[Brought to you by Dan's Typing Service (sm). The tape was missing a
few sections, and there was occasionally too much background noise, but
he got mostly everything.  This transcript is completely unofficial and
may well be wrong in places.  We'll publish the official court
reporter's transcript when it arrives.]

The players:

Scott McIntosh:  Justice Dept. attorney, defending the Government
Myron Bright: Judge, visiting from the 8th Circuit (Fargo, N. Dakota)
Betty Fletcher: Judge, 9th Circuit Court of Appeals
Thomas Nelson: Judge, 9th Circuit Court of Appeals
Cindy Cohn:  Prof. Bernstein's attorney.

Date:  December 8, 1997
Location:  9th Circuit Court of Appeals, Mission Street, San Francisco
Time:  Approx 11:00 AM


McIntosh: [XXX] by my count uh something like 400 pages of briefs from
the parties and the amici in this case, and we obviously can't cover all
of the terrain that's covered by those briefs in the time we have this
morning.

But I do think it's possible to get to the heart of the case. And the
way to do that is to be clear about what _is_ at issue here and what is
_not_ at issue.

What's not at issue in this case is whether the government can undertake
to prevent the free public exchange of information and ideas about the
discipline of cryptography. That's _not_ at issue, because the Export
Administration Regulations which are being challenged in this case make
no attempt to do that. To the contrary, the export regulations at issue
in this case go out of their way to make---to place outside the
regulatory scheme the exchange of information and ideas about
cryptography. As you know from our briefs, all publicly available
technology, which is the term of art used by the the EAR, all publicly
av---uh available information about cryptography is simply outside the
scope of EAR and its licensing requirements. That means, for example,
that information presented in academic settings, in courses of
instruction, in seminars, in universities, information that has been
published or will be published in books, magazines, um, information
that's a product of basic academic research in an academic setting, all
of that information is freely available and is not subject to licensing
by the federal government. You will find in the record in this case
literally hundreds of pages of books---taken from books and articles and
university syllabuses, devoted to the subject of cryptography, and the
government does not require a license for any of them.

Bright: Well then what's the big deal about uh putting the source code
on computers?

McIntosh: Well your honor, source code, here we here we move from what
isn't at issue to what is at issue. What the government controls are
products that are capable of encrypting data. It controls hardware, or
it controls the export of, hardware and software that can encrypt data.
And the parties have stipulated that encryption source code is capable
of being used by a computer to encrypt data.

Bright: Well, I'm I notice you cite uh the Lakewood uh City of Lakewood
case uh as supporting your position. Of course, City of Lakewood case
was a restriction on uh uh newspapers only indirectly. It was uh the uh
type of uh of uh units that are out on the street in which people buy
newspapers. And so that wasn't strictly First Amendment, but it was
something that related to speech uh print in this case print.

McIntosh: Uh that's correct, your honor. They were the newsstands as you
say that were used for distributing um newspapers and other periodicals.
Uh what the Court---the distinction the Court tried to draw in Lakewood,
and the distinction on which we rest in this case, is the distinction
between regulations, a licensing scheme, that is focused narrowly and
specifically at expression or activities commonly associated with
expression, and on the other hand a general licensing scheme, which may
happen to include within its reach particular activities that can be
undertaken for expressive purposes, but which is nonetheless a general
scheme that doesn't single out in a narrow and specific fashion
expressive activities. That's the distinction the Court is drawing in
Lakewood. That's the distinction this court drew in the Roulette case.

Fletcher: Let me ask you another question, uh apropos what you say this
case does not include. At some earlier point, uh the government was
contending that uh his---Bernstein's papers on this subject were subject
to licensing. Now they've abandoned that position.

McIntosh: The, the, there are, uh let me just preface this by saying of
course we're dealing with a facial challenge here. It doesn't turn on
the particular circumstances of Professor Bernstein. But, turning to
those circumstances, Professor Bernstein um ultimately presented six---
we we're back on the prior regulatory scheme, the State Department
scheme---presented the State Department six distinct documents if you
will. Um two of them, two of those documents, represent the source code
itself for his Snuffle encryption program.

Fletcher: And you still think those are subject to licensing.

McIntosh: The the software itself is subject to licensing. There's no
dispute---

Fletcher: No, but the---the paper describing it---

McIntosh: I'm sorry, your honor. The---source code in printed form, in a
book, a magazine, some other printed medium, is not subject to
licensing. But the in---the the the request, the inquiry that he was
making, wasn't directed at least as I understand it to the to the
physical piece of paper itself. It was directed to the software in some
electronic, in some medium, whether it be a diskette, or on the
Internet, some other computer-ready medium---

Fletcher: There must be some confusion in the record about that then.

McIntosh: I I I don't think there's any---I think at this point, your
honor, there's no debate between the parties as to what the government,
with respect to Professor Bernstein, what the government thinks is
subject to licensing and what the government thinks is not. Uh Ms. Cohn
can uh correct me if I'm wrong on that. But I think it's fairly clear at
this point his software, the source code for the the Snuffle program, is
subject to the licensing requirements of the EAR as encryption software.

Fletcher: Is the---This, this is a very puzzling case, at least to to my
mind. If you can put it all down on paper and all that's involved is a
lotta hard work to put it into uh software or hardware, why does the
government do this?

