Reno v. American Civil Liberties Union

*849 Justice Stevens

delivered the opinion of the Court.

At issue is the constitutionality of two statutory provisions enacted to protect minors from “indecent” and “patently offensive” communications on the Internet. Notwithstanding the legitimacy and importance of the congressional goal of protecting children from harmful materials, we agree with the three-judge District Court that the statute abridges “the freedom of speech” protected by the First Amendment.1

b — <

The District Court made extensive findings of fact, most of which were based on a detailed stipulation prepared by the parties. See 929 F. Supp. 824, 830-849 (ED Pa. 1996).2 The findings describe the character and the dimensions of the Internet, the availability of sexually explicit material in that medium, and the problems confronting age verification for recipients of Internet communications. Because those findings provide the underpinnings for the legal issues, we begin with a summary of the undisputed facts.

The Internet

The Internet is an international network of interconnected computers. It is the outgrowth of what began in 1969 as a *850 military program called “ARPANET,”3 which was designed to enable computers operated by the military, defense contractors, and universities conducting defense-related research to communicate with one another by redundant channels even if some portions of the network were damaged in a war. While the ARPANET no longer exists, it provided an example for the development of a number of civilian networks that, eventually linking with each other, now enable tens of millions of people to communicate with one another and to access vast amounts of information from around the world. The Internet is “a unique and wholly new medium of worldwide human communication.”4

The Internet has experienced “extraordinary growth.”5 The number of “host” computers — those that store information and relay communications — increased from about 300 in 1981 to approximately 9,400,000 by the time of the trial in 1996. Roughly 60% of these hosts are located in the United States. About 40 million people used the Internet at the time of trial, a number that is expected to mushroom to 200 million by 1999.

Individuals can obtain access to the Internet from many different sources, generally hosts themselves or entities with a host affiliation. Most colleges and universities provide access for their students and faculty; many corporations provide their employees with access through an office network; many communities and local libraries provide free access; and an increasing number of storefront “computer coffee shops” provide access for a small hourly fee. Several major national “online services” such as America Online, CompuServe, the Microsoft Network, and Prodigy offer access to their own extensive proprietary networks as well as a link to the much larger resources of the Internet. These com *851 mercial online services had almost 12 million individual subscribers at the time of trial.

Anyone with access to the Internet may take advantage of a wide variety of communication and information retrieval methods. These methods are constantly evolving and difficult to categorize precisely. But, as presently constituted, those most relevant to this case are electronic mail (e-mail), automatic mailing list services (“mail exploders,” sometimes referred to as “listservs”), “newsgroups,” “chat rooms,” and the “World Wide Web.” All of these methods can be used to transmit text; most can transmit sound, pictures, and moving video images. Taken together, these tools constitute a unique medium — known to its users as “cyberspace” — located in no particular geographical location but available to anyone, anywhere in the world, with access to the Internet.

E-mail enables an individual to send an electronic message — generally akin to a note or letter — to another individual or to a group of addressees. The message is generally stored electronically, sometimes waiting for the recipient to check her “mailbox” and sometimes making its receipt known through some type of prompt. A mail exploder is a sort of e-mail group. Subscribers can send messages to a common e-mail address, which then forwards the message to the group’s other subscribers. Newsgroups also serve groups of regular participants, but these postings may be read by others as well. There are thousands of such groups, each serving to foster an exchange of information or opinion on a particular topic running the gamut from, say, the music of Wagner to Balkan politics to AIDS prevention to the Chicago Bulls. About 100,000 new messages are posted every day. In most newsgroups, postings are automatically purged at regular intervals. In addition to posting a message that can be read later, two or more individuals wishing to communicate more immediately can enter a chat room to engage in real-time dialogue — in other words, by typing messages to one another that appear almost immediately on *852 the others’ computer screens. The District Court found that at any given time “tens of thousands of users are engaging in conversations on a huge range of subjects.”6 It is “no exaggeration to conclude that the content on the Internet is as diverse as human thought.”7

The best known category of communication over the Internet is the World Wide Web, which allows users to search for and retrieve information stored in remote computers, as well as, in some cases, to communicate back to designated sites. In concrete terms, the Web consists of a vast number of documents stored in different computers all over the world. Some of these documents are simply files containing information. However, more elaborate documents, commonly known as Web “pages,” are also prevalent. Each has its own address — “rather like a telephone number.”8 Web pages frequently contain information and sometimes allow the viewer to communicate with the page’s (or “site’s”) author. They generally also contain “links” to other documents created by that site’s author or to other (generally) related sites. Typically, the links are either blue or underlined text — sometimes images.

Navigating the Web is relatively straightforward. A user may either type the address of a known page or enter one or more keywords into a commercial “search engine” in an effort to locate sites on a subject of interest. A particular Web page may contain the information sought by the “surfer,” or, through its links, it may be an avenue to other documents located anywhere on the Internet. Users generally explore a given Web page, or move to another, by clicking a computer “mouse” on one of the page’s icons or links. Access to most Web pages is freely available, but some allow access only to those who have purchased the right from a *853 commercial provider. The Web is thus comparable, from the readers’ viewpoint, to both a vast library including millions of readily available and indexed publications and a sprawling mall offering goods and services.

From the publishers’ point of view, it constitutes a vast platform from which to address and hear from a worldwide audience of millions of readers, viewers, researchers, and buyers. Any person or organization with a computer connected to the Internet can “publish” information. Publishers include government agencies, educational institutions, commercial entities, advocacy groups, and individuals.9 Publishers may either make their material available to the entire pool of Internet users, or confine access to a selected group, such as those willing to pay for the privilege. “No single organization controls any membership in the Web, nor is there any single centralized point from which individual Web sites or services can be blocked from the Web.”10

Sexually Explicit Material

Sexually explicit material on the Internet includes text, pictures, and chat and “extends from the modestly titillating to the hardest-core.”11 These files are created, named, and posted in the same manner as material that is not sexually explicit, and may be accessed either deliberately or unintentionally during the course of an imprecise search. “Once a provider posts its content on the Internet, it cannot prevent that content from entering any community.”12 Thus, for example,

*854 “when the UCR/California Museum of Photography posts to its Web site nudes by Edward Weston and Robert Mapplethorpe to announce that its new exhibit will travel to Baltimore and New York City, those images are available not only in Los Angeles, Baltimore, and New York City, but also in Cincinnati, Mobile, or Beijing — wherever Internet users live. Similarly, the safer sex instructions that Critical Path posts to its Web site, written in street language so that the teenage receiver can understand them, are available not just in Philadelphia, but also in Provo and Prague.”13

Some of the communications over the Internet that originate in foreign countries are also sexually explicit.14

Though such material is widely available, users seldom encounter such content accidentally. “A document’s title or a description of the document will usually appear before the document itself . . . and in many cases the user will receive detailed information about a site’s content before he or she need take the step to access the document. Almost all sexually explicit images are preceded by warnings as to the content.”15 For that reason, the “odds are slim” that a user would enter a sexually explicit site by accident.16 Unlike communications received by radio or television, “the receipt of information on the Internet requires a series of affirmative steps more deliberate and directed than merely turning a dial. A child requires some sophistication’ and some ability to read to retrieve material and thereby to use the Internet unattended.”17

Systems have been developed to help parents control the material that may be available on a home computer with In *855 ternet access. A system may either limit a computer’s access to an approved list of sources that have been identified as containing no adult material, it may block designated inappropriate sites, or it may attempt to block messages containing identifiable objectionable features. “Although parental control software currently can screen for certain suggestive words or for known sexually explicit sites, it cannot now screen for sexually explicit images.”18 Nevertheless, the evidence indicates that “a reasonably effective method by which parents can prevent their children from accessing sexually explicit and other material which parents may believe is inappropriate for their children will soon be widely available.”19

