concurring in part and dissenting in part.

I agree that § 10(a) is constitutional and that § 10(b) is unconstitutional, and I join Parts I, II, III, and V, and the judgment in part. I am not persuaded, however, that the asserted “important differences” between §§ 10(a) and 10(c), ante, at 760, are sufficient to justify striking down § 10(c). I find the features shared by § 10(a), which covers leased access channels, and § 10(c), which covers public access channels, to be more significant than the differences. For that reason, I would find that § 10(c) also withstands constitutional scrutiny.

Both §§ 10(a) and 10(c) serve an important governmental interest: the well-established compelling interest of protecting children from exposure to indecent material. See Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 126 (1989); Ginsberg v. New York, 390 U. S. 629, 639-640 (1968). Cable television, like broadcast television, is a medium that is uniquely accessible to children, see ante, at 744-745, and, of course, children have equally easy access to public access channels as to leased access channels. By permitting a cable operator to prevent transmission of patently offensive sex-related programming, §§ 10(a) and 10(c) further the interest of protecting children.

Furthermore, both provisions are permissive. Neither presents an outright ban on a category of speech, such as we struck down in Sable Communications of Cal., Inc. v. FCC, supra. Sections 10(a) and 10(c) leave to the cable operator the decision whether to broadcast indecent programming, and, therefore, are less restrictive than an absolute governmental ban. Certainly § 10(c) is not more restrictive than § 10(a) in this regard.

It is also significant that neither § 10(a) nor § 10(c) is more restrictive than the governmental speech restriction we upheld in FCC v. Pacifica Foundation, 438 U. S. 726 (1978). I agree with Justice Breyer that we should not yet under*780take fully to adapt our First Amendment doctrine to the new context we confront here. Because we refrain from doing so, the precedent established by Pacifica offers an important guide. Section 10(c), no less than § 10(a), is within the range of acceptability set by Pacifica. See ante, at 744-747.

The distinctions upon which the Court relies in deciding that § 10(c) must fall while § 10(a) survives are not, in my view, constitutionally significant. Much emphasis is' placed on the differences in the origins of leased access and public access channels. To be sure, the leased access channels covered by § 10(a) were a product of the Federal Government, while the public access channels at issue in § 10(c) arose as part of the cable franchises awarded by municipalities, see ante, at 761-762,. but I am not persuaded that the difference in the origin of the access channels is sufficient to justify upholding § 10(a) and striking down § 10(c). The interest in protecting children remains the same, whether on a leased access channel or a public access channel, and allowing the cable operator the option of prohibiting the transmission of indecent speech seems a constitutionally permissible means of addressing that interest. Nor is the fact that public access programming may be subject to supervisory systems in addition to the cable operator, see ante, at 761-763, sufficient in my mind to render § 10(c) so ill tailored to its goal as to be unconstitutional.

Given the compelling interest served by § 10(c), its permissive nature, and its fit within our precedent, I would hold § 10(c), like § 10(a), constitutional.

with whom Justice Ginsburg joins, concurring in part, concurring in the judgment in part, and dissenting in part.

The plurality opinion, insofar as it upholds § 10(a) of the 1992 Cable Act, is adrift. The opinion treats concepts such as public forum, broadcaster, and common carrier as mere labels rather than as categories with settled legal signifi-*781canee; it applies no standard, and by this omission loses sight of existing First Amendment doctrine. When confronted with a threat to free speech in the context of an emerging technology, we ought to have the discipline to analyze the case by reference to existing elaborations of constant First Amendment principles. This is the essence of the ease-by-case approach to ensuring protection of speech under the First Amendment, even in novel settings. Rather than undertake this task, however, the plurality just declares that, all things considered, § 10(a) seems fine. I think the implications of our past cases for these cases are clearer than the plurality suggests, and they require us to hold § 10(a) invalid. Though I join Part III of the opinion (there for the Court) striking down § 10(b) of the Act, and concur in the judgment that § 10(c) is unconstitutional, with respect I dissent from the remainder.