McIntosh: Well, your honor, the the short answer is that the government,
uh not simply under the current regulations but under the prior
regulations in the State Department, has permitted the export of printed
materials. And that practice, permitting that export, has not thus far
compromised the government's intelligence-gathering capabilities. Um if
if in fact it had resulted in the inability of the government to engage
in intelligence-gathering activities, I think you would see either the
government giving up the game, or, more realistically, trying to
exercise control over the the paper materials themselves---

Fletcher: Well, after the barn doors are open and the pigs have fled, I
don't uh---

McIntosh: You, your honor, the the reason that the government is is
engaged in the control of encryption software is preci---notwithstanding
the paper exception---is precisely because, in the real world, that
doesn't open the barn doors in any way that thus far has actually
compromised the government's ability to c---to to engage in intelligence
activities. The the um. Eh um I think it's a matter of common sense. If
you were to say to the American software industry, you have a choice,
you can export books which have printed source code in them, or you can
export diskettes that have computer-ready programs on them, which one is
going to result in broader use abroad, which one is gonna result in
broader uh demand for American products abroad, I think the answer's
obvious. Um the the assumption behind um all of the public debates on
these export controls is that the export controls if they were lifted
would result in some fairly material expansion of American [XXX]
software being exported abroad.

Bright: This hasn't been argued either side, but you're objecting to the
software being distributed being used uh to go overseas---and this is a
prior restraint case, as I gathered, and Bernstein has not applied for a
license on his uh program, is that right?

McIntosh: He has not applied for a license that's correct.

Bright: Now the question I have, and as I say it bothers me. And maybe
you've got an answer, maybe you haven't. Uh does the First Amendment uh
apply to overseas beyond the borders of this country? Nobody's argued
that point, and and that seems to be something that at least is of a
concern to me.

McIntosh: Uh it it's clear under this court's decision in the Bullfrog
Films case, your honor---uh this is from about ten years ago---that at
least to some extent the First Amendment does have applications to
communications abroad uh and the government has not contended in this
court that there is no First Amendment issue at all simply because of a
foreign locus. I do think there is room for argument uh that the foreign
setting the national security concerns uh would result in a a less uh
extensive degree of First Amendment protection. But we don't think that
you need to resolve that, because we think under conventional First
Amendment standards the government's um activities here are
constitutional.

Fletcher: Then you're not contending that the---our government's
capacity to protect its communications is compromised. What you're
asserting is that what's compromised is your capacity to eavesdrop on
others.

McIntosh: If if if you like it it's the so-called signals intelligence
capabilities of the of the United States, the ability to obtain
electronic intelligence from from foreign intelligence targets. That's
what---that's what these regulations are designed to protect, is that
capacity. And I think what's important to understand is that that
capacity that the concern about the government's, that capacity, is what
drives these regulations not only with respect to hardware and
encryption object code but even encryption source code. The President's
executive order and his memorandum are very clear on that and the
structure of the regulations is very clear on that. All of these
products are regulated because of their capacity to make computers
encrypt data.

Bright: Uh you say that uh you're concerned with the function of the
c---of uh the source code. And yet and it's and you're saying that
that's not uh communication in the First Amendment sense. Is that your
position?

McIntosh: I I I wouldn't put it quite that way, your honor. Let me let
me try to put it this way. The activity that's being regulated here is
the export of encryption software, more particularly here the export of
encryption source code. It is possible and the government does not
dispute that you---that one can distribute source code for expressive
purposes. That is that is true. It is also true, and I think undisputed,
that one can distribute source code, and source code can be used, for
wholly non-expressive purposes.

Bright: But in this case Bernstein says he wants to use this source code
for expressive pur---purposes. He's gonna put it in a book or he may put
it in a book, and now he wants to put it in the computer so that foreign
uh scientists and experts can communicate and discuss uh the importance
of this source code and how it may be amplified and so forth to be used
commercially. Why isn't that expression and---why are you saying that
the fact that it's functional takes it out of the First Amendment uh
context?

McIntosh: No, your honor. I I I want to be clear on that. We do not
contend that it simply take---taken out of the First Amendment context.
Our contention is a different one. It is that when the government
regulates any activity for reasons that are unrelated to the expression
involved, for reasons that are not based on some disagreement with the
views or ideas or information being conveyed, that a very different
First Amendment analysis applies.

Bright: Well isn't that, wasn't that the analysis however, you said a
different reason? And that's how, is that how you distinguish this case
uh from uh the uh [XXX] Plain Dealer case.

McIntosh: Well, your honor.

Bright: Lake---the Lakes---Lakewood City---

McIntosh: Lakewood case. Yes your honor. Um uh that's part of it, your
honor, um I think the the fact that the government is not---is plainly
not undertaking suppress information about cryptography, that's not the
purpose of these regulations, places this case at a considerable remove
from the sort of core prior restraint cases like the Pentagon Papers
case, where the whole object of the prior restraint was to keep people
from learning information. That's not what the government's regulations
are trying to---

Fletcher: They're just keeping them from using it.

McIntosh: Not using the information, your honor. Using the software,
because---

Fletcher: That, that's just words, counsel.