Age Verification

The problem of age verification differs for different uses of the Internet. The District Court categorically determined that there “is no effective way to determine the identity or the age of a user who is accessing material through e-mail, mail exploders, newsgroups or chat rooms.”20 The Government offered no evidence that there was a reliable way to screen recipients and participants in such forums for *856 age. Moreover, even if it were technologically feasible to block minors’ access to newsgroups and chat rooms containing discussions of art, politics, or other subjects that potentially elicit “indecent” or “patently offensive” contributions, it would not be possible to block their access to that material and “still allow them access to the remaining content, even if the overwhelming majority of that content was not indecent.”21

Technology exists by which an operator of a Web site may condition access on the verification of requested information such as a credit card number or an adult password. Credit card verification is only feasible, however, either in connection with a commercial transaction in which the card is used, or by payment to a verification agency. Using credit card possession as a surrogate for proof of age would impose costs on noncommercial Web sites that would require many of them to shut down. For that reason, at the time of the trial, credit card verification was “effectively unavailable to a substantial number of Internet content providers.” 929 F. Supp., at 846 (finding 102). Moreover, the imposition of such a requirement “would completely bar adults who do not have a credit card and lack the resources to obtain one from accessing any blocked material.”22

Commercial pornographic sites that charge their users for access have assigned them passwords as a method of age verification. The record does not contain any evidence concerning the reliability of these technologies. Even if passwords are effective for commercial purveyors of indecent material, the District Court found that an adult password requirement would impose significant burdens on noncommercial sites, both because they would discourage users from accessing their sites and because the cost of creating and *857 maintaining such screening systems would be “beyond their reach.”23

In sum, the District Court found:

“Even if credit card verification or adult password verification were implemented, the Government presented no testimony as to how such systems could ensure that the user of the password or credit card is in fact over 18. The burdens imposed by credit card verification and adult password verification systems make them effectively unavailable to a substantial number of Internet content providers.” Ibid. (finding 107).

II

The Telecommunications Act of 1996, Pub. L. 104-104, 110 Stat. 56, was an unusually important legislative enactment. As stated on the first of its 103 pages, its primary purpose was to reduce regulation and encourage “the rapid deployment of new telecommunications technologies.” The major components of the statute have nothing to do with the Internet; they were designed to promote competition in the local telephone service market, the multichannel video mar *858 ket, and the market for over-the-air broadcasting. The Act includes seven Titles, six of which are the product of extensive committee hearings and the subject of discussion in Reports prepared by Committees of the Senate and the House of Representatives. By contrast, Title V — known as the “Communications Decency Act of 1996” (CDA) — contains provisions that were either added in executive committee after the hearings were concluded or as amendments offered during floor debate on the legislation. An amendment offered in the Senate was the source of the two statutory provisions challenged in this case.24 They are informally de *859 scribed as the “indecent transmission” provision and the “patently offensive display” provision.25

The first, 47 U. S. C. § 223(a) (1994 ed., Supp. II), prohibits the knowing transmission of obscene or indecent messages to any recipient under 18 years of age. It provides in pertinent part:

“(a) Whoever—

“(1) in interstate or foreign communications—

“(B) by means of a telecommunications device knowingly—

“(i) makes, creates, or solicits, and

“(ii) initiates the transmission of,

“any comment, request, suggestion, proposal, image, or other communication which is obscene or indecent, knowing that the recipient of the communication is under 18 years of age, regardless of whether the maker of such communication placed the call or initiated the communication;

“(2) knowingly permits any telecommunications facility under his control to be used for any activity prohibited by paragraph (1) with the intent that it be used for such activity,

“shall be fined under Title 18, or imprisoned not more than two years, or both.”

The second provision, § 223(d), prohibits the knowing sending or displaying of patently offensive messages in a manner that is available to a person under 18 years of age. It provides:

*860 “(d) Whoever—

“(1) in interstate or foreign communications knowingly—

“(A) uses an interactive computer service to send to a specific person or persons under 18 years of age, or “(B) uses any interactive computer service to display in a manner available to a person under 18 years of age, “any comment, request, suggestion, proposal, image, or other communication that, in context, depicts or describes, in terms patently offensive as measured by contemporary community standards, sexual or excretory activities or organs, regardless of whether the user of such service placed the call or initiated the communication; or

“(2) knowingly permits any telecommunications facility under such person’s control to be used for an activity prohibited by paragraph (1) with the intent that it be used for such activity,

“shall be fined under Title 18, or imprisoned not more than two years, or both.”

The breadth of these prohibitions is qualified by two affirmative defenses. See § 223(e)(5).26 One covers those who take “good faith, reasonable, effective, and appropriate actions” to restrict access by minors to the prohibited communications. § 223(e)(5)(A). The other covers those who *861 restrict access to covered material by requiring certain designated forms of age proof, such as a verified credit card or an adult identification number or code. § 223(e)(5)(B).

I — I HH

On February 8, 1996, immediately after the President signed the statute, 20 plaintiffs27 filed suit against the Attorney General of the United States and the Department of Justice challenging the constitutionality of §§ 223(a)(1) and 223(d). A week later, based on his conclusion that the term “indecent” was too vague to provide the basis for a criminal prosecution, District Judge Buckwalter entered a temporary restraining order against enforcement of § 223(a)(1)(B)(ii) insofar as it applies to indecent communications. A second suit was then filed by 27 additional plaintiffs,28 the two cases *862 were consolidated, and a three-judge District Court was convened pursuant to § 561 of the CDA.29 After an evidentiary hearing, that court entered a preliminary injunction against enforcement of both of the challenged provisions. Each of the three judges wrote a separate opinion, but their judgment was unanimous.

Chief Judge Sloviter doubted the strength of the Government’s interest in regulating “the vast range of online material covered or potentially covered by the CDA,” but acknowledged that the interest was “compelling” with respect to some of that material. 929 F. Supp.,' at 853. She concluded, nonetheless, that the statute “sweeps more broadly than necessary and thereby chills the expression of adults” and that the terms “patently offensive” and “indecent” were “inherently vague.” Id., at 854. She also determined that the affirmative defenses were not “technologically or economically feasible for most providers,” specifically considering and rejecting an argument that providers could avoid liability by “tagging” their material in a manner that would allow potential readers to screen out unwanted transmissions. Id., at 856. Chief Judge Sloviter also rejected the Government’s suggestion that the scope of the statute could be narrowed by construing it to apply only to commercial pornographers. Id., at 854-855.

Judge Buckwalter concluded that the word “indecent” in § 223(a)(1)(B) and the terms “patently offensive” and “in context” in § 223(d)(1) were so vague that criminal enforcement of either section would violate the “fundamental constitutional principle” of “simple fairness,” id., at 861, and the specific protections of the First and Fifth Amendments, id., at 858. He found no statutory basis for the Government’s argument that the challenged provisions would be applied only to “pornographic” materials, noting that, unlike obscenity, “indecency has not been defined to exclude works of serious literary, artistic, political or scientific value.” Id., at 863. *863 Moreover, the Government’s claim that the work must be considered patently offensive “in context” was itself vague because the relevant context might “refer to, among other things, the nature of the communication as a whole, the time of day it was conveyed, the medium used, the identity of the speaker, or whether or not it is accompanied by appropriate warnings.” Id., at 864. He believed that the unique nature of the Internet aggravated the vagueness of the statute. Id., at 865, n. 9.