I

Two provisions of the 1992 Act, §§ 10(a) and (c), authorize the operator of a cable system to exclude certain programming from two different kinds of channels. Section 10(a) concerns leased access channels. These are channels the cable operator is required by federal law to make available to unaffiliated programmers without exercising any control over program content. The statute allows a cable operator to enforce a written and published policy of prohibiting on these channels any programming it “reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner as measured by contemporary community standards,” speech we can refer to as “indecent programming.”

Section 10(c) involves public, educational, and governmental access channels (or PEG access channels, as they are known). These are channels set aside for use by members of the public, governmental authorities, and local school systems. As interpreted by the Federal Communications Commission (FCC), § 10(c) requires the agency to make regu*782lations enabling cable operators to prohibit indecent programming on PEG access channels. See ante, at 734-736 (quoting statutory provisions in full and discussing interpretive regulations).*

Though the two provisions differ in significant respects, they have common flaws. In both instances, Congress singles out one sort of speech for vulnerability to private censorship in a context where content-based discrimination is not otherwise permitted. The plurality at least recognizes this as state action, ante, at 737, avoiding the mistake made by the Court of Appeals, Alliance for Community Media v. FCC, 56 F. 3d 105, 112-121 (CADC 1995). State action lies in the enactment of a statute altering legal relations between persons, including the selective withdrawal from one group of legal protections against private acts, regardless of whether the private acts are attributable to the State. Cf. Hunter v. Erickson, 393 U. S. 385, 389-390 (1969) (state action under the Fourteenth Amendment).

The plurality balks at taking the next step, however, which is to advise us what standard it applies to determine whether the state action conforms to the First Amendment. Sections 10(a) and (c) disadvantage nonobscene, indecent programming, a protected category of expression, Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 126 (1989), on the basis of its content. The Constitution in general does not tolerate content-based restriction of, or discrimination against, speech. R. A. V. v. St. Paul, 505 U. S. 377, 382 (1992) (“Content-based regulations are presumptively invalid”); Carey v. Brown, 447 U. S. 455, 461-463 (1980); Police Dept. of Chicago v. Mosley, 408 U. S. 92, 96 (1972). In the *783realm of speech and expression, the First Amendment envisions the citizen shaping the government, not the reverse; it removes “governmental restraints from the arena of public discussion, putting the decision as to what views shall be voiced largely into the hands of each of us, in the hope that use of such freedom will ultimately produce a more capable citizenry and more perfect polity.” Cohen v. California, 403 U. S. 15, 24 (1971). “[E]ach person should decide for himself or herself the ideas and beliefs deserving of expression, consideration, and adherence. Our political system and cultural life rest upon this ideal.” Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 641 (1994). We therefore have given “the most exacting scrutiny to regulations that suppress, disadvantage, or impose differential burdens upon speech because of its content.” Id., at 642.

Sections 10(a) and (c) are unusual. They do not require direct action against speech, but do authorize a cable operator to deny the use of its property to certain forms of speech. As a general matter, a private person may exclude certain speakers from his or her property without violating the First Amendment, Hudgens v. NLRB, 424 U. S. 507 (1976), and if §§ 10(a) and (c) were no more than affirmations of this principle they might be unremarkable. Access channels, however, are property of the cable operator, dedicated or otherwise reserved for programming of other speakers or the government. A public access channel is a public forum, and laws requiring leased access channels create common-carrier obligations. When the government identifies certain speech on the basis of its content as vulnerable to exclusion from a common carrier or public forum, strict scrutiny applies. These laws cannot survive this exacting review. However compelling Congress’ interest in shielding children from indecent programming, the provisions in these cases are not drawn with enough care to withstand scrutiny under our precedents.

*784r — H 1 — I

Before engaging the complexities of cable access channels and explaining my reasons for thinking all of § 10 unconstitutional, I start with the most disturbing aspect of the plurality opinion: its evasion of any clear legal standard in deciding these cases. See ante, at 741 (disavowing need to “declare which, among the many applications of the general approach that this Court has developed over the years, we are applying here”).