[laughter]

McIntosh: No no, your honor. I don't think so. Uh Professor Bernstein is
free for example, is---Professor Bernstein I think he said in his brief
that he has uh a colleague in New Zealand, if I'm remembering correctly,
who he'd like to uh present this information to, as he puts it, his
ideas about cryptography. Well he can send his source code in printed
form to his colleague in New Zealand, or any other colleague abroad,
practically any purposes. He's free to do so. The government isn't
trying to stop him from doing that. He can go to a seminar or a lecture
series in Paris or Oslo or anywhere else he wants to. He can talk about
his theories, he can present, pardon me, present them to his colleagues,
he can for that matter hand out printed versions of the source code at
at his seminars---

Nelson: Couldn't he export himself? In other words, couldn't he go to
London and enter into an agreement with uh---make up uh a uh partnership
with uh, become Bernstein and Widget LLC and uh actually make these uh uh
disks over there and sell them any place he wants?

McIntosh: Well if, your honor, if he undertook---uh I I think there'd be
a question if his goals were to get into the the marketing of this
software abroad [XXX] that's not---

Nelson: All he's taking with him is his head.

McIntosh: I'm s---pardon me, your honor?

Nelson: All he takes with him is his head and what's in it.

McIntosh: I understand, your honor. There are um restrictions on
providing so-called technical assistance to foreigners uh with the
purpose of bringing about the manufacture of controlled products abroad.
And it might well be---I don't claim expertise on this particular
area---but it might well be if he undertakes the kind of scheme you're
talking about, go abroad, uh help set up a company abroad that starts
cranking out these diskettes, that that itself would be technical
assistance within the meaning of that regulation. It's 15 CFR 744.9.
That might well be a controlled activity that needs a license.

Nelson: Let me see if if I understand uh how you get to your your
content-neutral analysis. Uh contrary to a famous philosopher you're
saying that the medium is not the message. You have to separate the
message from the medium. That here when when we're talking the---not
just the code but the the disk in which that code is embedded that
that's the medium and the government doesn't care what message is on
that medium, it's the medium it's concerned with.

McIntosh: Uh I think it's clear both from the President's um executive
order here and from the structure of the regulations that the
government's object is not---the government is not concerned with the
message. Um---

Nelson: So it doesn't---it doesn't help to start with cases that say,
uh, computer code is language, uh language is speech, speech is
expression, ergo the First Amendment leaps on this thing full blown.

McIntosh: The---I think that's absolutely correct, your honor. I mean
the point is that when the government's regulatory goals are unrelated
to suppressing that information that you don't simply apply kind of
full-blown---

Nelson: Let's let's say let's say that you're right. Uh you would agree
that uh in the face of a First Amendment challenge the government has
to defend the regulation, is that correct?

McIntosh: I'm sorry, your honor---

Nelson: That in the face of a First Amendment challenge, the government
has to defend the regulation.

McIntosh: Yes, I I think that's---

Nelson: It's not gonna lose it, it's got to defend it.

McIntosh: That's correct your honor.

Nelson: All right, so the battleground you've chosen is to say that it's
uh content-neutral, intermediate scrutiny, and in your view then it
survives intermediate scrutiny.

McIntosh: That's correct your honor.

Nelson: The problem I have with that---and tell me tell me where where
I'm I'm uh I'm off---it seems to me, if as the parties seem to agree,
that software that is is at least as capable as what we're talking about
here uh is fully available uh in Europe and other places and Asia then
what is the government accomplishing? Is what it's accomplishing here
really narrowly tailored to its view, when what it's doing is to some
degree at least uh hampering uh some elements of First Amendment
expression?

McIntosh: Uh let me say several things about foreign availability, your
honor. First the President gave very specific consideration to the issue
of foreign availability. And he determined that nothwithstanding the
actual or apparent availability of encryption software, and hardware for
that matter, abroad that uncontrolled exports of encryption products in
this country could still compromise this country's national security and
foreign policy interests in ways which are not happening as the as the
current state of affairs exists. Uh as this court's decision in the
Mandel case indicates, I think that kind of determination by the
President is one that is entitled to a very concrete deference. Beyond
that I think you need to bear in mind the exact nature of this
regulatory scheme. This is not a simple ban on the export of encryption
products. This is a licensing scheme. And all this scheme is designed to
do is to put the government in a position so it can evaluate on a
case-by-case basis---

Fletcher: Were there any particular standard that I have seen?

McIntosh: Well your honor, I---the the standard, the standards are those
associated with the national concer---national security problems posed
by foreign use of encryption---

Nelson: Can the test be standard?

McIntosh: Well, your honor, there there's something, um the nature of
the inquiry here, uh the nature of the of the administrative process, is
one that gets into matters that are for obvious reasons extraordinarily
confidential and secret. Um the the you [XXX]---you are familiar I'm
sure with the state secret privilege. And one of the things the state
secrets privilege applies to with greatest force is information about
this country's intelligence-gathering capabilities. I think it's
difficult in any sort of public forum whether it's administrative or
judicial uh for the government to get into any level of detail about---

Fletcher: That may be, counsel, but when you wave national security
against the First Amendment, the government has a higher burden to prove
its case.

McIntosh: Uh uh I I certainly don't suggest that merely---the mere
assertion of national security ends the case, your honor---

Fletcher: That's about all that happened as far as the court's knowledge
is concerned.

McIntosh: Well, your honor, I I I think, you certainly have before you
the President's own appraisal of the national security consequences. Um
you have your decision in Mandel, which reflects I think a legitimate
concern about the relative capacity of the courts and the executive
branch to make these kinds of national security assessments. And, as I
was trying to say a moment ago, you have a licensing scheme which doesn't
prohibit altogether exports of encryption programs, nor does it make
foreign availability entirely irrelevant. It leaves the Department of
Commerce with discretion in particular cases to take account of foreign
availability if that appears to be relevant to the national security
risks. It does so quite explicitly.