Judge Dalzell’s review of “the special attributes of Internet communication” disclosed by the evidence convinced him that the First Amendment denies Congress the power to regulate the content of protected speech on the Internet. Id., at 867. His opinion explained at length why he believed the CDA would abridge significant protected speech, particularly by noncommercial speakers, while “[p]erversely, commercial pornographers would remain relatively unaffected.” Id., at 879. He construed our cases as requiring a “medium-specific” approach to the analysis of the regulation of mass communication, id., at 873, and concluded that the Internet — as “the most participatory form of mass speech yet developed,” id., at 883 — is entitled to “the highest protection from governmental intrusion,” ibid.30

*864 The judgment of the District Court enjoins the Government from enforcing the prohibitions in § 223(a)(1)(B) insofar as they relate to “indecent” communications, but expressly preserves the Government’s right to investigate and prosecute the obscenity or child pornography activities prohibited therein. The injunction against enforcement of §§ 223(d)(1) and (2) is unqualified because those provisions contain no separate reference to obscenity or child pornography.

The Government appealed under the CDA’s special review provisions, § 561, 110 Stat. 142-143, and we noted probable jurisdiction, see 519 U. S. 1025 (1996). In its appeal, the Government argues that the District Court erred in holding that the CDA violated both the First Amendment because it is overbroad and the Fifth Amendment because it is vague. While we discuss the vagueness of the CDA because of its relevance to the First Amendment overbreadth inquiry, we conclude that the judgment should be affirmed without reaching the Fifth Amendment issue. We begin our analysis by reviewing the principal authorities on which the Government relies. Then, after describing the overbreadth of the CDA, we consider the Government’s specific contentions, including its submission that we save portions of the statute either by severance or by fashioning judicial limitations on the scope of its coverage.

IV

In arguing for reversal, the Government contends that the CDA is plainly constitutional under three of our prior decisions: (1) Ginsberg v. New York, 390 U. S. 629 (1968); (2) FCC v. Pacifica Foundation, 438 U. S. 726 (1978); and (3) Renton v. Playtime Theatres, Inc., 475 U. S. 41 (1986). A close look at these cases, however, raises — rather than relieves— doubts concerning the constitutionality of the CDA.

In Ginsberg, we upheld the constitutionality of a New York statute that prohibited selling to minors under 17 years of age material that was considered obscene as to them even if not obscene as to adults. We rejected the defendant’s broad *865 submission that “the scope of the constitutional freedom of expression secured to a citizen to read or see material concerned with sex cannot be made to depend on whether the citizen is an adult or a minor.” 390 U. S., at 636. In rejecting that contention, we relied not only on the State’s independent interest in the well-being of its youth, but also on our consistent recognition of the principle that “the parents’ claim to authority in their own household to direct the rearing of their children is basic in the structure of our society.”31

In four important respects, the statute upheld in Ginsberg was narrower than the CDA. First, we noted in Ginsberg that “the prohibition against sales to minors does not bar parents who so desire from purchasing the magazines for their children.” Id., at 639. Under the CDA, by contrast, neither the parents’ consent — nor even their participation— in the communication would avoid the application of the statute.32 Second, the New York statute applied only to commercial transactions, id., at 647, whereas the CDA contains no such limitation. Third, the New York statute cabined its definition of material that is harmful to minors with the requirement that it be “utterly without redeeming social importance for minors.” Id., at 646. The CDA fails to provide us with any definition of the term “indecent” as used in § 223(a)(1) and, importantly, omits any requirement that the “patently offensive” material covered by § 223(d) lack serious literary, artistic, political, or scientific value. Fourth, the New York statute defined a minor as a person under the age *866 of 17, whereas the CDA, in applying to all those under 18 years, includes an additional year of those nearest majority.

In Pacifica, we upheld a declaratory order of the Federal Communications Commission, holding that the broadcast of a recording of a 12-minute monologue entitled “Filthy Words” that had previously been delivered to a live audience “could have been the subject of administrative sanctions.” 438 U. S., at 730 (internal quotation marks omitted). The Commission had found that the repetitive use of certain words referring to excretory or sexual activities or organs “in an afternoon broadcast when children are in the audience was patently offensive” and concluded that the monologue was indecent “as broadcast.” Id., at 735. The respondent did not quarrel with the finding that the afternoon broadcast was patently offensive, but contended that it was not “indecent” within the meaning of the relevant statutes because it contained no prurient appeal. After rejecting respondent’s statutory arguments, we confronted its two constitutional arguments: (1) that the Commission’s construction of its authority to ban indecent speech was so broad that its order had to be set aside even if the broadcast at issue was unprotected; and (2) that since the recording was not obscene, the First Amendment forbade any abridgment of the right to broadcast it on the radio.

In the portion of the lead opinion not joined by Justices Powell and Blackmun, the plurality stated that the First Amendment does not prohibit all governmental regulation that depends on the content of speech. Id., at 742-743. Accordingly, the availability of constitutional protection for a vulgar and offensive monologue that was not obscene depended on the context of the broadcast. Id., at 744-748. Relying on the premise that “of all forms of communication” broadcasting had received the most limited First Amendment protection, id., at 748-749, the Court concluded that the ease with which children may obtain access to broadcasts, *867 “coupled with the concerns recognized in Ginsberg,” justified special treatment of indecent broadcasting. Id., at 749-750.

As with the New York statute at issue in Ginsberg, there are significant differences between the order upheld in Pa-cifica and the CDA. First, the order in Pacifica, issued by an agency that had been regulating radio stations for decades, targeted a specific broadcast that represented a rather dramatic departure from traditional program content in order to designate when — rather than whether — it would be permissible to air such a program in that particular medium. The CDA’s broad categorical prohibitions are not limited to particular times and are not dependent on any evaluation by an agency familiar with the unique characteristics of the Internet. Second, unlike the CDA, the Commission’s declaratory order was not punitive; we expressly refused to decide whether the indecent broadcast “would justify a criminal prosecution.” 438 U. S., at 750. Finally, the Commission’s order applied to a medium which as a matter of history had “received the most limited First Amendment protection,” id., at 748, in large part because warnings could not adequately protect the listener from unexpected program content. The Internet, however, has no comparable history. Moreover, the District Court found that the risk of encountering indecent material by accident is remote because a series of affirmative steps is required to access specific material.

In Renton, we upheld a zoning ordinance that kept adult movie theaters out of residential neighborhoods. The ordinance was aimed, not at the content of the films shown in the theaters, but rather at the “secondary effects” — such as crime and deteriorating property values — that these theaters fostered: “Tt is th[e] secondary effect which these zoning ordinances attempt to avoid, not the dissemination of “offensive” speech.’” 475 U. S., at 49 (quoting Young v. American Mini Theatres, Inc., 427 U. S. 50, 71, n. 34 (1976)). According to the Government, the CDA is constitutional be *868 cause it constitutes a sort of “cyberzoning” on the Internet. But the CDA applies broadly to the entire universe of cyberspace. And the purpose of the CDA is to protect children from the primary effects of “indecent” and “patently offensive” speech, rather than any “secondary” effect of such speech. Thus, the CDA is a content-based blanket restriction on speech, and, as such, cannot be “properly analyzed as a form of time, place, and manner regulation.” 475 U. S., at 46. See also Boos v. Barry, 485 U. S. 812, 321 (1988) (“Regulations that focus on the direct impact of speech on its audience” are not properly analyzed under Renton); Forsyth County v. Nationalist Movement, 505 U. S. 123, 134 (1992) (“Listeners’ reaction to speech is not a content-neutral basis for regulation”).

