One of the “givens” in Josh Blackman’s effort to prevent various federal judges from issuing a temporary restraining order against the publication of the CAD code for 3D guns is that code is speech. Georgetown lawprof Marty Lederman rejects the notion, calling it a “red herring.”
The “freedom of speech” red herring
Defense Distributed challenged the State Department’s Internet-posting ban as an alleged abridgment of its freedom of speech. Its theory is based on the idea (as it argues in a brief) that its computer code “is expressive in that it can be read and edited by humans, who can also understand and adjust its output.”
It may be right about that fact: Some small number of human beings can in some sense “read” and “edit” this code, just as there are some who can “read” most computer programs; the code might even have certain properties associated with a language.
By calling those capable of reading code “some small number of human beings,” Lederman not only trivializes their numbers, as there are a great many people who read, write and understand code, even if few of them end up teaching law school, but trivializes the point. Is it only “speech” if it’s spoken to a full auditorium? It makes no difference how many people receive the speech. It’s speech because it communicates, not because any particular number of people hear, read or understand it.
Yet, this isn’t remotely the point. Some code causes the Constitution to appear on our computer screen. Some code causes Lederman’s law review articles to appear at SSRN. Some code makes a cute kitteh pic show up. And some code provides the means to cause a 3D printer to create a gun. It’s neither more nor less expressive because of the outcome of the communication, any more than verbal instructions that serve one purpose are less speech than that which serves another.
But Lederman shifts to the “purpose” of the speech, apparently unaware of Wobbly songs.
That’s not the primary purpose of posting it to the Internet, however: Presumably Defense Distributed expects and intends that 99.99 percent of the people who download it will not “read” or “edit” its code but will instead simply do with it what we all do with computer files every day, namely, stick them or download them into a computer so that they can perform the technological functions for which they were designed—in this case, to create operational weapons at the click of a mouse, without conveying any information to anyone.
The shift here seeks to narrow the read/write continuum by using intent to disclaim the nature of speech. For anyone who has ever watched a Youtube video on how to fix one’s car, the fallacy should be clear. Most people who “stick” this code into their computers won’t be making guns. They can’t. They don’t have the capacity to do so, and don’t actually want a crappy plastic gun.
But the communication is that they can say, “we have the technology.” Then there’s the communication upon which others will change, improve, the message. Maybe someone will use this CAD code to write code that will provide the means to save a life that Lederman values? But that doesn’t occur to him yet, as he’s stuck on the nefarious use of the code.
More importantly for constitutional purposes, the government’s reason for regulating the distribution of the code is not to suppress any potential “informational” value in rare cases where a recipient might “read” it. It is, instead, to prevent persons overseas from obtaining a tool of production the physical properties of which cause the easy creation of non-exportable weapons. (This is not, in other words, a restriction on speech because it (as Noah Feldman put it) “instructs the public how to commit a crime.”
The flat world of the internet allows things that the government would otherwise wish to prevent. Who knew? We say it here, and somebody in a far off land gets to see it because they have google and wifi too. Then again, it’s not as if people in other countries, say Russia, possess the mad coding skillz that would allow them to kick our butts in, say, hacking. Or creating a CAD of 3D guns. And if it’s expressed here, and read there, how then can the government execute its duty to restrict the information so it doesn’t fall into the hands of people who intend to do us harm?
This is a far better argument than trying to spin code as not being speech. Nobody wants terrorists bringing undetectable guns onto planes to do harm. Sure, there are other communications out there, the bomb-making instructions in the “Anarchist’s Cookbook,” for example, because as much as there is a potential for its use against us, it remains speech. And as such, it is protected by the First Amendment unless it falls within a prohibited class of speech.
The concerns raised by Lederman, as well as a great many others, about the potential spread of 3D guns may be grossly overstated, but they are still concerns. The argument that a gun whose primary feature (and many far more substantial deficits) is the ability to pass undetected through a magnetometer seems to serve only purposes that can do harm. And to ignore this concern is to be in denial, even if some will inevitably argue that this feature could also serve those who believe that guns in the hands of Americans serve as a check on governmental abuse.
