Copyright (c) 2003 Albany Law Journal of Science & Technology Albany Law Journal of Science & Technology 2003 13 Alb. L.J. Sci. & Tech. 273 LENGTH: 63502 words ARTICLE: COMPLICIT PUBLICATION: WHEN SHOULD THE DISSEMINATION OF IDEAS AND DATA BE CRIMINALIZED? Susan W. Brenner* * NCR Distinguished Professor of Law & Technology, University of Dayton School of Law, Dayton, Ohio. Website: http://www.cybercrimes.net. SUMMARY: ... This article is about controlling speech; more precisely, it is about the advisability of using criminal prohibitions to control the dissemination of ideas and information. ... The content of theories, ideologies and philosophies is abstract, and is therefore exceedingly unlikely to become the actus reus sufficient to support the imposition of criminal liability under the "speech act" doctrine. ... Since notional ideas do not lend themselves to prosecutions under a "speech act" approach, the only way to impose criminal liability for their dissemination is to make the articulation of certain types of notional ideas a crime. ... It is exceedingly unlikely that the symbolic expression of notional ideas in fiction, music and/or art could provide the basis for imposing criminal liability under the "speech act" doctrine. ... That is, personal information has negative value because it can be used to locate and harm someone. ... How is all this relevant to vulnerability reporting? This analysis is concerned with the extent to which criminal liability can be imposed for the dissemination of certain types of speech. ... It is this variation in how societies approach the criminalization of speech that creates the scenario described earlier, in which the owner of the Justice Files website moves it to a jurisdiction where its content, which is illegal in Washington, is lawful. ... TEXT: [*275] I. Introduction "Editor-in-chief of the official German news agency ... [and] head of the Wireless News Service ... . Fritzsche used the foregoing positions... to advocate, encourage, and incite the commission of the ... Crimes set forth in ... the Indictment ... ." n1 This article is about controlling speech; more precisely, it is about the advisability of using criminal prohibitions to control the dissemination of ideas and information. n2 This is not an article about the First Amendment, though it incorporates First Amendment principles into its analysis. Instead, it is concerned with how principles of substantive criminal law devised to deal with conduct occurring in the "real," physical world can, and should, be extrapolated to cyberspace. Cyberspace eliminates the pragmatic "filters" that have for so long constrained the dissemination of ideas and information; n3 it gives rise to new modes of communication and thereby raises difficult questions about the extent to which speech can, and should, be criminalized in American law. These are not merely national concerns; the capacity that cyberspace creates for the unfiltered, unfettered dissemination of information also raises difficult issues at the international level, since the content posted, for example, by American citizens can bleed across national boundaries and reach the citizens of countries where it is illegal. n4 Because their speech is legal in the United [*276] States, those responsible for posting the information cannot be prosecuted in this country; consequently, they cannot be extradited for prosecution in countries where their speech is illegal. n5 This, in turn, creates the possibility that the United States will become a "speech haven," n6 i.e., a nation that harbors those whose speech is outlawed elsewhere, and gives them the freedom and the resources that they need to broadcast their messages to the [*277] world. n7 If one assumes, as seems reasonable, that the United States would prefer not to become a haven for the Nazis of the new millennium, it becomes necessary to consider the extent to which the emerging varieties of, and opportunities for, disseminating information can be criminalized under American law. n8 [*278] The notion that criminal liability can be predicated upon what one says, as opposed to what one does, is far from new. n9 In the [*279] Anglo-American tradition, for example, the crime of blasphemy dates back to the seventeenth century, n10 and other "word crimes" were in existence long before that. n11 Blasphemy was one of the four "branches" of common law criminal libel; the others were obscenity, sedition, and defamation. "All four ... sought to ensure that speech did not violate established norms of respect and propriety." n12 Since courts of this era agreed that Christianity was part and "parcel of the common law of England," n13 it followed that speech that was repugnant to Christian values and violated the law, which could be punished as a criminal act. n14 This was an instance in which the law imposed the views of a dominant group upon an entire populace instead of protecting the values of many diverse groups, n15 a phenomenon that has been common throughout history. This perspective, which came to America with the colonists, n16 survived well into the nineteenth century, notwithstanding the adoption of the First Amendment. n17 By the end of the nineteenth century, the dominance of this view had begun to erode, due in large part to the increasing diversity of American society. n18 The influx of millions of immigrants in the nineteenth and early twentieth centuries gradually diluted the Anglo-Saxon, Protestant homogeneity of the original colonies. n19 America became a pluralist [*280] society in which no single group's views were "parcel" of the law. One consequence of this diversification of American society was an increasing intolerance for restrictions on speech, especially value-laden speech. This intolerance, coupled with other factors, resulted in the creation of free speech jurisprudence, free speech having received little or no attention from courts during the nineteenth century. n20 As a result of this new emphasis on free speech, many of the old laws criminalizing speech, including blasphemy [*281] and sedition laws, were either struck down as violating the First Amendment or fell into disuse for the same reason. n21 Contemporary American law governing the criminalization of speech is largely the product of evolved technology. Criminal prohibitions are adopted to control behaviors; laws criminalizing speech are therefore designed to control at least certain types of speech. n22 To understand why it has been deemed necessary to control speech, it is necessary to consider how the technologies of communication have evolved and how they interact with speech. In pre-literate societies, which have no communication technology, speech is merely oral, 2 severely limiting its dissemination: Face to face oral communication is restricted in scale; it can involve one speaker and one listener or expand to involve more than one speaker and multiple listeners, but its audience and its [*282] geographical reach are inevitably circumscribed by time and space. A diatribe delivered to a pre-literate audience (a) reaches only those who were physically capable of traveling to the location where it was delivered and (b) survives only imperfectly in the idiosyncratic memories of those who were present to hear it. These characteristics, plus the fact that one could not speak anonymously, mean that it is relatively easy for those in power to control speech by exerting informal pressures to discourage speech that they consider undesirable and/or by using force against those whose speech they