correct for what might be called "deducible" informainformation that can be developed "from known physical principles, careful reasoning, extrapolation from demonstrated capabilities of related technologies, and the large body of 1/ general information that is already in the public domain. There is no point in classifying such information; it must be assumed to be available to any nation or organization that makes a serious effort to obtain it. For slightly different reasons, we would reach the same conclusions for what we might call

"insightful" use of deducible information

based on deducible information

an uncommon intuition

such as the original idea that a chain reaction could be used to make a nuclear weapon or the mathematical insight that permits creation of a new class of cryptographic systems. It cannot be said that such insights will occur to another. But to attempt to restrict this type of insight is to chill an activity of great human importance; insights of this type often work to good. Moreover, as soon as the government relies on such an insight, as by building a nuclear weapon, the insight may become deducible to others.

In contrast, what can be reasonably classified is information that will not generally be accessible to the private thinker: the "non-deducible" technical choices made by the government, e.g. the key to a particular cipher; and information that can be

-7

gained by deduction or insight only from an elaborate experimental program of the type carried out or sponsored by the government. Thus, the principle we recommend need not necessarily prohibit the Congress from acting to restrain the Executive or perhaps industry from conducting large scale research programs in sensitive areas. Such large scale programs are likely to be essential to the development of information that can successfully be protected and is significant from a social or international viewpoint; even then, the information can only be protected for a limited period of time. Although there may be significant equal protection problems with restraints on such programs, such restraints are the least harmful from a civil liberties viewpoint.

Legislative implementation of our recommendation is

difficult and there are several important pitfalls in the way of this (or any other) reform in the area.

The first is the interplay between the provision at hand and the general espionage acts. It is possible, but not clearly established, that these acts do not apply to publication of information obtained from the government, except for those specific provisions using the word "publication." See New York Times, supra, (concurring opinion of Douglas, J.); H. Edgar & B. Schmidt, "The Espionage Statutes and Publication of Defense Information," 73 Colum. L. Rev. 929, 1032 ff (1973).