But to deny that code is speech isn’t the way to deal with the potential harm 3d guns could cause. If anything, it exposes code to all manner of mischief, censorship and abuse that will almost assuredly be regretted as the government seized upon the idea as a way to prevent the spread of disfavored ideas and communications.
Does this present a quandary, given that it protects speech that could, a couple steps down the line, cause harm? Sure, bearing in mind that it’s not the speech that does harm, but the downstream creation of a plastic gun (without a piece of metal as the law requires) that ends up in the hands of evil doers and is used to harm others.
Unfortunately, there is quite a bit of information available that serves similar purposes, as well as benign if not beneficial purposes. Dealing with it is hard, but denying that code is speech is not just ridiculous, but counterproductive and likely dangerous.
Update: In an update, Marty has generously sought to re-explain, maybe for my benefit (?), why he’s not saying code isn’t speech, but that this code isn’t speech. To make his point, he offers this explanation.
This part of Greenfield’s post demonstrates the point nicely:
Some code causes the Constitution to appear on our computer screen. Some code causes Lederman’s law review articles to appear at SSRN. Some code makes a cute kitten pic show up. And some code provides the means to cause a 3D printer to create a gun.
Let’s play “Which of These Things is Not Like the Others?”
Yes, of course code not only can itself be expressive, but can also facilitate other speech (just as ink and paper can do).
Now I appreciate videos as much as the next guy, but my sincere efforts to see a cognizable point here have been without avail. If code is speech, and this is code, then this is speech. It doesn’t become nonspeech because it can be used (or not) for nefarious purposes (or not).
“Some small number of human beings can in some sense “read” and “edit” this code…”
Those Arapahoe and Cherokee had better be careful what they say.
This was just his way of trivializing the problem so as to make most people shrug at the “loss” of speech. It’s unprincipled and irrelevant. It’s also false.
Braille also comes to mind. Or complicated math. Lederman has not put much thought into this.
If we outlaw Chebyshev polynomials, then only outlaws will use Chebyshev polynomials.
Wait, that might actually be a good idea.
Vive Laguerre!
https://www.youtube.com/watch?v=jm7cez97528
(perp-walked out the side door for brandishing a bad unregistered pun)
‘This is not, in other words, a restriction on speech because it (as Noah Feldman put it) “instructs the public how to commit a crime.”’
…and since when is that (even if true, which it isn’t) relevant to the First Amendment? Advocating for, or instructing in, criminal action is still protected under the First Amendment, so long as it doesn’t rise to the level of incitement. Even Paladin Press’ infamous “how to be a hit man” book wasn’t pulled for criminal reasons, but for liability reasons.
“a gun whose primary feature . . . is the ability to pass undetected through a magnetometer”
No, the primary feature is that anyone can make it at home with (relatively) common and very inexpensive technology. The ability to pass through a magnetometer is secondary.
Your first point is valid. Heck, a baseball bat swinging at a ball thrown with great velocity is called our national pastime. The same bad swinging at a person’s head is called attempted murder. Teaching someone to swing a bat isn’t the problem.
Your second point, on the other hand, isn’t accurate. You can’t make a viable 3D gun on a home printer. There are ten types of 3D printers now, very different in their approach, and it’s questionable whether any can reliably produce a gun that won’t blow up in your hand.
“Teaching someone to swing a bat isn’t the problem.”
No, of course not, but that can (as you point out) be used for good or ill. But even advocating for, and instructing in, explicitly criminal activity (like, say, the violent overthrow of the government) is protected by the 1A. Incitement is a 1A exception, but the bar for that is pretty high as you know.
On the second point, though, I don’t think the questionable viability of this weapon really changes the analysis (after all, if it doesn’t work at all, what does it matter if it goes through the detector undetected?). The point is that the technology is, if not here yet, at least very close, to allow more-or-less ordinary people to make their own guns (or what would otherwise be the regulated parts of guns, like the lower receiver of an AR15-pattern rifle), from scratch, at home. If that’s the case, it makes plain the point that gun-rights advocates have been arguing for years: gun control will be no more effective than Prohibition.