deem undesirable. The development of written language makes it somewhat more difficult to control speech because the speaker can use script to transcend the geographical and temporal limitations inherent in face-to-face oral communication. A dissident can transcribe his diatribe and send it to a geographically dispersed audience, who can read the diatribe and pass this preserved speech on to others. Script literacy does not, however, create significant control problems for those who are determined to ensure that speech stays within what they consider acceptable bounds: Each copy of the diatribe must be written, or transcribed, individually; the dissemination of speech remains limited in scale since this is a time-consuming process and there may be few literate dissidents. Additionally, because it is likely that only a small proportion of the populace will be literate, and because there is a good chance that each individual's handwriting is distinctive, it is relatively easy for those concerned with controlling speech to identify the authors of the diatribes and take action against them. All this changed with the availability of communication technologies, the first of which was the printing press. n23 Using a printing press, one can produce hundreds or even thousands of copies of a transcribed diatribe quickly, cheaply, and anonymously, with the copies being distributed over as wide a geographical area as physical transport allows. The copies, which perfectly record the substance of the diatribe, can be preserved or passed on to others; the substance of the diatribe survives to influence later readers, and perhaps shape future diatribes. It did not take those in power long to realize the dangers that printing posed to their ability to control speech: n24 The first printing press was brought to England [*283] in 1476, and by 1538, a Crown "proclamation required a license from the Privy Council or a bishop before any English book could be printed." n25 The licensing system lasted until 1694, n26 when it was abolished, common law courts began to use the common law crime of seditious libel - along with blasphemy and obscenity - to control speech. n27 Part II.A outlines the American history of these common law crimes and the other offenses that have been employed to control speech. These devices survived the adoption of the First Amendment and persisted well into the last century, long after communication technology had expanded well beyond the use of the printing press; n28 indeed, some of these devices are still in use. n29 This is not surprising because all pre-cyberspace technologies had one thing in common: Printing, cinema, radio, and television disseminate identical content that has been generated by or on behalf of a broadcaster to an audience that is composed of many individuals; the audience's role is passive, to receive the speech, as given, without the ability to respond or to modify. By the mid-twentieth century, these modalities had become the dominant [*284] modes of disseminating information and opinion. Economic forces dictate that those who decide what content to disseminate will identify with the social and political status quo, so there was little need to resort to formal measures - such as licensing or criminal prosecution - to ensure that this content stayed within "acceptable" bounds. n30 The attitudes and allegiances of those who controlled these various media outlets acted as an empirical "filter," ensuring that undesirable content was not broadcast to the American public. Criminal prosecution was needed only to deal with the outliers, i.e., with those who persisted in publishing "questionable" content on a limited scale, n31 and with their patrons. n32 3 The dominance of these modalities of mass communication is being undermined by the rise of the Internet and the use of cyberspace to communicate. n33 Cyberspace has the capacity to become [*285] the ultimate marketplace of ideas: Anyone with access to a computer connected to the Internet can "publish" ideas and information. The pragmatic, empirical "filters" that are an intrinsic aspect of existing mass media disappear, and one can share whatever speech they like with the world - recipes, diatribes, innermost thoughts, sexual exploits and fantasies, conspiracy theories, information, misinformation, etc. This creates new dilemmas for law, especially for American law with its First Amendment guarantee of free speech. The rise of cyberspace communication will test the American commitment to the notion of free speech because a complacential bromide has become a reality; for the first time in history speech is truly "free." This is a matter of concern for some who argue that it is necessary to adopt new, formal methods of controlling speech that is disseminated via the Internet, including the use of criminal sanctions, because the pragmatic, empirical "filters" that govern speech disseminated by radio, cinema, television, and conventional print media do not constrain cyberspace communication. n34 The viability of this argument is considered below: Part II reviews the criminalization of speech in Anglo-American history and Part III analyzes the extent to which criminal liability can, and should, be used to control cyberspace speech. III. Complicit Publication: Imputing Criminality to Communication "We educated no murderers. The contents of the articles which I wrote could not have educated murderers." n325 Cyberspace democratizes communication by eliminating the tacit, pragmatic filters that limit what is disseminated by the communication modalities that dominated until the late twentieth century, i.e., printing, cinema, radio and television. n326 Cyberspace makes speech truly "free" for the first time, and in so doing, creates a challenge for law, especially for American law with its professed commitment to free speech: Can social order co-exist with the perfectly unrestrained dissemination of ideas and information? To maintain social order, it is necessary to proscribe certain types of behavior. The proscribed behaviors are those that cannot be tolerated if organized social life is to exist because they produce socially undesirable "harms;" for the most part, they involve patterns of victimization, individuals preying on each other. The behaviors are proscribed by defining them as "criminal" and sanctioning those who flout the proscriptions by engaging in the outlawed conduct. n327 And although criminal proscriptions do vary across social systems, there is a notable level of consistency in a core set of prohibitions that are designed to protect the interests vital to every society: The safety of persons; the security of property; the stability of the government; and the sanctity of certain moral principles. n328 No society can survive if its members are free to harm each other at will, appropriate each other's property, [*336] undermine the political order, and/or flout the moral principles that the citizenry hold dear. Consequently, every society will formulate penal prohibitions defining (a) crimes against persons (e.g., murder, assault, rape); (b) crimes against property (e.g., theft, arson, fraud); (c) crimes against the state (e.g., treason, rioting, obstruction of justice); and (d) crimes against morality (e.g., gambling, defiling a place of worship). n329 There will be varying degrees of consistency in these prohibitions. By far the greatest degree of consistency will occur in the first two categories because they represent the malum in se crimes; these are the absolute prohibitions a society must