The real question is what are the consequences of the alternative regime
that the district court has tried to bring about, and that Professor
Bernstein is trying to accomplish, one in which there are _no_ export
controls on encryption software, regardless of the strength of the
software, regardless of the country to which it's destined, regardless
of the persons or organizations who are going to use it, and regardless
of the uses to which they're going to put it. I think the answer is, it
should I hope be obvious that there are legitimate reasons of national
security not to wind up in such a situation where the government cannot
exercise any control regardless of all those factors.

Bright: Does that uh put into issue the time, manner, and place
restrictions in this case, or if we decide that this is not a prior
restraint do you win at least at uh from the standpoint of this
litigation?

McIntosh: Well I think there there are two separate questions here.
There is the the prior restraint or licensing issue, and there is a a
separate although related question about the con---the standards
governing content-neutral regulations. They are the same standards that
would apply to a time, place, and manner uh regulation. So I think---I I
I don't suggest a simply one inquiry for you to make here. You you have
to inquire both into the content-neutrality issues that we've put
forward---

Bright: Are you saying that uh under time, manner, and place
restrictions uh the government's regulation is a valid one?

McIntosh: Under the standards that apply to time, place, and manner
regulations, which are which are the standards that apply generally to
content-neutral regulations, yes, it is our submission that the
government's regulations are valid.

Fletcher: Your argument seemed to say that if the government's purpose
was legitimate and pure that that should end it. And you don't talk
about the effect on---a chilling effect on First Amendment. For example,
there are all kinds of communications in today's world that need to be
encrypted, and they need to be strongly encrypted. And how does the
government respond to that chilling effect on First Amendment uh
communications?

McIntosh: Uh your honor, the the these regulations don't prohibit the
use of encryption by anyone.

Fletcher: Well but if you don't have the software how can you encrypt?

McIntosh: Well within this country anyone can purchase any encryption
software, and anyone who travels abroad from this country can bring with
them for their own personal use any encryption software they have
purchased in this country. There are specific provisions in the
regulations that provide for that. You don't have before you a statute
which purports to limit the use by American citizens of encryption
software.

Fletcher: So I could buy um Bernstein's source code, um, take it abroad,
and have it stolen from me in London?

McIntosh: Well I hope that wouldn't happen your honor. Uh but certainly
the risks posed by that uh random theft of your the one copy that's in
your uh briefcase uh are I think an order of magnitude different from
those that are presented if it is published or posted or otherwise
distributed uh to everyone in the world by the Internet. Uh and it is to
the latter kinds of risks that these regulations speak.

Nelson: I think I can I can detect some kind of a form behind your, the
government's assertion of of security. Let me see if if it's roughly
right. Uh let's uh let's assume that uh some expert looked at uh the
ability of people to encrypt information, from the dawn of time to
today, and decided that Mr. Bernstein on an arithmetic scale was about
at level 5, and no telling how high it'll go in time. If the government
then prohibits the export of Bernstein's level 5 program in direct
usable form by people overseas, then it's saying we aren't up to level 5
yet in our ability to de-encrypt uh this kind of information. So that in
your mind would be an admission of its uh intelligence-gathering
capability?

McIntosh: Well, I suppose some inferences can be drawn, your honor. Uh I
I wouldn't want to press them too hard. Um.

Nelson: I mean---is it that kind of a framework that we're talking
about? That that we know there are people in Europe that can do this
kind of work. And that the government is interested in the in the
information that they're encrypting, and is probably trying to decipher.
And so there's something here that if if if Snuffle is better than
what's in Europe then that gives some level of idea of what we can do
with European programs.

McIntosh: That that certainly is is the kind of inference that that
could be drawn, your honor. I don't---in a particular case it might or
might not be the the driving base of the decision, but it could be
drawn. Um well I I think I'd better leave it at that.

Uh I've exhausted my time. If the court has further questions I'd be
happy to answer them.

Fletcher: Well we'll hear from the other side and we then may have some
more questions for you.

McIntosh: Thank you your honor.

Cohn: May it please the court. My name is Cindy Cohn, and I'm from the
San Mateo law firm of McGlashan & Sarrail. We're here today representing
Professor Daniel Bernstein, plaintiff and appellee, and our job today is
to defend the district court's determination that the government's
cryptography licensing scheme is an unconstitutional prior restraint on
speech, in that it requires scientists to submit their ideas to faceless
government bureaucrats for evaluation, prior to publication, without any
standards or discretionary limits on the on the bureaucrat's actions. We
believe that this case can be decided with reference to four settled
Supreme Court precedents: City of Lakewood, Freedman v. Maryland, the
Pentagon Papers case, and Reno v. ACLU, which was decided by the Supreme
Court in its last term.

Bright: What was the first one you mentioned, I didn't catch it.

Cohn: City of Lakewood.

Bright: Oh that's the one I talked about.

Cohn: Which we've already discussed. That case, as as you you did
discuss with Mr. McIntosh, involved the licensing of newsracks, and a
licensing scheme that had no limits on the discretion of the licensing
bureaucrat to grant or not grant a newsrack license. And it set out to
test for whether a facial prior restraint could lie.

Bright: Does the record here show why Professor Bernstein uh did not
apply for a license, at least to find out whether he could get it?