One of the things that makes the hysteria around 3d printed guns so ridiculous is the fact that the technology to allow more-or-less ordinary people to make their own guns has been available pretty much since guns were invented. Guns are not that hard to make.
The only remotely unique quality of 3d printed guns is that someone took the time to figure out how much and what type of plastic is needed to make a gun that will probably not blow up when fired.
I think, SHG, there’s a more direct attack on Lederman’s concern that “Some small number of human beings can in some sense ‘read’ and ‘edit’ this code, just as there are some who can ‘read’ most computer programs.” And I think it’s more compelling:
When I take such a file and load it into a CAD program, I can visually view the representation without needing the ability to understand the code per se. The program interprets it for me and turns it into a visual diagram. And then I can grab a corner and move it around and change the shape of the object and save the results in a new file of CAD code. The code is the language that expresses the drawings, when translated through software.
(Note that I haven’t read the briefing in this litigation, so my understanding of the facts may be imperfect.)
To give an analogy, this is like saying a PDF file of an article, essay, or law brief is not speech, because only a small number of human beings can look at the raw PDF file that starts out:
%PDF-1.7
%
2 0 obj
<>stream
and understand the text that’s hidden inside there. But that doesn’t mean the code (here PDF) is not a representation of speech, and therefore itself speech.
A good point, which goes to my analogy to code manifesting as the Constitution on your screen.
There’s also the fact that “only a small minority of human beings” can understand the notation of, say, Gaussian differential geometry, or Demotic script. Indeed, if we were to limit freedom of speech to things Lederman can understand, our intellectual discourse would necessarily become extremely impoverished.
This was argued 20 years ago, when the government went after the guy who wrote PGP, the encryption program. They argued that the program was a munition, under the ITAR, and needed to be licensed for export/publication on the internet. I was working for the government and needed to exchange some proprietary information with my counterparts in the Italian government. I wanted to use PGP, but was told that I could not, without an export license. We eventually ended up sending 3 people to Italy to install the software there, and I believe that they eventually caved on the ITAR argument. I will go look it up
Tech raises a constant stream of problems that don’t fit neatly into our physical world conception of solutions. Each new one seems to be addressed as sui generis rather than merely the latest in the continuum of problems. We have yet to squarely face the reality and come up with principles to deal with it, and so the battles keep repeating themselves.
Well, it seems to me that the govt tries really hard to avoid having these cases resolved in a court of law. The PGP cases resulted in changes to the ITAR to make them moot, although two appeals courts made the initial determination that code was speech. And in this case, it seems that the govt finally caved, in order to avoid a showdown in a court about the matter. Our leaders make these wonderful laws and regulations, but then they don’t want to subject them to the scrutiny of a court.
Sometimes yes. Sometimes no. As with all grossly simplistic assumptions, it depends.
The government did eventually give up. In that case, the PGP source code was published in a book, which ITAR did not apply to, which was then shipped to other countries where each page of the book was scanned and the scanned source code compiled.
The PGP cases were Karn v State, much of the filed material is at http://www.ka9q.net/export/
When Karn and friends turned the source into a book and asked the government to explain why the book and the disk were treated differently under the export control regs, the State Dept chose to rewrite the regs to moot the question.
I’m wondering how long until there are calls for In the Line of Fire to be banned, considering John Malkovich shows quite clearly how to manufacture an undetectable gun, as well as how to get bullets past a metal detector. Which is the other part of this that no one has mentioned: sure the gun may be undetectable, but what about the bullets?
Aluminum, copper, lead, tin, titanium and zinc, and alloys such as brass and bronze are non-magnetic. Bullets quite commonly are cast of lead and it is certainly possible to produce cartridge casings from copper or brass (or, for muzzle-loaders or early breach-loader systems, even from paper).
Metal detectors do not rely on the presence of ferromagnetic materials. They rely on induced currents. Any sufficiently-conductive material (i.e., any metal) can be detected.