These precedents, then, surely do not require us to uphold the CDA and are fully consistent with the application of the most stringent review of its provisions.

V

In Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 557 (1975), we observed that “[e]ach medium of expression ... may present its own problems.” Thus, some of our cases have recognized special justifications for regulation of the broadcast media that are not applicable to other speakers, see Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969); FCC v. Pacifica Foundation, 438 U. S. 726 (1978). In these cases, the Court relied on the history of extensive Government regulation of the broadcast medium, see, e. g., Red Lion, 395 U. S., at 399-400; the scarcity of available frequencies at its inception, see, e. g., Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 637-638 (1994); and its “invasive” nature, see Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 128 (1989).

Those factors are not present in cyberspace. Neither before nor after the enactment of the CDA have the vast democratic forums of the Internet been subject to the type *869 of government supervision and regulation that has attended the broadcast industry.33 Moreover, the Internet is not as “invasive” as radio or television. The District Court specifically found that “[cjommunications over the Internet do not ‘invade’ an individual’s home or appear on one’s computer screen unbidden. Users seldom encounter content ‘by accident.’ ” 929 F. Supp., at 844 (finding 88). It also found that “[ajlmost all sexually explicit images are preceded by warnings as to the content,” and cited testimony that “ ‘odds are slim’ that a user would come across a sexually explicit sight by accident.” Ibid.

We distinguished Pacifica in Sable, 492 U. S., at 128, on just this basis. In Sable, a company engaged in the business of offering sexually oriented prerecorded telephone messages (popularly known as “dial-a-porn”) challenged the constitutionality of an amendment to the Communications Act of 1934 that imposed a blanket prohibition on indecent as well as obscene interstate commercial telephone messages. We held that the statute was constitutional insofar as it applied to obscene messages but invalid as applied to indecent messages. In attempting to justify the complete ban and criminalization of indecent commercial telephone messages, the Government relied on Pacifica, arguing that the ban was necessary to prevent children from gaining access to such messages. We agreed that “there is a compelling interest in protecting the physical and psychological well-being of minors” which extended to shielding them from indecent messages that are not obscene by adult standards, 492 U. S., at *870 126, but distinguished our “emphatically narrow holding” in Pacifica because it did not involve a complete ban and because it involved a different medium of communication, id., at 127. We explained that “the dial-it medium requires the listener to take affirmative steps to receive the communication.” Id., at 127-128. “Placing a telephone call,” we continued, “is not the same as turning on a radio and being taken by surprise by an indecent message.” Id., at 128.

Finally, unlike the conditions that prevailed when Congress first authorized regulation of the broadcast spectrum, the Internet can hardly be considered a “scarce” expressive commodity. It provides relatively unlimited, low-cost capacity for communication of all kinds. The Government estimates that “[a]s many as 40 million people use the Internet today, and that figure is expected to grow to 200 million by 1999.”34 This dynamic, multifaceted category of communication includes not only traditional print and news services, but also audio, video, and still images, as well as interactive, real-time dialogue. Through the use of chat rooms, any person with a phone line can become a town crier with a voice that resonates farther than it could from any soapbox. Through the use of Web pages, mail exploders, and newsgroups, the same individual can become a pamphleteer. As the District Court found, “the content on the Internet is as diverse as human thought.” 929 F. Supp., at 842 (finding 74). We agree with its conclusion that our cases provide no basis for qualifying the level of First Amendment scrutiny that should be applied to this medium.

H-i >

Regardless of whether the CDA is so vague that it violates the Fifth Amendment, the many ambiguities concerning the scope of its coverage render it problematic for purposes of the First Amendment. For instance, each of the two parts *871 of the CDA uses a different linguistic form. The first uses the word “indecent,” 47 U. S. C. § 223(a) (1994 ed., Supp. II), while the second speaks of material that “in context, depicts or describes, in terms patently offensive as measured by contemporary community standards, sexual or excretory activities or organs,” § 223(d). Given the absence of a definition of either term,35 this difference in language will provoke uncertainty among speakers about how the two standards relate to each other36 and just what they mean.37 Could a speaker confidently assume that a serious discussion about birth control practices, homosexuality, the First Amendment issues raised by the Appendix to our Pacifica opinion, or the consequences of prison rape would not violate the CDA? This uncertainty undermines the likelihood that the CDA has been carefully tailored to the congressional goal of protecting minors from potentially harmful materials.

The vagueness of the CDA is a matter of special concern for two reasons. First, the CDA is a content-based regulation of speech. The vagueness of such a regulation raises *872 special First Amendment concerns because of its obvious chilling effect on free speech. See, e. g., Gentile v. State Bar of Nev., 501 U. S. 1030, 1048-1051 (1991). Second, the CDA is a criminal statute. In addition to the opprobrium and stigma of a criminal conviction, the CDA threatens violators with penalties including up to two years in prison for each act of violation. The severity of criminal sanctions may well cause speakers to remain silent rather than communicate even arguably unlawful words, ideas, and images. See, e. g., Dombrowski v. Pfister, 380 U. S. 479, 494 (1965). As a practical matter, this increased deterrent effect, coupled with the “risk of discriminatory enforcement” of vague regulations, poses greater First Amendment concerns than those implicated by the civil regulation reviewed in Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U. S. 727 (1996).

The Government argues that the statute is no more vague than the obscenity standard this Court established in Miller v. California, 413 U. S. 15 (1973). But that is not so. In Miller, this Court reviewed a criminal conviction against a commercial vendor who mailed brochures containing pictures of sexually explicit activities to individuals who had not requested such materials. Id., at 18. Having struggled for some time to establish a definition of obscenity, we set forth in Miller the test for obscenity that controls to this day:

“(a) whether the average person, applying contemporary community standards would find that the work, taken as a whole, appeals to the prurient interest; (b) whether the work depicts or describes, in a patently offensive way, sexual conduct specifically defined by the applicable state law; and (c) whether the work, taken as a whole, lacks serious literary, artistic, political, or scientific value.” Id., at 24 (internal quotation marks and citations omitted).

*873 Because the CDA’s “patently offensive” standard (and, we assume, arguendo, its synonymous “indecent” standard) is one part of the three-prong Miller test, the Government reasons, it cannot be unconstitutionally vague.

The Government’s assertion is incorrect as a matter of fact. The second prong of the Miller test — the purportedly analogous standard — contains a critical requirement that is omitted from the CDA: that the proscribed material be “specifically defined by the applicable state law.” This requirement reduces the vagueness inherent in the open-ended term “patently offensive” as used in the CDA. Moreover, the Miller definition is limited to “sexual conduct,” whereas the CDA extends also to include (1) “excretory activities” as well as (2) “organs” of both a sexual and excretory nature.

The Government’s reasoning is also flawed. Just because a definition including three limitations is not vague, it does not follow that one of those limitations, standing by itself, is not vague.38 Each of Miller’s additional two prongs — (1) that, taken as a whole, the material appeal to the “prurient” interest, and (2) that it “lac[k] serious literary, artistic, political, or scientific value” — critically limits the uncertain sweep of the obscenity definition. The second requirement is particularly important because, unlike the “patently offensive” and “prurient interest” criteria, it is not judged by contemporary community standards. See Pope v. Illinois, 481 U. S. 497, 500 (1987). This “societal value” requirement, absent in the CDA, allows appellate courts to impose some limitations and regularity on the definition by setting, as a matter of law, a national floor for socially redeeming value. The Government’s contention that courts will be able to give such legal limitations to the CDA’s standards is belied by Miller’s own rationale for having juries determine whether material *874 is “patently offensive” according to community standards: that such questions are essentially ones of fact.39

In contrast to Miller and our other previous cases, the CDA thus presents a greater threat of censoring speech that, in fact, falls outside the statute’s scope. Given the vague contours of the coverage of the statute, it unquestionably silences some speakers whose messages would be entitled to constitutional protection. That danger provides further reason for insisting that the statute not be overly broad. The CDA’s burden on protected speech cannot be justified if it could be avoided by a more carefully drafted statute.