The plurality begins its flight from standards with a number of assertions nobody disputes. I agree, of course, that it would be unwise “to declare a rigid single standard, good for now and for all future media and purposes,” ante, at 742. I do think it necessary, however, to decide what standard applies to discrimination against indecent programming on cable access channels in the present state of the industry. We owe at least that much to public and leased access programmers whose speech is put at risk nationwide by these laws.

In a similar vein, we are admonished, these cases are complicated, not simple; the importance of contextual review, we are told, cannot be evaded by recourse to simple analogies. Ante, at 739-743, 748. All this is true, but use of a standard does not foreclose consideration of context. Indeed, if strict scrutiny is an instance of “judicial formulas so rigid that they become a straitjacket that disables government from responding to serious problems,” ante, at 741, this is a grave indictment of our First Amendment jurisprudence, which relies on strict scrutiny in a number of settings where context is important. I have expressed misgivings about judicial balancing under the First Amendment, see Burson v. Freeman, 504 U. S. 191, 211-212 (1992) (concurring opinion); Simon & Schuster, Inc. v. Members of N. Y. State Crime Victims Bd., 502 U. S. 105, 124-125 (1991) (opinion concurring in judgment), but strict scrutiny at least confines the balancing process in a manner protective of speech; it does not disable government from addressing serious problems, *785but does ensure that the solutions do not sacrifice speech to a greater extent than necessary.

The plurality claims its resistance to standards is in keeping with our case law, where we have shown a willingness to be flexible in confronting novel First Amendment problems. The cases it cites, ante, at 740-741, however, demonstrate the opposite of what the plurality supposes: In each, we developed specialized or more or less stringent standards when certain contexts demanded them; we did not avoid the use of standards altogether. Indeed, the creation of standards and adherence to them, even when it means affording protection to speech unpopular or distasteful, is the central achievement of our First Amendment jurisprudence. Standards •are the means by which we state in advance how to test a law’s validity, rather than letting the height of the bar be determined by the apparent exigencies of the day. They also provide notice and fair warning to those who must predict how the courts will respond to attempts to suppress their speech. Yet formulations like strict scrutiny, used in a number of constitutional settings to ensure that the inequities of the moment are subordinated to commitments made for the long run, see Simon & Schuster, supra, at 115-116; Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983), mean little if they can be watered down whenever they seem too strong. They mean still less if they can be ignored altogether when considering a case not on all fours with what we have seen before.

The plurality seems distracted by the many changes in technology and competition in the cable industry. See ante, at 741-742; ante, at 776-777 (Souter, J., concurring). The laws challenged here, however, do not retool the structure of the cable industry or (with the exception of § 10(b)) involve intricate technologies. The straightforward issue here is whether the Government can deprive certain speakers, on the basis of the content of their speech, of protections af*786forded all others. There is no reason to discard our existing First Amendment jurisprudence in answering this question.

While it protests against standards, the plurality does seem to favor one formulation of the question in these cases: namely, whether the Act “properly addresses an extremely important problem, without imposing, in light of the relevant interests, an unnecessarily great restriction on speech.” Ante, at 743. (Though the plurality frowns on any effort to settle on a form of words, it likes this formulation well enough to repeat it; see ante, at 741.) This description of the question accomplishes little, save to clutter our First Amendment case law by adding an untested rule with an uncertain relationship to the others we use to evaluate laws restricting speech. The plurality cannot bring itself to apply strict scrutiny, yet realizes it cannot decide these cases without uttering some sort of standard; so it has settled for synonyms. “[C]lose judicial scrutiny,” ibid., is substituted for strict scrutiny, and “extremely important problem,” ante, at 743, or “extraordinary proble[m],” ante, at 741, is substituted for “compelling interest.” The admonition that the restriction not be unnecessarily great in light of the interest it serves, ante, at 743, is substituted for the usual narrow tailoring requirements. All we know about the substitutes is that they are inferior to their antecedents. We are told the Act must be “appropriately tailored,” ante, at 741, “sufficiently tailored,” ante, at 743, or “carefully and appropriately addressed,” ante, at 748, to the problems at hand — anything, evidently, except narrowly tailored.