establish to maintain a modicum of social order since they outlaw the [*337] direct infliction of "harm" by one person upon another or others. n330 4 There will also be consistency as to a core of offenses in the third category because every society must protect its basic governing processes; n331 but there will be more overall deviation in this category because societies vary in the extent to which they are willing to tolerate political dissidence. n332 The least amount of consistency will exist as to offenses in the fourth category; these prohibitions are the product of a society's values and religious principles, and therefore tend to be more idiosyncratic in nature. n333 Criminal proscriptions targeting the content of speech have historically played a minor role in Anglo-American law. Content-based proscriptions did not really appear until the first great "information revolution" - the proliferation of the printing press - made it possible to disseminate ideas and information to a wide audience. n334 Perceiving this as a threat to the existing social order, British law reacted by creating two new crimes against the [*338] state (seditious and criminal libel) and two new crimes against morality (obscene and blasphemous libel). n335 As Part II explained, these crimes became part of American law, but only one obscenity - survives, the others having foundered on the First Amendment or fallen into disuse. The second great "information revolution" has just begun but already ideas and information can be disseminated faster, more inexpensively and on a wider scale than has ever been possible before. In addition, at this point, it is simply not possible to predict what cyberspace communication will become in, say, the next century; unlike printing, which is a static technology, computer-facilitated communication is a dynamic technology that can evolve into new and ever-more complex forms. Any attempt to develop law and policy in this area is, therefore, necessarily an exercise in speculation. n336 It is, however, reasonable to postulate that the pragmatic constraints limiting the amount and type of speech that is disseminated will be significantly reduced in importance, if they do not disappear entirely. That brings us back to the question posed at the beginning of this section: Can social order coexist with the perfectly unrestrained dissemination of ideas and information? Or to put it another way, as the dissemination of ideas and information is constrained less and less by pragmatic filters, will societies find it necessary to proscribe the dissemination of at least certain types of speech on the premise that it creates a risk of "harm" to persons, property, the state or morality? The best analytical test case for considering how these questions can, and should, be answered is American society because of its constitutionally embedded commitment to "free" speech. The sections below undertake this analysis, examining two basic issues: (1) Should criminal prohibitions be used to limit the dissemination of ideas and information via cyberspace?; and (2) Can criminal prohibitions be put to this use given the American guarantee of free speech? These issues are addressed in reverse order, [*339] since the nature of the available proscriptions is an important factor in deciding whether speech should be outlawed. A. Can Cyberspace Speech be Criminalized? "The evidence has not furnished the slightest proof ... that [Fritzsche] ... committed [a] ... crime ... through his transmission of news; there is not the slightest evidence to show that he has instigated a single person to murder, cruelties, deportations, killing of hostages, massacre of Jews, or other crimes ... ." n337 "The German Command ... had the idea of penalizing the persons who listened to these broadcasts ... asking the French authorities... to institute the most stringent penalties, even going so far as to prescribe the death penalty for persons repeating news heard on the foreign radio service." n338 As Part II demonstrated, two different approaches have been used in criminalizing speech. One approach criminalizes speech based on its content. For example, using the phrase "Pig 5 Madonna," or tying mentions "of God and the Madonna ... to explicit images of sex," constitutes criminal blasphemy. n339 The rationale for content-based criminalization is that the speech in question is either "dangerous" (seditious speech) or "unworthy" (blasphemy and obscenity). n340 Since the content of the speech is [*340] somehow antithetical to the maintenance of social order, it must be outlawed. n341 This proposition dictated the creation of the four common law-derived content offenses discussed in Part II.A. One was a crime against the state (seditious libel), two were crimes against morality (obscenity and blasphemy), and the fourth (criminal libel) was a crime against the state (intended to prevent breaches of the peace caused by defamatory remarks). n342 The second approach - the "speech act" approach - criminalizes speech when it serves as the actus reus of an independent criminal offense. Here, speech qua speech is not criminalized; instead, criminal liability is imposed for speech because it is the conduct by which one violates a freestanding criminal prohibition. In State v. Tally, n343 for example, Tally was convicted of being an accomplice to murder based on his telling the killers where to find their victim. n344 As Part II.B explained, the speech-act approach allows criminal liability to be imposed for offenses that effectively target speech (e.g., perjury), as well as for those that do not (e.g., contempt). The division of labor represented by these approaches has been a satisfactory way of addressing "undesirable" speech, at least until recently. Computer technology and computerfacilitated communication, however, raise issues that cannot easily be resolved by this framework, as illustrated by United States v. Elcom, Ltd. n345 Dmitry Sklyarov, an employee of a Russian software company called Elcom Ltd., n346 was indicted on and charged with violating two provisions of the Digital Millennium Copyright Act ("DMCA"); n347 specifically, 1201(b)(1)(A), which makes it unlawful [*341] to traffic technology designed to circumvent the rights of a copyright owner, and 1201(b)(1)(C), which makes it unlawful to traffic technology marketed for use in circumventing technology that protects the rights of a copyright owner. n348 The first count of the indictment alleged that Elcom Ltd. and Sklyarov conspired to traffic technology that was designed to circumvent technology that protects the rights of a copyright owner. n349 Furthermore, it alleged that as an Elcom Ltd. employee, Sklyarov wrote a program called the Advanced ebook Processor (AEBPR) that allows users to "remove any and all limitations on an ebook purchaser's ability to copy, distribute, print, have the text read audibly by the computer, or any other limitation imposed by the publisher or distributor of the ebook." n350 The AEBPR program on its face seems to be legitimate software. However, the program also produced negative effects. The conversion accomplished by the AEBPR program enabled a purchaser of an ebook to engage in 'fair use' of an ebook without infringing the copyright laws, for example, by allowing the lawful owner of an ebook to read it on another computer, to make a back-up copy, or to print the ebook in paper form. The same technology, however, also allowed a user to engage in copyright infringement by making and distributing unlawful copies of the ebook. n351 