Cohn: Well, the gov---the the determination that Professor Bernstein
would have to apply for a license was essentially the determination of
whether he would get one in this case. The government, although it's not
printed and although it's not bound to it, has a pretty consistent
policy of not granting licenses for strong encryption such as Professor
Bernstein's Snuffle. In addition it's our contention that he shouldn't
have to apply for a license here.

Bright: I understand your position is he doesn't have to apply, that
he's got a First Amendment right. Now um---

Fletcher: Your first argument, futility?

Cohn: Yes, your honor.

Okay. Um the second case is is Freedman v. Maryland. And that case laid
out the procedural requirements that the government must put in place if
it wants to institute a prior restraint scheme. Indeed, we we believe
that if the court finds that City of Lakewood applies, that there is a
close enough nexus to expression or conduct commonly associated with
expression here, and that the Freedman v. Maryland procedural
protections are not in place, which the government doesn't contest, that
it need go no further in its analysis. In FW/PBS, a recent Supreme Court
case involving a general licensing scheme for businesses that had
especially onerous restrictions for adult businesses, Justice O'Connor
writing for the plurality held that, once the court finds a facial prior
restraint and a lack of procedural protections, it need go no further,
following settled Supreme Court precedent that there is no need to go
any further once it---it need not find any further problems once it's
found a initial co---facial constitutional problem.

The third case and the fourth case are raised by the government's
defenses here. The Pentagon Papers case sets up the standards for proof
of harm, proof of causation, and proof of imminence that the government
must meet if it wants to const---if it wants to institute a facial prior
restraint on---a licensing scheme. And it's important to take a look at
that that---we think that opinion is very instructive here because in
that case, as the as in this case, the government was claiming a
national security interest. In that case, unlike this case, our nation
was at war. But yet the government---but yet the court held in the
Stewart opinion---well there are many opinions in the Pentagon Papers
case---but the Stewart opinion holds that the government must
demonstrate a direct, immediate, and irreparable damage to our nation or
its people from publication of of the material sought to be enjoined.
That's a high standard. And indeed the Pentagon Papers case took a close
look at the proof that the government presented in that case and held
that it didn't meet the standard. In this case there's no proof
presented in the record.

Fletcher: Has the government offered to submit proof under seal?

Cohn: It has not, your honor. Indeed they moved for summary judgment
below without making an offer of proof whatsoever so it's the---there
were cross-motions for summary judgment. They did not make an offer
proof below, other than the President's assertion and their assertion
which is at best throughout their papers that publication of this
material _may_ cause harm. That's exactly what the government said in
the Pentagon Papers, and the Supreme Court was clear that simply an
assertion that there may be some possible future harm from publication
of materials was not sufficient to meet the First Amendment standards.

And the final case is Reno v. ACLU. And we believe that it's in this
case that the Supreme Court's done much of of the work necessary in this
case. It took a look at the government's claim that it's okay to forbid
publication in one medium as long as you allow it in another. And
harkening back to City of Ladue, previous Supreme Court cases involving
media discrimination by the government, it said that the Internet is a
fully protected medium. And the fact of where where Professor Bernstein
chooses to publish his ideas, the Internet, should make no difference in
your analysis of this case.

The the government bases its arguments on some factual assertions here
that we think ignore some of the good fact-finding work that Judge Patel
did down below in the district court. The---obviously this is a facial
constitutional challenge, and this court isn't bound by the factual
findings of Judge Patel, but we think it's important to pr---that this
court take a look at at what what Judge Patel found in hearing, in her
opinion about the facts of the case. The judge---the judge found that,
the district court found that the regulations do have an impact on
academic speech, and that that speech is often undertaken by academics,
and that indeed these restrictions restrict the most common expressive
activities of scholars, and that includes teaching, and that includes
publishing.

Bright: Well, isn't this a case where the government is saying, the
function of source code which goes into the computer is such that if we
allow that source code into the computer and then it's it's converted so
uh the the message can go all over the world. And the government is
saying essentially we're looking at the function and what runs the
computer, and even though it may be speech-related it is not aiming it
at the speech itself, and says the reason we're not aiming it at at
speech itself is because there's so many other uh places that uh the
speech can be uh had and read and communicated. Now as as applied to
function alone, does that involve in your view a prior restraint, or can
that be construed to be content neutral?

Cohn: Well, your honor, the test for whether um a prior restraint facial
challenge can lie was set up in City of Lakewood. And and that test is
whether the regulations um have a close enough nexus to expression or
conduct commonly associated expression to present a real threat of
censorship. Um we think that, regardless of the government's aim here,
the record is clear, and the district court found, that these
regulations have a clear effect on expression. And so therefore a prior
restraint analysis must apply and Freedman v. M---

Bright: Well what expression are we talking about? The expression, is it
the code or is it cryptology, the ultimate message that would go out,
the secret message if I can use that term.

Cohn: Uh I I think it's both. Uh the code itself, for people like
Professor Bernstein and many people in the room here today, is itself an
expression. People who work in this area of science, people who work in
areas of science that are um assisted by computer programs, read those
programs, just like you or I might read a brief that might be Greek to
some person who's not had legal training. These people read that source
code---

Nelson: Some of them are.

[laughter]

Cohn: And some of them are, exactly. Um I hope ours weren't to you. But
um that that that the people like Professor Bernstein and his colleagues
all over the world read the code. They read the code just like we would
read read a brief. Code is published in books and has been for many many
years because people read it.

Bright: Why isn't the ability to read it in a book sufficient for
academic discourse. Why does it have to be on disk?