Typical metal detectors don’t rely on ferromagnetism to work. They work by inducing electrical eddy currents in conductors (like metals) which in turn produce their own magnetic fields. This works quite handily on non-magnetic metals.
Which is good, because a lot of knives are made entirely of stainless steel, which is generally also non-magnetic.
Of course, projectiles don’t really have to be metal at all.
They do, however, have to be dense. The only way you’re going to get a plastic with a specific gravity rivaling even steel (forget lead) is to use a metallic filler, thus defeating the point.
Shaped stone pellet with a hard plastic coating to travel down a gun barrel without damaging said barrel. The plastic layer can be just thick enough to coat the stone and engage the rifling. Possibly not conveniently mass producible but someone who would want this wouldn’t be concerned about mass production. The coating could even be done with a 3D printer although molding would probably be easier. Use this in a firearm designed for caseless ammunition and you are good to go, at least from an ammunition standpoint. Designing the gun is an exercise I don’t have time or desire to engage in at the moment but using caseless makes it simpler as you can use a breech-load design with no need for spent cartridge removal.
Now that I’ve put this blog on the watch-list and made SHG nervous about me I’ll go for the rest of the day.
It’s my fault for allowing this idiotic tangent to go so far. I should never have let it happen. This is the end of it. Comments are now going on moderation with extreme prejudice. Enough nonsense for one day.
Or the cryptic code of real property in metes and bounds or the lot and block survey system. Not very dissimilar to the code that drives these 3D printers.
Even less people can read Egyptian heiroglyphics.
Defense Distributed also sells a CNC machine that can mill AR lower receivers and 1911 frames. These are the once cost prohibitive machines that firearm manufacturers currently use. They require code as well.
It could be the exact same code, in fact.
This is how we make prototypes in industry. The same machine will make them out of ABS, with the same code.
So, is this just about 3D printer code? Or is it 3D printed firearm code? Or is it CNC code? All code? This is messy.
Hopefully Lederman is confined to classroom con law.
Rxc, Ken and Adam,
There were three more or less related cases. Adam mentions Phil Karn’s case, but the other two, Bernstein v. DOJ in the 9th Circuit, and Junger v. Daley in the 6th Circuit, both were broader challenges than Phil’s. (Neither of the other two involved PGP itself, but both involved encryption software code regulated by ITAR, then EAR.)
The final dispositions of each varied according to circumstances, but in each, the courts answered in the affirmative the threshold question of whether encryption software code was a form of speech, thus subject to a First Amendment prior restraint analysis.
For the reasons given above, and many more, it’s pretty clear that the suits by the Nine AGs are futile. They can’t stop the the eventual spread tech info, besides which it’s not more dangerous than other problems like illegal purchases and distribution of regular handguns. So, it seems the biggest effect of suits is PR. What I don’t understand is whether the AGs know they are basically engaged in a PR/political exercise, or if really think they’re the good guys riding to the rescue (which seems to be an occupational hazard in the AG business).
What the fuck does this have to do with the post?
Scott: Just for the record, in case your readers haven’t read my post: That (“this code isn’t speech”) is not what I wrote, either. Thanks.
Hopefully, anyone who reads my post will read yours as well.
01011001 01101111 01110101 00100000 01110011 01100101 01100101 01101101 00100000 01101100 01101001 01101011 01100101 00100000 01100001 00100000 01101110 01101001 01100011 01100101 00100000 01100101 01101110 01101111 01110101 01100111 01101000 00100000 01110000 01100101 01110010 01110011 01101111 01101110 00100000 01100110 01101111 01110010 00100000 01100001 00100000 01101100 01100001 01110111 00100000 01110000 01110010 01101111 01100110 01100101 01110011 01110011 01101111 01110010 00101100 00100000 01100010 01110101 01110100 00100000 01111001 01101111 01110101 00100111 01110010 01100101 00100000 01110111 01110010 01101111 01101110 01100111 00101110
There are 10 kinds of people in this world. . .