The indictment accused Elcom of selling the AEBPR program on its website to people who could use it to circumvent copy protection on eBooks. n352 After Sklyarov's prosecution was deferred because he agreed to cooperate with the government, his employer moved to dismiss the indictment on due process and First Amendment [*342] grounds. n353 The company argued that the DMCA violated the First Amendment "because it constitutes a content-based 6 restriction on speech that is not sufficiently tailored to serve a compelling government interest." n354 The government, in turn, argued that the DMCA does not implicate First Amendment protections because (a) the DMCA targets the act of selling technology and the sale of circumvention technology is not "speech," and (b) Sklyarov's computer code, AEBPR, is not speech protected by the First Amendment. n355 The district court rejected the government's arguments and held that the prosecution did implicate speech protected by the First Amendment. n356 Since the First Amendment was implicated, the court turned to the question of whether the DMCA is a content-based or content-neutral regulation. n357 Elcom Ltd. claimed the DMCA is a content-based regulation because it bans instructive computer code, which is speech entitled to First Amendment protection. n358 In response, the government argued that DMCA is a content-neutral regulation subject to less rigorous scrutiny than a content-based regulation for purposes of the First Amendment. n359 After noting that "the principal inquiry in determining whether a statute is content-neutral is whether the government has adopted a regulation of speech because of agreement or disagreement with the message it conveys," n360 [*343] the court agreed with the government. n361 At the outset, it found that nothing in the legislative history of the DMCA evinced a Congressional intent to target speech because of its expressive content. Rather, Congress sought ways to further electronic commerce and protect intellectual property rights, while at the same time protecting fair use. In order to balance these priorities, Congress sought to ban trafficking in any technology or device that could be used to circumvent technological restrictions that served to protect the rights of copyright owners. n362 Elcom argued that these technologies "necessarily involve computer code" and that "'it is precisely the content of the code that causes the government to regulate it.'" n363 The district court explained that this argument went too far. In the digital age, more and more conduct occurs through the use of computers and over the Internet. Accordingly, more and more conduct occurs through "speech" by way of messages typed onto a keyboard or implemented through the use of computer code ... . The mere fact that this conduct occurs at some level through expression does not elevate all such conduct to the highest levels of First Amendment protection. Doing so would turn centuries of our law and legal tradition on its head, eviscerating the carefully crafted balance between protecting free speech and permissible governmental regulation. n364 The court concluded that the DMCA does not regulate speech; it merely bans trafficking in devices. n365 Therefore, it held that the DMCA is a content-neutral regulation because it bans code "not because of what the code says, but rather because of what the code [*344] does." n366 The court upheld the DMCA under the less demanding scrutiny given to content-neutral regulations. n367 With all due respect to the Elcom court, it was wrong. The provisions of the DMCA at issue in Elcom are content-based criminal regulations of speech. To say that the DMCA bans code "not because of what it says but because of what it does" n368 evades the real issue, which is whether, in a technologically driven world, it becomes necessary to criminalize certain types of speech because of its content. The court avoided this issue because it did not want to deal 7 with the consequences of finding that the DMCA is a content-based regulation. As a contentbased regulation, the DMCA might not have survived the rigorous scrutiny to which such regulations are subjected under current law. n369 [*345] To understand why the Elcom court erred when it held that the DMCA provisions are not content-based regulations of speech, it is necessary to consider a variation on the facts in the case. The DMCA makes it an offense to "offer to the public, provide, or otherwise traffic in any technology, product, service, device, component, or part thereof that (A) is primarily designed ... for the purpose of circumventing ... rights of a copyright owner ... [or] (B) has only limited commercially significant purpose or use other than [circumventing copyright] protection." n370 Assume that instead of marketing a program, Elcom had posted on its website the code that could be used to crack embedded copyright restrictions on eBooks. n371 Also, assume that the code qualifies as technology or a component or part thereof n372 that is primarily designed to circumvent copyright protection or has only limited commercially significant uses other than circumventing copyright protection. By posting the code, Elcom has "offered it to the public" or "provided" it or otherwise "trafficked" in it within the scope of the DMCA. n373 Therefore, Elcom's conduct violates the DMCA, and it is prosecuted for posting the code. Here, the prosecution clearly targets the content of speech. The code functions as a set of instructions on how to bypass technology that maintains copyright restrictions on eBooks. The instructions convey information from one human being to another, and are [*346] therefore speech protected by the First Amendment. n374 By posting the code on the website, Elcom, or, perhaps more precisely, Sklyarov, n375 communicates this information to anyone who cares to read it. Elcom is essentially publishing the results of scientific inquiry, an act that is protected by the First Amendment. n376 Of course, the act of publishing this information is also a violation of the DMCA, which brings us back to the issue the Elcom court avoided: How should the law deal with conduct that is presumptively protected by the First Amendment, but that is potentially capable of assisting in the commission of criminal activity? Is it permissible to criminalize speech that is merely capable of facilitating unlawful acts? The problem, which did not exist prior to the rise of cyberspace communication, is that the potential assistance is ambiguous, subtle, and diffuse in nature. Instead of selling an erstwhile bank robber the combination he can use to open a bank safe, one posts code on the Internet, knowing it can be used to override the encryption that protects copyrighted material. Selling the combination to the safe is a one-to-one transaction that is analogous to the formation of a criminal contract. The seller knows the robber needs this combination to open either a specific safe (e.g., the First National Bank safe) or to open a safe of a particular type (e.g., the Guardian Series 3000 Safe). Because the seller knows that the combination will be used to engage in specific criminal activity, selling it to the robber is a "speech act" n377 because it aids and abets a crime that will be committed in the future. n378 If the crime is not committed, the seller can still be prosecuted for attempting to aid and abet its commission. n379 [*347] Posting code is a much more ambiguous act than selling the combination to a bank safe. It is not a one-to-one transaction in which