Cohn: Well I think that it's im---there---the question is, first of all,
does the government get to decide which medium Professor Bernstein
publishes ideas in, or not? Because there's no question that electronic
media is a legitimate and well-used medium for expression, just as paper
is. And does the government get to sit and tell Professor Bernstein you
can publish, because your ideas are on this subject, you can publish
them on paper, but you can't publish them electronically. The answer is
that that question requires stringent First Amendment analysis. And we
think of course the answer is no in this particular instance.

Um the---back to your your question cause I think that there was a an
um---you asked a second question which is is it the message itself that
is encrypted that is protected expression. We contend that that's that's
important too, and that the ability to send a message such that it can't
be read by anybody but your intended recipient is an important part of
the First Amendment, and that the ability of the government to restrict
the development of the technology that lets us have electronic envelopes
is itself imbued with First Amendment difficulties, and itself could be
a reason to trigger First Amendment analysis. We think there are two
two---

Bright: Well they're not trying to stop encrypted material. They're
trying to stop the computer from sending it out to the world at large.
And is that the same thing as a newsstand in the Lakewood City case,
City of Lakewood?

Cohn: Well the government is trying to prevent people from getting this
code---

Bright: Yes.

Cohn: ---because if they have this code they can use electronic
envelopes. Um I I think that the government's um---I think that this
is---I think the code itself is pure speech. It's a text. It's closer to
the First Amendment than the newsracks were in Lakewood. No one
seriously considered that the newsracks in Lakewood were themselves
speech. It was only that they they were part of the process of conveying
speech to readers. In this case the code itself is speech. And the
record is replete with examples of people who read code, who write code,
who develop it, and who who who have been chilled in their further
development of this particular kind of code because of these
regulations. And Judge---the district court's opinion re---reflects
that.

Nelson: If you meet the government on its own ground, in other words
it's brushed aside uh everything you you said, and said no, this is uh
content-neutral, intermediate scrutiny, we win. Uh if you accept their
premise and meet them on their own playing field, what do you think is
the basic flaw with the government's position?

Cohn: The basic flaw with the government's position is that they have a
licensing scheme.

Bright: I didn't hear that.

Cohn: The basic flaw with the government's position is that they have
chosen to implement their concerns here through a prepublication
licensing scheme. Like Mr.---Mr. uh McIntosh said that that it was
important to note that this did---wasn't a complete ban, that it was a
licensing scheme, and that there were government agents with discretion
to decide whether someone could publish or whether they could not. We
think that's exactly the point. That when the government sets up a
bureaucrat to decide who gets to speak and who doesn't get to speak on a
particular topic, they have to meet the very strict standards of prior
restraint.

Nelson: So the fact that it's a licensing scheme keeps it from being
content-neutral.

Cohn: No. I think it's content-based in addition. But you asked me for
the first problem and I I gave you the first one. Th---this is a
content-based scheme. This is a content-based scheme because it singles
out cryp---software on the subject of cryptography for especially
onerous treatment. If you look at the way the Export Administration
Regulations are structured as a whole, all other software on all other
topics is treated under the under the rubric of the phrase technology.
And the technology provisions contain pretty good First Amendment
cutouts. If you're involved in fundamental research, if you're uh---
people like Professor Bernstein's work on Snuffle would be allowed if it
were categorized as technology like all other software that's controlled
under the Export Administration Regulations. But in promulgating the
cryptography-specific regulations the government said if your software
is about this subject you don't get the First Amendment cutouts.

Nelson: Well I can I can see the government's arguments on why that
isn't so. I---not necessarily agree with them, but I can see their
arguments. But let's assume they're correct: in spite of your licensing
answer, this is a content-neutral scheme. Given intermediate scrutiny,
does it fail to meet the scrutiny?

Cohn: Yes it does.

Nelson: How does it fail?

Cohn: It fails intermediate scrutiny because the government hasn't shown
that its scheme directly and materially furthers its interest. And the
reason for that is the reason that you raised in discussing with Mr.
McIntosh before. The foreign availability problem, that there is
cryptography available all over the world. We have some really bright
cryptographers in this country but we do not hold a monopoly on this
knowledge. We do not hold a monopoly on the ability to take this
knowledge and turn it into functioning software. In the record there's
clear---there's clear evidence from people who have done surveys of the
cryptography that's available abroad and shown that extremely strong
cryptography is widely available abroad. The second reason that they
fail intermediate scrutiny---

Bright: Did the district court uh uh address intermediate scrutiny?

Cohn: It did not.

Bright: I didn't think so.

Cohn: It did not, because it found that it was a prior restraint, and
that Lakewood applied, and and following on the FW/PBS---

Bright: Well if if uh if you're uh. If if we should decide that there is
no prior restraint, and the issue now is uh the adequacy or inadequacy
of intermediate scrutiny, doesn't this case have to go back for that
kind of a determination in the distr---in the district court, rather
than our making the determination?

Cohn: I think that that that this court could send it back to the
district court for further findings. But I don't think it has to. I
think the record here is very clear, and indeed the foreign availability
is one problem in in meeting intermediate scrutiny test, but the second
one, the print exception, doesn't require any findings of fact, doesn't
require any factual knowledge. It's---there's no question, and the
government admits, that if---that Professor Bernstein could make
200---200000 copies of Snuffle, which by the way is two pages, it's 550
words---and take it on an airplane, and distribute it over Baghdad
today, and not be in violation of the export restrictions. But if he if
he were to take one copy of the program and electronically send it to
his colleague in New Zealand, he would be prosecuted as an arms
exporter.