I agree. Heheh.
01001101 01100101 00101100 00100000 01110100 01101111 01101111 00101110
01010111 01100101 01101100 01101100 00100000 01110011 01100001 01101001 01100100 00101110
It has just occurred to me that there are some firms that are selling ceramic knives, which contain no metal. These knives are very sharp, and some of them that are sold in places like Aldi cannot be detected by an airport screener. Why haven’t we seen a movement to ban them? The Brits are starting to argue that knives with sharp points should be banned – why not also ceramic knives, if they can serve as “ghost knives” and be used for nefarious purposes.
For that matter, why not big heavy books, too? You can certainly hit someone in the face or on the head with a large dictionary, or kill them with a blow to the neck. Large books should be banned, too. Only thin paperbacks, made out of flimsy tissue paper, should be allowed. That would really drive society forward. (/sarc)
Knowing a little about tech: the files that make up the Blu-ray disc of Infinity War you just preordered are much more ‘code’-like than these gun files, which are really more like blueprints. As someone else already pointed out, PDFs aren’t documents, they’re fully fledged programs. It’s a great analogy: PDFs descended from PostScript, which is a language for telling a laser printer how to print a document.
With all due respect to Mr. Lederman, I don’t think he has an argument, and instead is just flailing about grabbing at anything resembling an argument in an attempt to justify his opinion. Because GUNS ARE BAD.
I edited out the first sentence of your comment because it was completely idiotic and I will not have this post take yet another dive down another nonlawyer rabbit hole.
Thy rod and thy staff, they comfort me.
Thanks.
SHG–I suggest tomorrow’s post be about cooking. Everyone knows about cooking. It would be fun to discuss copper v. iron skillets. Maybe you could mention doughnuts? Or is it donuts?
Et tu, Skink?
Cast iron and gas cooktops are the only way to go.
Delicious pastrami only gets you so far.
Ok, I’ll be good for the rest of the week.
I’d settle for the rest of the day.
Considering that it’s Friday and I don’t usually get on much on the weekends. Hmmm…?
Wait. Today is Friday?
The argument is being distorted by guns, and gun control. If it was computer code for anything else the case law from PGP should make it a settled issue. Instead we have full blown tantrums and predictions of the collapse of society if evil plastic guns aren’t blocked by any means necessary. This is bringing out the same rhetoric as the claims that the 1st amendment is being “weaponized” because the unfavored group is benefitting
As I see it the 3d printing is just a convenient pretext to ban home gunsmithing and increase government control.
The practicalities of 3d printing guns makes them more of a proof of concept than a serious tool, but the rhetorical bogeyman makes an excellent pretext to ban “80%” receivers and unserialized guns thus facilitating the long term goals of registration and confiscation. The anti-gun groups are more concerned with banning guns than any constitutional issues, just as the extreme left is eager to censor speech to gain temporary advantage, regardless of potential long term harm.
As a technical aside, cartridge casings are usually brass, but steel cases are common for lower cost, pistol cartridges are sometimes aluminum to save weight and some plastic cased ammunition is on the market. Also Philip Luty has a book on improvised ammunition to use in your improvised gun, and Luty’s work is plain old written word, which has clear and strong free speech protection.
As a technical person I’d say there is a fine distinction involved here. Code is only speech while human beings are interpreting it. Once code is running on a computer, code is a set of binding instructions which the machine ABSOLUTELY MUST follow. The computer has no power to critique the code, make ethical decisions about the code, suggest variations on how the instructions should be followed, etc. The process is essentially mechanical, and the code is part and parcel of the machine. The code might as well be a gear, pulley, or belt, and there is nothing remotely like consciousness happening at all when a machine runs code.
Whether a judge can be induced to understand this fine distinction and it’s implications is another matter, of course, one more suited to your skills than mine.
It’s the CAD files they blocked. The files convey design intent. The same thing could be done with ink on a napkin These were not executables.
I know. I didn’t want to get into the difference between data and code because that was not my major point, but that’s definitely a nit work picking.