the seller knows, or can reasonably be deemed to know, that particular, recognized individuals would use the information to commit a specific crime. Posting information is a one-to-many transaction akin to publishing an article or a book. The person posting the information distributes it to an undefined universe of anonymous persons with no way of knowing precisely who will access it and the uses to which they may put it. n380 Another difficulty is that the information being disseminated may have both legitimate and illegitimate uses. This is theoretically true of the combination to the bank safe when used legitimately by bank staff, but the circumstances surrounding its transmission to one with 8 acknowledged criminal purposes eliminates any argument that it is disseminated for legitimate purposes. Comparable circumstances are usually lacking when information is posted on a website. In the hypothetical given above, for example, Elcom may have posted the code that can be used to bypass copyright restrictions on eBooks: (a) it wants to facilitate the criminal circumvention of copyright restrictions on eBooks because; (b) it wants to bring this information to the attention of computer scientists and thereby increase the body of knowledge in this area; or (c) this is a marketing device that is intended to demonstrate the technical expertise of the company's employees. The posted code is speech protected by the First Amendment. n381 In alternatives (b) and (c), the speech is disseminated for purely lawful purposes, and cannot provide a basis for imposing criminal liability. n382 In alternative (a), Elcom "spoke" - [*348] disseminated expressive content - with the knowledge that the information would be used to engage in unlawful activity and, presumably, with the purpose that it would be put to this end. Unfortunately, unless Elcom admits that alternative (a) was the sole reason for posting the code, n383 there is no way to prove its actions constitute a "speech act" - aiding and abetting violations of eBook copyright protections. n384 There is no demonstrable link, no meeting of the criminal minds, between Elcom and those who commit such violations. The DMCA's approach to these issues is to treat certain types of speech, and more specifically, certain types of information, as tools that must be outlawed because they can be used to commit crimes. In effect, the DMCA analogizes information of this type to "burglar's tools," the possession and manufacture of which is illegal in many states. n385 This is why the DMCA is in fact a content-based regulation of speech. Like the libel offenses that English law created as a reaction to the proliferation of the printing press, n386 the DMCA criminalizes a category of speech based purely on its content. In both instances, criminal liability is imposed for the act of uttering the speech. There is no inquiry into whether the person responsible for the speech acted with the intent to facilitate the commission of a criminal act. The English libel offenses outlawed certain categories of speech because they presumptively threatened the stability of the state (seditious libel) or offended [*349] prevailing notions of morality (blasphemy and obscenity). n387 In a similar fashion, the DMCA outlaws a category of speech to protect private property rights, i.e., intellectual property rights. n388 The question is whether this is a constitutionally permissible strategy. It is interesting to contrast the DMCA's approach with the treatment that is given to another analogous category of speech, which includes books and websites that provide instructions for creating bombs, poisons, and other devices of destruction. n389 Bomb making instructions in print have been available for decades, yet Congress has not made any effort to outlaw them. n390 In 1995, Senator Dianne Feinstein proposed legislation that would have made it a federal crime to use the Internet to "demonstrate the making ... of an explosive ... or to distribute ... information pertaining to ... the manufacture ... of such an explosive," but the proposal died in committee and has never been revived. n391 The Department of Justice was asked to evaluate the [*350] constitutionality of the Feinstein proposal, since it was clearly a contentbased regulation of speech. n392 In a report issued in 1997, the Department of Justice concluded that the First Amendment would impose substantial constraints on any attempt to indiscriminately proscribe the dissemination of bombmaking information. n393 "The government generally may not ... punish persons either for advocating lawless action or for disseminating truthful information - including information that would be dangerous if used that such persons have obtained lawfully." n394 The report also pointed out that the Feinstein proposal could be construed to impose culpability if the person disseminating the information has reason to know that some unidentified, unspecified recipient thereof will use the information for an unlawful purpose, or if such an outcome is the "natural consequence" of publication of 9 the information. Because that construction could cover virtually all public dissemination of such information, it would raise serious constitutional questions. n395 Comparing the failed history of criminalizing the dissemination of bomb making information with the DMCA suggests a paradox. The First Amendment does not permit Congress to outlaw the general dissemination of bomb making information when the natural consequence of that dissemination is that the information will be used for an unlawful purpose, i.e., to inflict injury and destruction upon persons and property. However, the First [*351] Amendment does permit Congress to outlaw the general dissemination of information that instructs how to circumvent copyright protection technology when the natural consequence of that dissemination is that the information will be used to violate laws protecting intellectual property rights. Presumably, this is a matter of less moral import than the infliction of injury, death, and September 11th types of property destruction. The Department of Justice would no doubt argue that there is no paradox, and that the difference between the two scenarios lies not in the consequences that may result from disseminating the information, but from the nature of the information itself. As noted earlier, the Department of Justice argued in Elcom that computer code is not speech and is therefore not protected by the First Amendment, an argument courts have rejected. n396 It has been argued that while code within the scope of the DMCA may be "speech," it is not speech that is entitled to full First Amendment protection because it is not solely concerned with communication between human beings. n397 According to this logic, the code is not entitled to full First Amendment protection because at least some portion of it is concerned with facilitating communication between computers, and the First Amendment does not protect "machine speech." n398 This reasoning is problematic for at least two reasons. It introduces an element of uncertainty to the application of the DMCA's criminal prohibitions. If Elcom posts a program on its website, the First Amendment does not apply because a program is machine speech. However, if the company posts code that humans must use to achieve a result, this is speech because the code communicates information from one human being to another. [*352] The arbitrariness inherent in this solution is unacceptable because, aside from anything else, it