VII

We are persuaded that the CDA lacks the precision that the First Amendment requires when a statute regulates the content of speech. In order to deny minors access to potentially harmful speech, the CDA effectively suppresses a large amount of speech that adults have a constitutional right to receive and to address to one another. That burden on adult speech is unacceptable if less restrictive alternatives would be at least as effective in achieving the legitimate purpose that the statute was enacted to serve.

In evaluating the free speech rights of adults, we have made it perfectly clear that “[s]exual expression which is indecent but not obscene is protected by the First Amendment.” Sable, 492 U. S., at 126. See also Carey v. Population Services Int’l, 431 U. S. 678, 701 (1977) (“[Wjhere obscenity is not involved, we have consistently held that the *875 fact that protected speech may be offensive to some does not justify its suppression”). Indeed, Pacifica itself admonished that “the fact that society may find speech offensive is not a sufficient reason for suppressing it.” 438 U. S., at 745.

It is true that we have repeatedly recognized the governmental interest in protecting children from harmful materials. See Ginsberg, 390 U. S., at 639; Pacifica, 438 U. S., at 749. But that interest does not justify an unnecessarily broad suppression of speech addressed to adults. As we have explained, the Government may not “reduc[e] the adult population ... to .. . only what is fit for children.” Denver, 518 U. S., at 759 (internal quotation marks omitted) (quoting Sable, 492 U. S., at 128).40 “[Rjegardless of the strength of the government’s interest” in protecting children, “[t]he level of discourse reaching a mailbox simply cannot be limited to that which would be suitable for a sandbox.” Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 74-75 (1983).

The District Court was correct to conclude that the CDA effectively resembles the ban on “dial-a-porn” invalidated in Sable. 929 F. Supp., at 854. In Sable, 492 U. S., at 129, this Court rejected the argument that we should defer to the congressional judgment that nothing less than a total ban would be effective in preventing enterprising youngsters from gaining access to indecent communications. Sable thus made clear that the mere fact that a statutory regulation of speech was enacted for the important purpose of protecting children from exposure to sexually explicit material does not foreclose inquiry into its validity.41 As we pointed out last *876 Term, that inquiry embodies an “overarching commitment” to make sure that Congress has designed its statute to accomplish its purpose “without imposing an unnecessarily great restriction on speech.” Denver, 518 U. S., at 741.

In arguing that the CDA does not so diminish adult communication, the Government relies on the incorrect factual premise that prohibiting a transmission whenever it is known that one of its recipients is a minor would not interfere with adult-to-adult communication. The findings of the District Court make clear that this premise is untenable. Given the size of the potential audience for most messages, in the absence of a viable age verification process, the sender must be charged with knowing that one or more minors will likely view it. Knowledge that, for instance, one or more members of a 100-person chat group will be a minor — and therefore that it would be a crime to send the group an indecent message — would surely burden communication among adults.42

The District Court found that at the time of trial existing technology did not include any effective method for a sender to prevent minors from obtaining access to its communications on the Internet without also denying access to adults. The Court found no effective way to determine the age of a user who is accessing material through e-mail, mail exploders, newsgroups, or chat rooms. 929 F. Supp., at 845 (findings 90-94). As a practical matter, the Court also found *877 that it would be prohibitively expensive for noncommercial— as well as some commercial — speakers who have Web sites to verify that their users are adults. Id., at 845-848 (findings 95-116).43 These limitations must inevitably curtail a significant amount of adult communication on the Internet. By contrast, the District Court found that “[d]espite its limitations, currently available user-based software suggests that a reasonably effective method by which parents can prevent their children from accessing sexually explicit and other material which parents may believe is inappropriate for their children will soon be widely available.” Id., at 842 (finding 73) (emphases added).

The breadth of the CDA’s coverage is wholly unprecedented. Unlike the regulations upheld in Ginsberg and Pa-cifica, the scope of the CDA is not limited to commercial speech or commercial entities. Its open-ended prohibitions embrace all nonprofit entities and individuals posting indecent messages or displaying them on their own computers in the presence of minors. The general, undefined terms “indecent” and “patently offensive” cover large amounts of non-pornographic material with serious educational or other value.44 Moreover, the “community standards” criterion as applied to the Internet means that any communication avail *878 able to a nationwide audience will be judged by the standards of the community most likely to be offended by the message.45 The regulated subject matter includes any of the seven “dirty words” used in the Pacifica monologue, the use of which the Government’s expert acknowledged could constitute a felony. See Olsen Testimony, Tr. Vol. V, 53:16-54:10. It may also extend to discussions about prison rape or safe sexual practices, artistic images that include nude subjects, and arguably the card catalog of the Carnegie Library.

For the purposes of our decision, we need neither accept nor reject the Government’s submission that the First Amendment does not forbid a blanket prohibition on all “indecent” and “patently offensive” messages communicated to a 17-year-old — no matter how much value the message may contain and regardless of parental approval. It is at least clear that the strength of the Government’s interest in protecting minors is not equally strong throughout the coverage of this broad statute. Under the CDA, a parent allowing her 17-year-old to use the family computer to obtain information on the Internet that she, in her parental judgment, deems appropriate could face a lengthy prison term. See 47 U. S. C. § 223(a)(2) (1994 ed., Supp. II). Similarly, a parent who sent his 17-year-old college freshman information on birth control via e-mail could be incarcerated even though neither he, his child, nor anyone in their home community found the material “indecent” or “patently offensive,” if the college town’s community thought otherwise.

*879 The breadth of this content-based restriction of speech imposes an especially heavy burden on the Government to explain why a less restrictive provision would not be as effective as the CDA. It has not done so. The arguments in this Court have referred to possible alternatives such as requiring that indecent material be “tagged” in a way that facilitates parental control of material coming into their homes, making exceptions for messages with artistic or educational value, providing some tolerance for parental choice, and regulating some portions of the Internet — such as commercial Web sites — differently from others, such as chat rooms. Particularly in the light of the absence of any detailed findings by the Congress, or even hearings addressing the special problems of the CDA, we are persuaded that the CDA is not narrowly tailored if that requirement has any meaning at all.

VIII

In an attempt to curtail the CDA’s facial overbreadth, the Government advances three additional arguments for sustaining the Act’s affirmative prohibitions: (1) that the CDA is constitutional because it leaves open ample “alternative channels” of communication; (2) that the plain meaning of the CDA’s “knowledge” and “specific person” requirement significantly restricts its permissible applications; and (3) that the CDA’s prohibitions are “almost always” limited to material lacking redeeming social value.