These restatements have unfortunate consequences. The first is to make principles intended to protect speech easy to manipulate. The words end up being a legalistic cover for an ad hoc balancing of interests; in this respect the plurality succeeds after all in avoiding the use of a standard. Second, the plurality’s exercise in pushing around synonyms for the words of our usual standards will sow confusion in the courts bound by our precedents. Those courts, and lawyers in the *787communications field, now will have to discern what difference there is between the formulation the plurality applies today and our usual strict scrutiny. I can offer little guidance, except to note the unprotective outcome the plurality reaches here. This is why comparisons and analogies to other areas of our First Amendment case law become a responsibility, rather than the luxury the plurality considers them to be. The comparisons provide discipline to the Court and guidance for others, and give clear content to our standards — all the things I find missing in the plurality’s opinion. The novelty and complexity of these cases is a reason to look for help from other areas of our First Amendment jurisprudence, not a license to wander into uncharted areas of the law with no compass other than our own opinions about good policy.

Another troubling aspect of the plurality’s approach is its suggestion that Congress has more leeway than usual to enact restrictions on speech where emerging technologies are concerned, because we are unsure what standard should be used to assess them. Justice Souter recommends to the Court the precept, “‘First, do no harm,’” ante, at 778. The question, though, is whether the harm is in sustaining the law or striking it down. If the plurality is concerned about technology’s direction, it ought to begin by allowing speech, not suppressing it. We have before us an urgent claim for relief against content-based discrimination, not a dry run.

I turn now to the issues presented, and explain why strict scrutiny is warranted.

III

A

Cable operators deliver programming from four sources: retransmission of broadcast stations; programming purchased from professional vendors (including national services like ESPN and Nickelodeon) and delivered by satellite; pro*788grams created by the cable operator itself; and access channels (PEG and leased), the two kinds of programming at issue here. See Mueller, Note, Controversial Programming on Cable Television’s Public Access Channels: The Limits of Governmental Response, 38 DePaul L. Rev. 1051, 1056-1057 (1989) (hereinafter Mueller). See also Turner Broadcasting, 512 U. S., at 628-629.

PEG access channels grew out of local initiatives in the late 1960’s and early 1970’s, before the Federal Government began regulating cable television. Mueller 1061. Local franchising was the first form of cable regulation, arising from the need of localities to control access to public rights-of-way and easements and to minimize disruption to traffic and other public activity from the laying of cable lines. See D. Brenner, M. Price, & M. Meyerson, Cable Television and Other Nonbroadcast Video § 3.01[3] (1996) (hereinafter Brenner); Turner Broadcasting, supra, at 628 (“[T]he cable medium may depend for its very existence upon express permission from local governing authorities”). A local government would set up a franchise authority to oversee the cable system and to negotiate a franchise agreement specifying the cable operator’s rights and obligations. See Brenner § 3.01; § 3.01[4] (discussing States where local franchising has now been displaced by state regulation). Cf. 47 U. S. C. § 522(10) (defining franchise authority). A franchise, now mandatory under federal law except for systems operating without them prior to 1984, § 541(b), is an authorization, akin to a license, by a franchise authority permitting the construction or operation of a cable system. § 522(8). From the early 1970’s onward, franchise authorities began requiring operators to set aside access channels as a condition of the franchise. See Mueller 1061-1062; D. Agosta, C. Rogoff, & A. Norman, The Participate Report: A Case Study of Public Access Cable Television in New York State 24 (1990) (hereinafter Agosta), attached as Exhibit K to Joint Comments for the Alliance for *789Community Media et al., filed with the FCC under MM Docket No. 92-258 (hereinafter FCC Record).