hinders predictability. One of the basic principles of criminal law is that individuals must be given fair notice of what is forbidden so they can conform their conduct to the demands of the law. n399 The level of ambiguity inherent in the DMCA's criminalization of information-as-inchoate-tool means that individuals can be left uncertain as to whether they risk violating the law if they choose to disseminate information about subjects that fall within the statute's compass. n400 In some ways, this is a de facto return to the common law rule that eliminated prior restraints on speech but allowed speech to be prosecuted if it was determined to be unlawful after the fact. n401 Another problem with the reasoning outlined above is that it ignores the effect computer technology and cyberspace will have on speech. It is as if the Department of Justice announced in 1910 that the First Amendment applied to written and face-to-face speech, but not to speech communicated via the telephone because this was "machine speech," a form of artificial speech. To use what may be a better analogy, it is as if the Attorney General announced in 1850 that the First Amendment did not apply to the contents of telegraphic communication because telegraphic communication was purely "machine speech," i.e., telegraphic instruments [*353] communicating to each other via a set of code. n402 Technology allows speech to assume new and different forms, and speech has never been defined as being outside the First Amendment simply because it takes an unfamiliar form. n403 10 The content-based criminalization used in the DMCA is a simplistic solution to a complex problem: Preventing the dissemination of ideas or information from being used to cause certain types of harm. n404 The DMCA focuses on copyright infringement, a relatively traditional type of harm, which is essentially theft, the ultimate infliction of which occurs in a conventional manner. The only novel aspect is that the infliction of this harm is facilitated by cyberspace communication, that is, by the online transmission of burglar's tools. n405 This, however, does not exhaust the criminal potential of online speech. The increasingly free speech created by cyberspace communication has already given rise to scenarios in which the dissemination of ideas and information threatens to: (a) inflict new types of harm; or (b) create new ways to inflict traditional harms. Since speech is at the heart of these scenarios, using content-based criminalization to attack them would effectively eviscerate the First Amendment because it would require [*354] outlawing numerous types of speech. These scenarios - and others that will likely emerge in the future - should be criminally proscribed and sanctions should be applied to those who perpetrate them because they do produce socially unacceptable harms. The Department of Justice recognized the only logical alternative in its report on the dissemination of bombmaking information: The "speech act" approach does not criminalize speech for content, but rather the conduct by which one commits an independent noncontent-based offense. n406 The "speech act" approach avoids First Amendment issues because it focuses on the traditional benchmarks of criminal activity - an act taken with the intent to produce a result - instead of on the content of speech. n407 Unfortunately, applying the "speech act" approach to conduct occurring in and via cyberspace is not as straightforward a process as it might seem. Like all criminal law doctrines, the "speech act" approach evolved to deal with conduct occurring in the "real world." n408 Real world conduct is relatively unambiguous because it involves synchronous one-to-one interactions. n409 For example, if an alienated bank teller sells the combination to the safe in her bank to someone whom she knows intends to rob the bank, it is reasonable to infer that she acted with the purpose of bringing about a socially proscribed harm - the theft of property belonging to the bank. Moreover, it is consequentially reasonable, assuming [*355] the above facts are proven beyond a reasonable doubt, to hold her liable as an accomplice to the consummated bank robbery. Cyberspace complicates the analysis by adding the potential for asynchronous one-to-many and many-to-many interactions to the synchronous one-to-one interactions that produced the current body of criminal law. The issues that can arise from these new modes of interaction are illustrated by a variation on the hypothetical used earlier. n410 Assume that instead of selling the bank safe combination to someone whom she knows intends to rob the bank, the alienated teller posts it on a website. If an aspiring bank robber who knows nothing about the alienated teller or her motivations for posting the combination finds it on the website and uses it to rob the bank, can the teller reasonably be held liable as an accomplice under the "speech act" approach? What if instead of posting the combination on the website, she posts information about the bank's operations, i.e., when employees arrive in the morning and leave at night, the number of security guards stationed in the bank at any given time, and the security systems in place at the bank? If the bank is robbed, can she reasonably be held to have been an accomplice to the robbers? Can the traditional "speech act" approach be used to answer these questions? If not, how should the law approach apportioning criminal liability in these and comparable situations? The sections below address these issues. They consider how the criminal law should go about imposing liability for the dissemination of speech. The analysis is based on the premise that criminal liability can be imposed only when speech is the actus reus of an independent, noncontent-based criminal prohibition. It also divides speech into three analytical categories: Notional ideas, applied ideas, and data. n411 11 In this schema, "ideas" refer to abstractions generated by the human intellect. Ideas are constructs of the human mind. "Data" is content that has been compiled or structured by the human mind, but that has an existence independent of the human [*356] mind. n412 The differences between the three analytical categories are illustrated by the following example: n413 Notional idea: There are physiological and psychological differences between human races. The members of some races are stronger or smarter than others. Data: Statistical evidence showing that more whites than blacks have college and advanced degrees and that the average earnings of whites far exceed those of blacks. Applied ideas: Whites are "superior" to blacks (opinion) and therefore whites should receive certain advantages over blacks. Blacks should be treated as second-class citizens (advocacy). n414 This division is in no way intended to imply that one category of speech is more important than any others. It is a device used to explore the possibility that certain types of speech may have an inherently greater capacity for facilitating criminal conduct than other types. If this is true, it is a matter that the criminal law needs to take into account; that is, if certain types of speech have an inherently greater capacity to facilitate criminal conduct than others, this is an empirical circumstance which needs to be considered in the articulation and application of criminal doctrines governing conduct mediated through cyberspace. If, of course, this is not true, then the sections below simply become a useful organizational structure. B. Should Cyberspace Speech