Bright: So, from your standpoint, for complete affirmance we have to
determine the district court was right uh in ruling that this was a
prior restraint of free speech. Is that correct?

Cohn: Um I think you could affirm on different grounds if you if you
wanted to. I think the record is sufficient to allow you to affirm---

Bright: Well um we could but we wouldn't have to on the---

Cohn: You would not have to.

I think it's also important in considering this case to take a
look---the government's claiming that because---the government in
essence here wants you to create a new category of lesser protected
speech. And that's speech that can be read by humans and used by humans
but can also be functional if given to a machine. Now the government
conceded just moments ago that software can be used to convey ideas
without also conveying, controlling a computer. And that software that's
conveyed merely to other people to read is still controlled. Um we
submit that at best their argument triggers some of the intertwined
cases, the cases where there's a speech element and perhaps a lesser
protected speech element intertwined in the same speech. But we say that
analysis of those cases also leads you to prior restraint, full prior
restraint analysis. In this---in this court the [XXX] case involving the
selling of peddler's license for selling of t-shirts that had messages
on them um is a case in which the, there there was speech, there was
lesser-protected commercial speech involved, but yet this court found
that full prior restraint analysis had to apply. And because the scheme
at issue didn't have limits on the discretion of the agent---the the the
sherriffs, the agency bureaucrats who were administering it, it failed.

So we think that even if you---even if you accept the government's
argument, that that functionality or capability should somehow add a
lesser-than-speech element to the speech here, the cases clearly
indicate that you've got to do full prior restraint analysis. And
there's no question that this scheme fails prior restraint analysis.
There are no limits on the discretion of the agency bureaucrat. We saw
that in this case. Because Professor---as as Judge Fletcher noted
earlier, when Professor Bernstein first submitted his materials to the
agency, they told him his paper, his scientific paper, was controlled.
And it wasn't until after we sued that they issued a clarification
saying that they didn't really mean that. Um so the ab---the abuse of
discretion is clear in the record here. There are several other
declarants in the record who can talk ab---who talk about the abuse of
discretion, and and uh the chilling effect that this scheme has on them.
That's why---that's why prior restraint analysis is is is more is is
applied in licensing scheme cases, because of the chilling effect,
because it stops speech before it happens. That's how this case is
different than the Roulette case that the government cites over and over
again. The Roulette case involved uh a a scheme that criminalized
sitting on sidewalks. But there was no question that there would be
judicial review of each of these determinations, because a person would
be prosecuted. It was a subsequent punishment scheme. So we submit that
Roulette is not the correct place to look for the proper analysis. The
correct place to look is previous licensing schemes.

If the court has no further questions I'll sit down.

Fletcher: This is an odd case in that uh we say that the essence of
prior restraint is that you mustn't stop speech before it happens. We
know what the speech is here. It's kind of looking at the effects of
stopping it.

Cohn: We do but this---that's part of why this is a facial challenge,
because there are other people out there who haven't had the benefit of
four years of pro bono representation in the courts, who would like to
publish their ideas, who are chilled from doing so. And we won't know
about them, you'll never know about them, unless they're lucky enough to
to to have the the benefit that that Professor Bernstein's had. That's
why it's important that this is a facial challenge and that it be
decided as a facial challenge, because you'll never see, you'll never
see the people who are chilled by this. And the record's clear that
there are others out there who are.

Fletcher: We'll give you a few minutes, counsel.

McIntosh: Thank you, your honor. Uh Ms. Cohn said that this regulatory
scheme puts the government in a position to decide who can speak and who
can't speak. If I accomplish nothing else this morning, I want to make
clear that that is absolutely, utterly incorrect. This regulatory scheme
doesn't stop anyone from speaking, including Professor Bernstein---

Fletcher: It does---it does prevent them speaking in certain mediums.

McIntosh: It does in a particular medium, your honor, the Internet, that
I guess in his in his instance. But the the the critical point for prior
restraint analysis is there's no attempt here by the government, and no
capacity by the government with these regulations, to keep Professor
Bernstein's ideas bottled up inside his head or anywhere else. There are
any number of ways that he can distribute those ideas to anyone he wants
to for academic purposes, which is his stated purpose. Now---

Nelson: Except on the Internet.

McIntosh: Except on the Internet. There's a particular medium that's
being restricted here. [XXX] And so of course---

Nelson: If he writes a letter to the to the editor of the New York
Times, Unabomber style, and puts his [source code in the letter], he can't
do that because they put out an Internet edition.

McIntosh: Well no he he wouldn't certainly be subject to any
restrictions, your honor. And they could publish it, the New York Times
could print it. Now if they posted the source code on the web that would
be subject to export controls. But what's important, your honor---

Nelson: Don't they---don't they have an edition that goes on the web?

McIntosh: They do, your honor. I don't know whether it's comprehensive.
But they certainly do.

Nelson: So so I guess my point is that that he could write to the Boise
Statesman and be perfectly safe as to the the Boise Statesman. If he
writes the the same letter to the New York Times, somebody's in trouble.

McIntosh: Well he's not in trouble, your honor. He's not in trouble.

Fletcher: Even if he knows it's going to go on the Internet?

McIntosh: Pardon me, your honor?

Fletcher: Even if he knows that it's going to go on the Internet?