The FCC entered the arena in 1972, requiring the cable companies servicing the country’s largest television markets to set aside four access channels (one each for public, educational, governmental, and leased programming) by a date certain, and to add channel capacity if necessary to meet the requirement. Cable Television Report and Order, 36 F. C. C. 2d 141, 189-198 (1972). See also In re Amendment of Part 76 of the Commission’s Rules and Regulations Concerning the Cable Television Channel Capacity and Access Channel Requirements of Section 76.251, 59 F. C. C. 2d 294, 303, 321 (1976) (modifying the 1972 rules). We struck down the access rules as beyond the FCC’s authority under the Communications Act of 1934. FCC v. Midwest Video Corp., 440 U. S. 689, 708-709 (1979).

When Congress turned its attention to PEG access channels in 1984, it recognized that “reasonable third-party access to cable systems will mean a wide diversity of information sources for the public — the fundamental goal of the First Amendment — without the need to regulate the content of programming provided over cable.” H. R. Rep. No. 98-934, p. 30 (1984). It declined, however, to set new federal mandates or authorize the FCC to do so. Since “[a]lmost all recent franchise agreements provide for access by local governments, schools, and non-profit and community groups” over some channels, the 1984 Act instead “continued] the policy of allowing cities to specify in cable franchises that channel capacity and other facilities be devoted to such use.” Ibid.

Section 611 of the Communications Act of 1934, added by the Cable Communications Policy Act of 1984 (1984 Act), authorized local franchise authorities to require cable operators to set aside channel capacity for PEG access when seeking new franchises or renewal of old ones. 47 U. S. C. § 531(b). *790Franchise authorities may enforce franchise agreements, § 531(c), but they lack the power to impose requirements beyond those authorized by federal law, § 531(a). But cf. § 557(a) (grandfathering as valid all pre-1984 franchise agreements for the remainder of their term). Federal law also allows a franchise authority to “require adequate assurance that the cable operator will provide adequate public, educational, and governmental access channel capacity, facilities, or financial support.” § 541(a)(4)(B). Prior to the passage of § 10(c) of the 1992 Act, the cable operator, save for implementing provisions of its franchise agreement limiting obscene or otherwise constitutionally unprotected cable programming, § 544(d), was forbidden any editorial control over PEG access channels. 47 U. S. C. § 531(e) (1988 ed.).

Congress has not, in the 1984 Act or since, defined what public, educational, or governmental access means or placed substantive limits on the types of programming on those channels. Those tasks are left to franchise agreements, so long as the channels comport in some sense with the industry practice to which Congress referred in the statute.

My principal concern is. with public access channels (the P of PEG). These are the channels open to programming by members of the public. Petitioners here include public access programmers and viewers who watch their shows. By contrast, educational and governmental access channels (the E and G of PEG) serve other speakers. Under many franchises, educational channels are controlled by local school systems, which use them to provide school information and educational programs. Governmental access channels are committed by the cable franchise to the local municipal government, which uses them to distribute information to constituents on public affairs. Mueller 1065-1066. No local governmental entity or school system has petitioned for relief in these cases, and none of the petitioners who are viewers has asserted an interest in viewing educational or governmental programming or briefed the relevant issues.

*791B

The public access channels established by franchise agreements tend to have certain traits. They are available at low or no cost to members of the public, often on a first-come, first-served basis. Brenner § 6.04[3][a]-[b], at 6-38. The programmer on one of these channels most often has complete control over, as well as liability for, the content of its show. Ibid.; Mueller 1064. The entity managing the technical aspects of public access, such as scheduling and transmission, is not always the cable operator; it may be the local government or a third party that runs the access centers, which are facilities made available for the public to produce programs and transmit them on the access channels. Brenner § 6.04[7], at 6-48.

Public access channels meet the definition of a public forum. We have recognized two kinds of public fora. The first and most familiar are traditional public fora, like streets, sidewalks, and parks, which by custom have long been open for public assembly and discourse. Perry, 460 U. S., at 45; Hague v. Committee for Industrial Organization, 307 U. S. 496, 515 (1939) (opinion of Roberts, J.). “The second category of public property is the designated public forum, whether of a limited or unlimited character — property that the State has opened for expressive activity by part or all of the public.” International Soc. for Krishna Consciousness, Inc. v. Lee, 505 U. S. 672, 678 (1992).