be Criminalized? The criminalization - or, perhaps more properly, the attempted criminalization - of speech in cyberspace has two aspects, a national aspect and a transnational aspect. At the national level, countries are using criminal sanctions - along with firewalls and other measures - in an effort to control the speech that reaches their citizens via cyberspace. n703 That is the topic this article addresses; it is concerned with whether the United States of America can and should criminalize the use of cyberspace to disseminate certain types of speech. The sections above analyzed the extent to which criminal sanctions can be used to this end. This section considers whether they should be used to this end. Earlier, it was explained that criminal proscriptions are used to control behaviors that cannot be tolerated if a society is to survive. n704 This proposition is obvious as to certain types of conduct. No society, for example, can allow its citizens to inflict physical injury on each other to rape, assault and murder each other - at their individual whim. The same is true of property; some level of security in one's property is an essential part of any society. A society must, therefore, outlaw intolerable behaviors and sanction those who obstinately engage in them. These rules provide a baseline of stability and predictability that is essential for normal social functions. Historically, they have for the most part been directed toward conduct, not speech. Generally, disruptive physical acts are a more destabilizing force than mere speech. However, speech, too, has been outlawed. As Part II.A explained, [*423] Anglo-American law has used criminal proscriptions to control speech that authorities have regarded as "dangerous." As that section also explained, these proscriptions were a reaction to a technological innovation - the printing press - that made speech more "free" than it had ever been before. But as centuries passed and pragmatic constraints evolved to regulate the amount and types of speech that could be disseminated via the dominant modalities of mass communication - i.e., printing, radio, television and the cinema - the perceived need for prohibitions of this type eroded and disappeared, at least in American law. n705 12 A new technology has emerged, one that has the capacity to make speech truly "free" by eliminating these constraints. n706 That capacity is causing consternation in some segments of American society, as well as in other societies. For example, in some countries, it has produced efforts to use criminal sanctions to control the potentially unbridled speech made possible by cyberspace. n707 For American law, there are two possible responses to this phenomenon: One is to reinstitution content-based prohibitions in an attempt to discourage speech that is seen as socially intolerable; in my opinion, this is what the DMCA does. n708 The other is the "speech act" approach, discussed in Part III.A, whereby speech is punished when it becomes the actus reus by which someone violates an independent, non-contentbased criminal prohibition. This approach can be used to impose direct liability for a substantive offense and imputed liability for facilitating the commission of such an offense. n709 This is the approach American law has followed for at least sixty years, n710 and it is the approach we must continue to follow. Embarking upon the proscription of speech based on its content is inconsistent with the principles and values contemporary American society claims to honor, and certainly claimed to honor while pragmatic filters meant that speech was "free" in principle but constrained in fact. The rise of cyberspace has undermined the effectiveness of these filters, a trend that will only accelerate. The [*424] consequence of this trend is that speech that seems valueless, offensive, and perhaps even "dangerous," is entering the public discourse in a way that would never have been possible thirty years ago. n711 While some may find this disturbing, we are not obliged to "listen" to this speech. We can simply ignore it or use other means of technology to shut it out. However distasteful or offensive some of this speech may be, it is merely speech. Unless it becomes the actus reus of a crime, speech is "harm-less" and should not become the basis for imposing criminal liability. n712 American society is beginning to confront the world's first real marketplace of ideas. It is learning, along with the rest of the world, that as in any marketplace, some of the goods are shoddy, distasteful, even disgusting. n713 The best way, indeed, the only way to preserve freedom of thought is to allow odious speech to be heard and rejected. n714 There is another aspect to the criminalization of speech in cyberspace debate: The transnational aspect. Cyberspace transcends national boundaries, but law is still defined by physical territory. This creates situations in which one nation, outraged by speech a citizen of another nation has posted on a website hosted in that country, seeks to prosecute the noncitizen for speech that violates its law but is quite legal in the other country. n715 This scenario raises a number of legal issues, such as jurisdiction and the need to harmonize laws in an era of cyberspace communication, [*425] which are not pertinent to the issue currently under consideration. It also raises a practical issue that is relevant to this discussion. It was much easier to control speech when the printing press produced the first content-based criminalization of speech laws in the Anglo-American tradition. n716 Printing presses can be found and confiscated or destroyed, as can the products they create. It is relatively easy to accomplish this because printing presses and printed matter exist in the real, physical world; they necessarily exist in a fixed, spatial location that is part of a territorial domain in which certain laws hold sway. When seditious libel became a crime in England, n717 it was a relatively straightforward process to find the printing presses that were being used to publish sedition, confiscate the presses, along with any available seditious literature, and prosecute the printer. The process was unproblematic because law and territory coincided; the press, the printer and the literature all were physically present in a jurisdiction in which the publishing of this particular speech was a crime. That is no longer true. Websites are hosted on servers that are physically located in a jurisdiction, but the content they disseminate is available around the world. If the speech a website displays is illegal under the law of the jurisdiction in which the server hosting it is located, and if the operator of the website is also physically present in that jurisdiction, there 13 is again a coincidence of law and territory. The operator can be apprehended and prosecuted for violating local law. The operator of the website can, however, use cyberspace to prevent law and territory from coinciding and thereby make prosecution difficult, if not impossible. Assume, for the sake of analysis, that the state of Washington enacts a statute, which makes it a crime to post personal information about law enforcement officers on a website, and the Justice Files, a website that is hosted on a server in the state of Washington, displays personal information about law enforcement officers. n718 If the operator of the website remains physically in the state of Washington and keeps the site on the original server, there is a coincidence of law and territory and he