McIntosh: Uh um I'm not sure how he would know that, your honor.

Fletcher: Well let's assume that. He knows that the New York Times puts
all these things on the Internet.

Nelson: Hey I've read it on the Internet so I assume he might know that
too.

[laughter]

McIntosh: Well if the New York Times were undertaking to to engage in
conduct [XXX] export of source code then---and he was aware of that then
that might have---he might be subject to regulation in that regard. But,
your honor---

Nelson: Would the New York Times be subject to regulation?

McIntosh: The the export of encryption software requires a license as a
general matter. And therefore the posting in some unrestricted forum of
encryption source code on the Internet is subject to a licensing
requirement. If if controls are placed on the avail---on access so that
it's limited to domestic access then there's no export and therefore no
licensing requirement.

Nelson: So that would have to be policed after the fact?

McIntosh: I'm sorry, your honor?

Nelson: That would have to be policed after the fact? I assume the the
New York Times doesn't have a licensing bureau.

McIntosh: I I I I assure you your honor the government does not spend
its time reading the New York Times online to try to bring bring
prosecutions for the publication of electronic uh source code. Uh---

Nelson: Doesn't that make that make the the scheme even less effective?
In other words, if if uh Bernstein wanted some help all he had to do was
write to the Times and say if you publish this uh you're going to jail. 
And I can assure you they'd---he'd get lots of help. Wouldn't be pro
bono.

[laughter]

McIntosh: I think he's he's he he has done very well by his pro bono
counsel, your honor. I don't think he needs the Times's assistance. Um
but if I could just pursue this a little bit further. We've been talking
about different avenues for communicating information, and about the
effect of restricting a particular avenue of communication while leaving
other avenues open, press, print, other avenues. Here too the
distinction between content-based and content-neutral regulations is
absolute critical for you to keep in mind. What the Supreme Court has
said in Ward, and what your honor Judge Fletcher said last year in a
case called Bland v. Fessler [79 F.3d 942 (9th Cir. 1996)], is that when
you have a content-neutral regulatory scheme, a scheme that is not
directed at the at the s---suppressing information, that the existence
of alternative avenues for disseminating information are very
significant. Recall Bland v. Fessler was a case involving automated uh
phone solicitation. And the government pointed out that, while
restrictions were placed on these automated uh phone solicitation
devices, there were other ways you could engage in phone solicitation or
more direct solicitation. And the plaintiffs in that case said, but but
this is the best way, this is better than the alternatives, this really
affects us, and the alternatives aren't as good. And under intermediate
scrutiny what this court said, and what the Supreme Court has said in
Ward, is as long as there are ample alternative avenues for
communication it doesn't get you anywhere simply to say that this one is
preferable from your standpoint.

Fletcher: Does the fact that it was commercial speech we were talking
about have anything to do with it?

McIntosh: There was more than com---I believe, your honor, more than
commercial speech at issue in that case. Uh I could be wrong about
that but I---

Fletcher: It was somebody who had a carpet business, and he wanted to
advertise his cleaning.

McIntosh: Uh there were uh again---I could stand corrected in this, your
honor, but I think there was more than that in the case. And in any
event the standards that were being applied, the intermediate scrutiny
standards, are the same ones that would apply to any content-neutral
government regulation.

Bright: On intermediate scrutiny I've asked uh uh your opposing counsel
since the district court didn't address it. What do we do with that
issue if we find that uh this regulation is content-neutral but subject
to the intermediate scrutiny analysis. Do we send this back, or do we
have to decide it?

McIntosh: I I I think this is the one of the rare points where opposing
counsel and I are in agreement. Uh it would be within your power to send
it back, um but it would likewise be within your power to decide it
yourself. Uh any review [XXX] subsequently engaged in if there were a
decision by the district court would be de novo, uh and the questions
that are presented are substantially questions of law. I don't think
there are genuine issues of material fact in this case.

Um now one final point. Judge Bright you asked whether this case,
whether the speech that we're dealing with here is the code itself or
the messages that are being encrypted. Um the district court's decision
rests on the source code itself, that is the speech that the district
court was concerned with, not the messages that are being encrypted by
software. And indeed, this also goes back to one of Professor---um
pardon me, one of Judge Fletcher's questions. Uh Professor Bernstein has
never asserted that he himself has been impaired in any way in _his_
ability to engage in encrypted conversations. As far as I know there's
nothing in the record to that effect. And so this simply isn't a case
for that reason as well as others that presents that question whether---

Fletcher: Well, uh, if in fact the government's fears are realized, it
is in effect preventing others from, uh free to conversing---with lots
of encryption.

McIntosh: Well the the government's object your honor is to preserve its
ability to maintain electronic surveillance of foreign intelligence
targets.

Fletcher: Right.

McIntosh: Uh there's no there's no attempt to to restrict at all in
these regulations the use of encryption of whatever strength, whatever
capabilities, domestically.

Nelson: Well if Coca-Cola wants to send an encrypted message to, from
Atlanta to its uh its European headquarters in Paris, wherever it is,
it can't use this this code system because it can't get the decoding
disk to Paris.

McIntosh: Okay. A license would be required for export. Now we we don't
know, your honor, because there's never been a license application here
um whether a license---

Nelson: Without---without getting a license it can't do that.

McIntosh: Yes. Any any export of encryption software generally speaking
requires a license.

If the court has no further questions.

Fletcher: Thank you, counsel. Uh the case we have just heard is
submitted. The court stands adjourned.