Public access channels fall in the second category. Required by the franchise authority as a condition of the franchise and open to all comers, they are a designated public forum of unlimited character. The House Report for the 1984 Act is consistent with this view. It characterized public access channels as “the video equivalent of the speaker’s soap box or the electronic parallel to the printed leaflet. They provide groups and individuals who generally have not had access to the electronic media with the opportunity to become sources of information in the electronic mar-

*792ketplace of ideas.” H. R. Rep. No. 98-934, supra, at 30. Public fora do not have to be physical gathering places, Rosenberger v. Rector and Visitors of Univ. of Va., 515 U. S. 819, 830 (1995), nor are they limited to property owned by the government, Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 800 (1985). Indeed, in the majority of jurisdictions, title to some of the most traditional of public fora, streets and sidewalks, remains- in private hands. 10A E. McQuillin, Law of Municipal Corporations § 30.32 (3d ed. 1990); Hague, supra, at 515 (“Wherever the title of streets and parks may rest, they have immemorially been held in trust for the use of the public and, time out of mind, have been used for purposes of assembly, communicating thoughts between citizens, and discussing public questions”). Public access channels are analogous; they are public fora even though they operate over property to which the cable operator holds title.

It is important to understand that public access channels are public fora created by local or state governments in the cable franchise. Section § 10(c) does not, as the Court of Appeals thought, just return rightful First Amendment discretion to the cable operator, see Alliance for Community Media, 56 F. 3d, at 114. Cable operators have First Amendment rights, of course; restrictions on entry into the cable business may be challenged under the First Amendment, Los Angeles v. Preferred Communications, Inc., 476 U. S. 488, 494 (1986), and a cable operator’s activities in originating programs or exercising editorial discretion over programs others' provide on its system also are protected, Turner Broadcasting, 512 U. S., at 636. But cf. id., at 656 (distinguishing discretion of cable operators from that of newspaper editors). Yet the editorial discretion of a cable operator is a function of the cable franchise it receives from local government. The operator’s right to exercise any editorial discretion over cable service disappears if its franchise is terminated. See 47 U. S. C. § 541(b) (cable service may not *793be offered without a franchise); §546 (prescribing procedures and standards for renewal). Cf. Brenner §3.07[9][a] (franchise terms of 15 years are the norm); §3.07[15] (typical franchise agreements recognize the absolute right of the franchiser to refuse renewal at expiration of term). If the franchise is transferred to another, so is the right of editorial discretion. The cable operator may own the cables transmitting the signal, but it is the franchise — the agreement between the cable operator and the local government — that allocates some channels to the full discretion of the cable operator while reserving others for public access.

In providing public access channels under their franchise agreements, cable operators therefore are not exercising their own First Amendment rights. They serve as conduits for the speech of others. Cf. PruneYard Shopping Center v. Robins, 447 U. S. 74, 87 (1980). Section 10(c) thus restores no power of editorial discretion over public access channels that the cable operator once had; the discretion never existed. It vests the cable operator with a power under federal law, defined by reference to the content of speech, to override the franchise agreement and undercut the public forum the agreement creates. By enacting a law in 1992 excluding indecent programming from protection but retaining the prohibition on cable operators’ editorial control over all other protected speech, the Federal Government at the same time ratified the public-forum character of public access channels but discriminated against certain speech based on its content.

The plurality refuses to analyze public access channels as public fora because it is reluctant to decide “the extent to which private property can be designated a public forum,” ante, at 742. We need not decide here any broad issue whether private property can be declared a public forum by simple governmental decree. That is not what happens in the creation of public access channels. Rather, in return for granting cable operators easements to use public rights-of-*794way for their cable lines, local governments have bargained for a right to use cable lines for public access channels. Justice Thomas resists public-forum analysis because he sees no evidence of a “formal easement.” Post, at 828. Under general principles of property law, no particular formalities are necessary to create an easement. Easements may be created by contract. 2 G. Thompson, Commentaries on the Modern Law of Real Property §§331-332 (1980); 3 H. Tiffany, The Law of Real Property § 776 (3d ed. 1939). A franchise agreement is a contract, and in those agreements the cable operator surrenders his power to exclude certain programmers from use of his property for specific purposes. A state court confronted with the issue would likely hold the franchise agreement to create a right of access equivalent to an easement in land. So one can even view these cases as a local government’s dedication of its own property interest to speech by members of the public. In any event, it seems to me clear that when a local government contracts to use private property for public expressive activity, it creates a public forum.