can be apprehended and prosecuted. But what if the operator of the Justice Files, knowing the statute will soon go into effect, moves it to, say, [*426] a server in Tahiti? n719 Now the website is no longer physically located in the state of Washington, even though the information it displays is still accessible in Washington and even though this information is displayed in violation of Washington law. If the operator of the website remains in Washington, the state of Washington might try to prosecute him for violating the statute, on the premise that he is physically within its territory and therefore subject to its laws. Such a prosecution would be problematic if the posting of the information is legal under Tahitian law; n720 and it would be even more problematic if the ostensible operator of the site disclaimed any responsibility for the new site, or if he left the state of Washington and took full responsibility for it. Law and territory no longer coincide. Lawyers and legislators around the world are grappling with how this scenario should be resolved. It presents issues of law and policy that, again, are not relevant to this discussion. However, the scenario is pertinent to this discussion as it illustrates that if the content of speech is criminalized by one territorial jurisdiction the speech can migrate to another, more hospitable jurisdiction. Consequently, it becomes much more difficult to use criminal liability to control speech than it was when England formulated criminal libel laws to control what appeared in print. This is an empirical reality. Should this empirical reality impact upon efforts to criminalize the dissemination of "undesirable" speech via cyberspace? The difficulty of enforcing criminal prohibitions is definitely not a reason to avoid instituting such prohibitions, but there is more to the scenario described above than the difficulty cyberspace creates for the enforcement of laws criminalizing speech. It illustrates how cyberspace creates a gap between "speech" and "conduct," at least with regard to the application of the criminal law. For the most part, criminal law is parochial, concerned with occurrences in the local community. That is because it is concerned with victimization, i.e., with a perpetrator's inflicting "harm" upon a victim. Historically, victimization involved a face-to-face transaction between the two; rape, murder, assault, most types of theft, arson and the [*427] like were all synchronous events, occurring in "real time" and in "real space." n721 As a result, criminal law, which is arguably the oldest form of law, evolved as bounded law; this, in turn, evolved into the concept of jurisdiction, which equates physical presence in a territory with the power to apply local law. This model of criminal law is likely to survive insofar as it is concerned with victimization-byconduct; that is, while some crimes - such as fraud and theft - are migrating into cyberspace, most are likely to remain a "real world" affair. n722 The primary reason for this is that traditional "crimes" tend to encompass visceral behavior, physical behavior: n723 People hurt each other; they damage each other's property; and they take each other's property. People are going to remain in the "real world," as is most of their property; indeed, the property that remains in the "real world" will continue to be the most valued property, since it is the only property one can enjoy in the physical reality where they will spend most of their conscious existence. This means that the criminal law's bounded, territorial focus is unlikely to prove problematic with regard to the "conduct" crimes; these are crimes that are predicated on the infliction of some "harm" in the physical world. Homicide, sexual assault, assault, arson, theft, fraud and the like will for the most part continue to be local crimes; when they do transcend national boundaries, scenarios such as the one hypothesized above are unlikely to arise. If 14 police determine that A, who is located in Tahiti, used a website hosted in Tahiti to murder B, who lived in Washington, it is almost certain that A will be extradited for prosecution in Washington because murder is a crime in every country. In other words, there tends to be a great deal of consensus across nations with regard to "conduct" crimes; so if A, who is now in Sri Lanka, uses a website to defraud B, who [*428] is now in Australia, the Australian authorities can probably have A extradited to stand trial in their country. There are, of course, differences in national criminal codes, but extradition should be no more difficult to obtain for those who commit cyber-mediated versions of "conduct" crimes than it is for those who commit their "real world" variants. A greater degree of consensus exists as to "conduct" crimes because they have undeniable, often immediate, "real world" effects. They are emblematic of the principle that societies must outlaw certain types of behavior or descend into anarchy and chaos. Speech, however, is very different. Its effects are not "real world" effects, at least not unless and until they are translated into conduct by those who "listen" to it. This is why societies vary in terms of the extent to which, and the ways in which, they criminalize the content of speech. Some, those that tend not to criminalize the content of speech, essentially view it as not-conduct, i.e., as something that does not and cannot produce immediate "real world" effects. The criminal laws of these societies therefore ignore speech except, perhaps, when it becomes the actus reus of a conventional, "conduct" crime. n724 Other societies regard certain types of speech as a form of inchoate crime, i.e., as effort that can produce the commission of a "conduct" crime. These societies criminalize the content of speech that falls into this category. Still other societies go even further in regarding certain types of speech as crime-in-itself; blasphemy prohibitions are an example of this approach, since they criminalize speech not for the "conduct" it might inspire but because its very content is anathema. n725 It is this variation in how societies approach the criminalization of speech that creates the scenario described earlier, in which the owner of the Justice Files website moves it to a jurisdiction where its content, which is illegal in Washington, is lawful. If, as seems likely, cyberspace continues to evolve and becomes an ever-more pervasive force in human communication, it will be exceedingly difficult for societies to enforce criminal prohibitions on speech, especially the more Draconian prohibitions. With cyberspace, [*429] speech, unlike conduct, can become free in fact as well as in aspiration. Empirically free speech challenges, indeed eludes, the bounded conception of criminal law. It is, perhaps, too much to hope that the nations of the world will realize this and will at the very least take a parsimonious approach to criminalizing the content of speech. IV. Conclusion In fine, this article argues against the criminalization of speech qua speech. It argues for confining the use of criminal sanctions against speech to those instances in which speech, in effect, becomes the "conduct" involved in committing a conventional crime. The article argues against criminalizing the content of speech, emphasizing the differences that exist between speech and the conduct targeted by traditional criminal prohibitions. It also points out how these differences are exacerbated by the use of cyberspace as the vehicle for disseminating speech. 15