The Government first contends that, even though the CDA effectively censors discourse on many of the Internet’s modalities — such as chat groups, newsgroups, and mail exploders — it is nonetheless constitutional because it provides a “reasonable opportunity” for speakers to engage in the restricted speech on the World Wide Web. Brief for Appellants 39. This argument is unpersuasive because the CDA regulates speech on the basis of its content. A “time, place, and manner” analysis is therefore inapplicable. See Consolidated Edison Co. of N. Y v. Public Serv. Comm’n of N Y., *880 447 U. S. 530, 536 (1980). It is thus immaterial whether such speech would be feasible on the Web (which, as the Government’s own expert acknowledged, would cost up to $10,000 if the speaker’s interests were not accommodated by an existing Web site, not including costs for data base management and age verification). The Government’s position is equivalent to arguing that a statute could ban leaflets on certain subjects as long as individuals are free to publish books. In invalidating a number of laws that banned leafletting on the streets regardless of their content, we explained that “one is not to have the exercise of his liberty of expression in appropriate places abridged on the plea that it may be exercised in some other place.” Schneider v. State (Town of Irvington), 308 U. S. 147, 163 (1939).

The Government also asserts that the “knowledge” requirement of both §§ 223(a) and (d), especially when coupled with the “specific child” element found in § 223(d), saves the CDA from overbreadth. Because both sections prohibit the dissemination of indecent messages only to persons known to be under 18, the Government argues, it does not require transmitters to “refrain from communicating indecent material to adults; they need only refrain from disseminating such materials to persons they know to be under 18.” Brief for Appellants 24. This argument ignores the fact that most Internet forums — including chat rooms, newsgroups, mail exploders, and the Web — are open to all comers. The Government’s assertion that the knowledge requirement somehow protects the communications of adults is therefore untenable. Even the strongest reading of the “specific person” requirement of § 223(d) cannot save the statute. It would confer broad powers of censorship, in the form of a “heckler’s veto,” upon any opponent of indecent speech who might simply log on and inform the would-be discoursers that his 17-year-old child — a “specific person . . . under 18 years of age,” 47 U. S. C. § 223(d)(1)(A) (1994 ed., Supp. II) — would be present.

*881 Finally, we find no textual support for the Government’s submission that material having scientific, educational, or other redeeming social value will necessarily fall outside the CDA’s “patently offensive” and “indecent” prohibitions. See also n. 37, supra.

IX

The Government’s three remaining arguments focus on the defenses provided in § 223(e)(5).46 First, relying on the “good faith, reasonable, effective, and appropriate actions” provision, the Government suggests that “tagging” provides a defense that saves the constitutionality of the CDA. The suggestion assumes that transmitters may encode their indecent communications in a way that would indicate their contents, thus permitting recipients to block their reception with appropriate software. It is the requirement that the good-faith action must be “effective” that makes this defense illusory. The Government recognizes that its proposed screening software does not currently exist. Even if it did, there is no way to know whether a potential recipient will actually block the encoded material. Without the impossible knowledge that every guardian in America is screening for the “tag,” the transmitter could not reasonably rely on its action to be “effective.”

For its second and third arguments concerning defenses— which we can consider together — the Government relies on the latter half of § 223(e)(5), which applies when the transmitter has restricted access by requiring use of a verified credit card or adult identification. Such verification is not only technologically available but actually is used by commercial providers of sexually explicit material. These providers, therefore, would be protected by the defense. Under the findings of the District Court, however, it is not economically feasible for most noncommercial speakers to employ such verification. Accordingly, this defense would not signifi *882 cantly narrow the statute’s burden on noncommercial speech. .Even with respect to the commercial pornographers that would be protected by the defense, the Government failed to adduce any evidence that these verification techniques actually preclude minors from posing as adults.47 Given that the risk of criminal sanctions “hovers over each content provider, like the proverbial sword of Damocles,”48 the District Court correctly refused to rely on unproven future technology to save the statute. The Government thus failed to prove that the proffered defense would significantly reduce the heavy burden on adult speech produced by the prohibition on offensive displays.

We agree with the District Court’s conclusion that the CDA places an unacceptably heavy burden on protected speech, and that the defenses do not constitute the sort of “narrow tailoring” that will save an otherwise patently invalid unconstitutional provision. In Sable, 492 U. S., at 127, we remarked that the speech restriction at issue there amounted to “‘burn[ing] the house to roast the pig.’” The CDA, casting a far darker shadow over free speech, threatens to torch a large segment of the Internet community.

X

At oral argument, the Government relied heavily on its ultimate fail-back position: If this Court should conclude that the CDA is insufficiently tailored, it urged, we should save the statute’s constitutionality by honoring the severability clause, see 47 U. S. C. § 608, and construing nonseverable terms narrowly. In only one respect is this argument acceptable.

A severability clause requires textual provisions that can be severed. We will follow § 608’s guidance by leaving con *883 stitutional textual elements of the statute intact in the one place where they are, in fact, severable. The “indecency” provision, 47 U. S. C. § 223(a) (1994 ed., Supp. II), applies to “any comment, request, suggestion, proposal, image, or other communication which is obscene or indecent.” (Emphasis added.) Appellees do not challenge the application of the statute to obscene speech, which, they acknowledge, can be banned totally because it enjoys no First Amendment protection. See Miller, 413 U. S., at 18. As set forth by the statute, the restriction of “obscene” material enjoys a textual manifestation separate from that for “indecent” material, which we have held unconstitutional. Therefore, we will sever the term “or indecent” from the statute, leaving the rest of § 223(a) standing. In no other respect, however, can § 223(a) or § 223(d) be saved by such a textual surgery.

The Government also draws on an additional, less traditional aspect of the CDA’s severability clause, 47 U. S. C. § 608, which asks any reviewing court that holds the statute facially unconstitutional not to invalidate the CDA in application to “other persons or circumstances” that might be constitutionally permissible. It further invokes this Court’s admonition that, absent “countervailing considerations,” a statute should “be declared invalid to the extent it reaches too far, but otherwise left intact.” Brockett v. Spokane Arcades, Inc., 472 U. S. 491, 503-504 (1985). There are two flaws in this argument.

First, the statute that grants our jurisdiction for this expedited review, § 561 of the Telecommunications Act of 1961, note following 47 U. S. C. § 223 (1994 ed., Supp. II), limits that jurisdictional grant to actions challenging the CDA “on its face.” Consistent with §561, the plaintiffs who brought this suit and the three-judge panel'that decided it treated it as a facial challenge. We have no authority, in this particular posture, to convert this litigation into an “as-applied” challenge. Nor, given the vast array of plaintiffs, the range of their expressive activities, and the vagueness of the stat *884 ute, would it be practicable to limit our holding to a judicially defined set of specific applications.

Second, one of the “countervailing considerations” mentioned in Brockett is present here. In considering a facial challenge, this Court may impose a limiting construction on a statute only if it is “readily susceptible” to such a construction. Virginia v. American Booksellers Assn., Inc., 484 U. S. 883, 397 (1988). See also Erznoznik v. Jacksonville, 422 U. S. 205, 216 (1975) (“readily subject” to narrowing construction). The open-ended character of the CDA provides no guidance whatever for limiting its coverage.