It’s kind of a big nit. These data sets are the language of the design engineer. The information and ideas communicated in these data sets direct trillions of dollars of human activity. References to these data sets are incorporated into contracts and purchase orders. These data sets are considered legally binding by those engaged in this commerce.
The difference between data and code is an important technical point because Leaderman incorrectly argues that a cad data set that is comprised of a 3 dimensional coordinate system is not really a data set at all. 3D Cad data of a firearm becomes “gun-creating software”.
Exactly. The CAD data is not code, and many different kinds of code, written in many different computing languages, could interpret that data to a 3D printer.
To the blog owner: This is a big deal to the tech community as judges and lawyers frequently don’t understand what technology is doing sufficiently to make good rulings or good pleadings on the subject and there have been some pretty bad decisions made; wrong not on the law, but wrong because the level of technical inaccuracy and technical misunderstanding is so high.
This is why I would like to hear something from you on the matter beyond “get off my lawn.” It’s a pretty important issue to us techies. (Just for the record, I’m neutral on the issue of guns.)
Oh. You thought this was all about you? Damn, are you going to be disappointed. So here’s the deal. This is a law blog, not a blog about whatever the fuck you want it to be. If you don’t like it, get lost. If you want to have a discussion about something tangential to my post, start your own blog and write about it to your heart’s content, but who the fuck do you think you are to tell me what you want me to do on my blog to accommodate whatever shit is floating in your head?
And even if, just if, you wanted to try to make your point in the context of this post, try doing so in the context and manner it would be informative to readers here rather than what you feel like spewing. Maybe part of the problem with lawyers and judges being so ill-informed about technology is that people knowledgeable about technology are incapable or disinterested in communicating what we’re not getting in such a way as it can be incorporated into what we’re doing. And bite me.
I’m interested in learning more about tech so my decisions are accurate, useful and fully informed. You haven’t helped, and you seem to have no clue why your comments have been unilluminating to lawyers and judges.
Perhaps you should put more effort into communicating your concerns to lawyers than whining that you’re not being treated as the center of the universe. And check your entitlement. You’re a guest here.
How does the distinction between who (a computer vs. a human) is interpreting speech have any bearing on the speech itself?
I’m not sure it does, but consider the following. I get some friends together and use speech to discuss my plan for robbing a bank. Since my friends can both interpret speech and critique speech, they either try to modify/improve the plan, or decide for ethical reasons that no plan to rob a bank should take place. Or they express a fear of police and prison as their reason for not agreeing to my plan. Maybe they suggest better means of raising money or offer me a loan.
A computer just goes out and robs the bank. To the computer, my code (legally “my speech)” is instantly and thoughtless obeyed. The computer just goes out on the network, exploits a flaw in the bank’s security, and brings me a billion dollars. There’s no interpretation of my speech, no attempt to improve my ideas, no ethical objections. The computer just does it, and once again, there is no thought about the subject whatsoever.
Obviously there are huge and profound differences between how a humans versus computers deal with “speech,” and I’d imagine this fact has major legal interpretations, but what they are (and whether they exist) aren’t my field of expertise, so perhaps Our Gracious Host will offer us an opinion on this issue. Meanwhile I’m going to drop the subject as I don’t want to try OGH’s patience.
It’s not OGH and you’re too late. Your comment was irrelevant to begin with, and now it’s just another I should have trashed in the first place. But I’m too nice, and no good deed goes unpunished.
“Blog Owner.” lol. This may be the best comment ever.
There are approximately 3.6 million software developers in the U.S. Hardly Lederman’s “small number of human beings” who can read and write code. There are approximately 1.2 million lawyers in the U.S., presumably a very “small number of human beings” under Lederman’s scale, who can read and write the “code” of law. Statutes and case law “instruct(s) the public how to [lawfully] commit a crime.” Or not. Given Lederman’s position, he has convinced me that — if we shop for the right federal judge — the entire practice of law might soon be deemed to be unconstitutional. And those once-lawyers might well become software developers, who will learn how to create 3D gun code.