This case is therefore unlike those in which we have construed a statute narrowly because the text or other source of congressional intent identified a clear line that this Court could draw. Cf., e. g., Brockett, 472 U. S., at 504-505 (invalidating obscenity statute only to the extent that word “lust” was actually or effectively excised from statute); United States v. Grace, 461 U. S. 171, 180-183 (1983) (invalidating federal statute banning expressive displays only insofar as it extended to public sidewalks when clear line could be drawn between sidewalks and other grounds that comported with congressional purpose of protecting the building, grounds, and people therein). Rather, our decision in United States v. Treasury Employees, 513 U. S. 454, 479, n. 26 (1995), is applicable. In that case, we declined to “dra[w] one or more lines between categories of speech covered by an overly broad statute, when Congress has sent inconsistent signals as to where the new line or lines should be drawn” because doing so “involves a far more serious invasion of the legislative domain.”49 This Court “will not rewrite a . . . law *885 to conform it to constitutional requirements.” American Booksellers, 484 U. S., at 397.50

XI

In this Court, though not in the District Court, the Government asserts that — in addition to its interest in protecting children — its “[e]qually significant” interest in fostering the growth of the Internet provides an independent basis for upholding the constitutionality of the CDA. Brief for Appellants 19. The Government apparently assumes that the unregulated availability of “indecent” and “patently offensive” material on the Internet is driving countless citizens away from the medium because of the risk of exposing themselves or their children to harmful material.

We find this argument singularly unpersuasive. The dramatic expansion of this new marketplace of ideas contradicts the factual basis of this contention. The record demonstrates that the growth of the Internet has been and continues to be phenomenal. As a matter of constitutional tradition, in the absence of evidence to the contrary, we presume that governmental regulation of the content of speech is more likely to interfere with the free exchange of ideas than to encourage it. The interest in encouraging freedom of expression in a democratic society outweighs any theoretical but unproven benefit of censorship.

For the foregoing reasons, the judgment of the District Court is affirmed.

It is so ordered.

‘‘Congress shall make no law . . . abridging the freedom of speech.” U. S. Const., Amdt. 1.

The Court made 410 findings, including 356 paragraphs of the parties’ stipulation and 54 findings based on evidence received in open court. See 929 F. Supp., at 830, n. 9, 842, n. 15.

An acronym for the network developed by the Advanced Research Project Agency.

Id., at 844 (finding 81).

Id., at 831 (finding 3).

Id., at 835 (finding 27).

Id., at 842 (finding 74).

Id., at 836 (finding 36).

“Web publishing is simple enough that thousands of individual users and small community organizations are using the Web to publish their own personal ‘home pages,’ the equivalent of individualized newsletters about that person or organization, which are available to everyone on the Web.” Id., at 837 (finding 42).

Id., at 838 (finding 46).

Id., at 844 (finding 82).

Ibid. (finding 86).

Ibid. (finding 85).

Id., at 848 (finding 117).

Id., at 844-845 (finding 88).

Ibid.

Id., at 845 (finding 89).

Id., at 842 (finding 72).

Ibid. (finding 73).

Id., at 845 (finding 90): “An e-mail address provides no authoritative information about the addressee, who may use an e-mail ‘alias’ or an anonymous remailer. There is also no universal or reliable listing of e-mail addresses and corresponding names or telephone numbers, and any such listing would be or rapidly become incomplete. For these reasons, there is no reliable way in many instances for a sender to know if the e-mail recipient is an adult or a minor. The difficulty of e-mail age verification is compounded for mail exploders such as listservs, which automatically send information to all e-mail addresses on a sender’s list. Government expert Dr. Olsen agreed that no current technology could give a speaker assurance that only adults were listed in a particular mail exploder’s mailing list.”

Ibid. (finding 93).

Id., at 846 (finding 102).

Id., at 847 (findings 104-106):

“At least some, if not almost all, non-commercial organizations, such as the ACLU, Stop Prisoner Rape or Critical Path AIDS Project, regard charging listeners to access their speech as contrary to their goals of making their materials available to a wide audience free of charge.

“There is evidence suggesting that adult users, particularly casual Web browsers, would be discouraged from retrieving information that required use of a credit card or password. Andrew Anker testified that HotWired has received many complaints from its members about HotWired’s registration system, which requires only that a member supply a name, e-mail address and self-created password. There is concern by commercial content providers that age verification requirements would decrease advertising and revenue because advertisers depend on a demonstration that the sites are widely available and frequently visited.”

See Exon Amendment No. 1268, 141 Cong. Rec. 15536 (1995). See also id.., at 15505. This amendment, as revised, became § 502 of the Telecommunications Act of 1996, 110 Stat. 133, 47 U. S. C. §§223(a)-(e) (1994 ed., Supp. II). Some Members of the House of Representatives opposed the Exon Amendment because they thought it “possible for our parents now to child-proof the family computer with these products available in the private sector.” They also thought the Senate’s approach would “involve the Federal Government spending vast sums of money trying to define elusive terms that are going to lead to a flood of legal challenges while our kids are unprotected.” These Members offered an amendment intended as a substitute for the Exon Amendment, but instead enacted as an additional section of the Act entitled “Online Family Empowerment.” See 110 Stat. 137, 47 U.S.C. § 230 (1994 ed., Supp. II); 141 Cong. Rec. 27881 (1995). No hearings were held on the provisions that became law. See S. Rep. No. 104-23, p. 9 (1995). After the Senate adopted the Exon Amendment, however, its Judiciary Committee did conduct a one-day hearing on “Cyberporn and Children.” In his opening statement at that hearing, Senator Leahy observed:

“It really struck me in your opening statement when you mentioned, Mr. Chairman, that it is the first ever hearing, and you are absolutely right. And yet we had a major debate on the floor, passed legislation overwhelmingly on a subject involving the Internet, legislation that could dramatically change — some would say even wreak havoc — on the Internet. The Senate went in willy-nilly, passed legislation, and never once had a hearing, never once had a discussion other than an hour or so on the floor.” Cyberporn and Children: The Scope of the Problem, The State of the Technology, and the Need for Congressional Action, Hearing on S. 892 before the Senate Committee on the Judiciary, 104th Cong., 1st Sess., 7-8 (1995).

Although the Government and the dissent break § 223(d)(1) into two separate “patently offensive” and “display” provisions, we follow the convention of both parties below, as well as the District Court’s order and opinion, in describing § 223(d)(1) as one provision.

In full, § 223(e)(5) provides:

“(5) It is a defense to a prosecution under subsection (a)(1)(B) or (d) of this section, or under subsection (a)(2) of this section with respect to the use of a facility for an activity under subsection (a)(1)(B) of this section that a person—

“(A) has taken, in good faith, reasonable, effective, and appropriate actions under the circumstances to restrict or prevent access by minors to a communication specified in such subsections, which may involve any appropriate measures to restrict minors from such communications, including any method which is feasible under available technology; or

“(B) has restricted access to such communication by requiring use of a verified credit card, debit account, adult access code, or adult personal identification number.”

American Civil Liberties Union; Human Rights Watch; Electronic Privacy Information Center; Electronic Frontier Foundation; Journalism Education Association; Computer Professionals for Social Responsibility; National Writers Union; Clarinet Communications Corp.; Institute for Global Communications; Stop Prisoner Rape; AIDS Education Global Information System; Bibliobytes; Queer Resources Directory; Critical Path AIDS Project, Inc.; Wildcat Press, Inc.; Declan McCullagh dba Justice on Campus; Brock Meeks dba Cyberwire Dispatch; John Troyer dba The Safer Sex Page; Jonathan Wallace dba The Ethical Spectacle; and Planned Parenthood Federation of America, Inc.

American Library Association; America Online, Inc.; American Booksellers Association, Inc.; American Booksellers Foundation for Free Expression; American Society of Newspaper Editors; Apple Computer, Inc.; Association of American Publishers, Inc.; Association of Publishers, Editors and Writers; Citizens Internet Empowerment Coalition; Commercial Internet Exchange Association; CompuServe Incorporated; Families Against Internet Censorship; Freedom to Read Foundation, Inc.; Health Sciences Libraries Consortium; Hotwired Ventures LLC; Interactive Digital Software Association; Interactive Services Association; Magazine Publishers of America; Microsoft Corporation; The Microsoft Network, L. L. C.; National Press Photographers Association; Netcom On-Line Communication Services, Inc.; Newspaper Association of America; Opnet, Inc.; Prodigy Services Company; Society of Professional Journalists; and Wired Ventures, Ltd.

110 Stat. 142-148, note following 47 U. S. C. § 223 (1994 ed., Supp. II).

See also 929 F. Supp., at 877: “Four related characteristics of Internet communication have a transcendent importance to our shared holding that the CDA is unconstitutional on its face. We explain these characteristics in our Findings of fact above, and I only rehearse them briefly here. First,'the Internet presents very low barriers to entry. Second, these barriers to entry are identical for both speakers and listeners. Third, as a result of these low barriers, astoundingly diverse content is available on the Internet. Fourth, the Internet provides significant access to all who wish to speak in the medium, and even creates a relative parity among speakers.” According to Judge Dalzell, these characteristics and the rest of the District Court’s findings “lead to the conclusion that Congress may not regulate indecency on the Internet at all.” Ibid. Because appellees do not press this argument before this Court, we do not consider it. Ap-pellees also do not dispute that the Government generally has a compelling interest in protecting minors from “indecent” and “patently offensive” speech.

390 U. S., at 639. We quoted from Prince v. Massachusetts, 321 U. S. 158, 166 (1944): “It is cardinal with us that the custody, care and nurture of the child reside first in the parents, whose primary function and freedom include preparation for obligations the state can neither supply nor hinder.”

Given the likelihood that many e-mail transmissions from an adult to a minor are conversations between family members, it is therefore incorrect for the partial dissent to suggest that the provisions of the CDA, even in this narrow area, “are no different from the law we sustained in Ginsberg.” Post, at 892.

Cf. Pacifica Foundation v. FCC, 556 F. 2d 9, 36 (CADC 1977) (Levanthal, J., dissenting), rev’d, FCC v. Pacifica Foundation, 438 U. S. 726 (1978). When Pacifica was decided, given that radio stations were allowed to operate only pursuant to federal license, and that Congress had enacted legislation prohibiting licensees from broadcasting indecent speech, there was a risk that members of the radio audience might infer some sort of official or societal approval of whatever was heard over the radio, see 556 F. 2d, at 37, n. 18. No such risk attends messages received through the Internet, which is not supervised by any federal agency.

Juris. Statement 3 (citing 929 F. Supp., at 881 (finding 3)).

“Indecent” does not benefit from any textual embellishment at all. “Patently offensive” is qualified only to the extent that it involves “sexual or excretory activities or organs” taken “in context” and “measured by contemporary community standards.”

See Gozlon-Peretz v. United States, 498 U. S. 395, 404 (1991) (“[W]here Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally and purposely in the disparate inclusion and exclusion” (internal quotation marks omitted)).

The statute does not indicate whether the “patently offensive” and “indecent” determinations should be made with respect to minors or the population as a whole. The Government asserts that the appropriate standard is “what is suitable material for minors.” Reply Brief for Appellants 18, n. 13 (citing Ginsberg v. New York, 390 U. S. 629, 633 (1968)). But the Conferees expressly rejected amendments that would have imposed such a “harmful to minors” standard. See S. Conf. Rep. No. 104-230, p. 189 (1996) (S. Conf. Rep.), 142 Cong. Rec. H1145, H1165-H1166 (Feb. 1, 1996). The Conferees also rejected amendments that would have limited the proscribed materials to those lacking redeeming value. See ibid.

Even though the word “trunk,” standing alone, might refer to luggage, a swimming suit, the base of a tree, or the long nose of an animal, its meaning is clear when it is one prong of a three-part description of a species of gray animals.

413 U. S., at 30 (Determinations of “what appeals to the ‘prurient interest’ or is ‘patently offensive’ ... are essentially questions of fact, and our Nation is simply too big and too diverse for this Court to reasonably expect that such standards could be articulated for all 50 States in a single formulation, even assuming the prerequisite consensus exists”). The CDA, which implements the “contemporary community standards” language of Miller, thus conflicts with the Conferees’ own assertion that the CDA was intended “to establish a uniform national standard of content regulation.” S. Conf. Rep., at 191.

Accord, Butler v. Michigan, 352 U. S. 380, 383 (1957) (ban on sale to adults of books deemed harmful to children unconstitutional); Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 128 (1989) (ban on “dial-a-porn” messages unconstitutional); Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 73 (1983) (ban on mailing of unsolicited advertisement for contraceptives unconstitutional).

The lack of legislative attention to the statute at issue in Sable suggests another parallel with this case. Compare 492 U. S., at 129-130 (“[A]side from conclusory statements during the debates by proponents of *876 the bill, as well as similar assertions in hearings on a substantially identical bill the year before,. . . the congressional record presented to us contains no evidence as to how effective or ineffective the FCC’s most recent regulations were or might prove to be. . . . No Congressman or Senator purported to present a considered judgment with respect to how often or to what extent minors could or would circumvent the rules and have access to dial-a-porn messages” (footnote omitted)), with n. 24, supra.

The Government agrees that these provisions are applicable whenever “a sender transmits a message to more than one recipient, knowing that at least one of the specific persons receiving the message is a minor.” Opposition to Motion to Affirm and Reply to Juris. Statement 4-5, n. 1.

The Government asserts that “[t]here is nothing constitutionally suspect about requiring commercial Web site operators ... to shoulder the modest burdens associated with their use.” Brief for Appellants 35. As a matter of fact, however, there is no evidence that a “modest burden” would be effective.

Transmitting obscenity and child pornography, whether via the Internet or other means, is already illegal under federal law for both adults and juveniles. See 18 U. S. C. §§ 1464-1465 (criminalizing obscenity); §2251 (criminalizing child pornography). In fact, when Congress was considering the CDA, the Government expressed its view that the law was unnecessary because existing laws already authorized its ongoing efforts to prosecute obscenity, child pornography, and child solicitation. See 141 Cong. Rec. 16026 (1995) (letter from Kent Markus, Acting Assistant Attorney General, U. S. Department of Justice, to Sen. Leahy).

Citing Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U. S. 520 (1993), among other cases, appellees offer an additional reason why, in their view, the CDA fails strict scrutiny. Because so much sexually explicit content originates overseas, they argue, the CDA cannot be “effective.” Brief for Appellees American Library Association et al. 33-34. This argument raises difficult issues regarding the intended, as well as the permissible scope of, extraterritorial application of the CDA. We find it unnecessary to address those issues to dispose of this case.

For the full text of § 223(e)(5), see n. 26, supra.

Thus, ironically, this defense may significantly protect commercial purveyors of obscene postings while providing little (or no) benefit for transmitters of indecent messages that have significant social or artistic value.

929 F. Supp., at 855-856.

As this Court long ago explained: “It would certainly be dangerous if the legislature could set a net large enough to catch all possible offenders, and leave it to the courts to step inside and say who could be rightfully detained, and who should be set at large. This would, to some extent, substitute the judicial for the legislative department of the government.” United States v. Reese, 92 U. S. 214, 221 (1876). In part because of these *885 separation-of-powers concerns, we have held that a severability clause is “an aid merely; not an inexorable command.” Dorchy v. Kansas, 264 U. S. 286, 290 (1924).

See also Osborne v. Ohio, 495 U. S. 103, 121 (1990) (judicial rewriting of statutes would derogate Congress’ “incentive to draft a narrowly tailored law in the first place”).