Treating access channels as public fora does not just place a label on them, as the plurality suggests, see ante, at 750. It defines the First Amendment rights of speakers seeking to use the channels. When property has been dedicated to public expressive activities, by tradition or government designation, access is protected by the First Amendment. Regulations of speech content in a designated public forum, whether of limited or unlimited character, are “subject to the highest scrutiny” and “survive only if they are narrowly drawn to achieve a compelling state interest.” Lee, 505 U. S., at 678. Unless there are reasons for applying a lesser standard, § 10(c) must satisfy this stringent review.

C

Leased access channels, as distinct from public access channels, are those the cable operator must set aside for un*795affiliated programmers who pay to transmit shows of their own without the cable operator’s creative assistance or editorial approval. In my view, strict scrutiny also applies to § 10(a)’s authorization to cable operators to exclude indecent programming from these channels.

Congress created leased access channels in the 1984 Act. Section 612 of the Act, as amended, requires a cable system with more than 36 channels to set aside a certain percentage of its channels (up to 15%, depending on the size of the system) “for commercial use by persons unaffiliated with the operator.” 47 U. S. C. § 532(b)(1). Commercial use means “provision of video programming, whether or not for profit.” § 532(b)(5). When an unaffiliated programmer seeks access, the cable operator shall set “the price, terms, and conditions of such use which are at least sufficient to assure that such use will not adversely affect the operation, financial condition, or market development of the cable system,” § 532(c)(1). Cf. 47 CFR §76.971 (1995) (rules governing terms and conditions of leased access). The price may not exceed the maximum charged any unaffiliated programmer in the same program category for the use of nonaccess channels. § 76.970. Aggrieved programmers have recourse to federal district court and the FCC (if there are repeated violations) to compel access on appropriate terms. 47 U.S. C. §§ 532(d), (e).

Before 1992, cable operators were forbidden editorial control over any video programming on leased access channels, and could not consider the content of the programming except to set the price of access, 47 U. S. C. § 532(c)(2) (1988 ed.). But cf. 47 U. S. C. § 532(h) (prohibiting programs that are obscene or otherwise unprotected under the Constitution on leased access channels). Section 10(a) of the 1992 Act modifies the no-discretion rule by allowing cable operators to reject, pursuant to a written and published policy, programs they reasonably believe to be indecent. § 532(h). Under § 10(b) of the Act, any indecent programming must be *796segregated onto one channel and blocked unless the subscriber requests that the channel be provided to him. § 532(j); 47 CFR §76.701 (1995).

Two distinctions between public and leased access channels are important. First, whereas public access channels are required by state and local franchise authorities (subject to certain federal limitations), leased access channels are created by federal law. Second, whereas cable operators never have had editorial discretion over public access channels under their franchise agreements, the leased access provisions of the 1984 Act take away channels the operator once controlled. Cf. Midwest Video, 440 U. S., at 708, n. 17 (federal mandates “compelling cable operators indiscriminately to accept access programming will interfere with their determinations regarding the total service offering to be extended to subscribers”). In this sense, § 10(a) now gives back to the operator some of the discretion it had before Congress imposed leased access requirements in the first place.

The constitutionality under Turner Broadcasting, 512 U. S., at 665-668, of requiring a cable operator to set aside leased access channels is not before us. For purposes of these cases, we should treat the cable operator’s rights in these channels as extinguished, and address the issue these petitioners present: namely, whether the Government can discriminate on the basis of content in affording protection to certain programmers. I cannot agree with Justice Thomas, post, at 821-822, that the cable operator’s rights inform this analysis.

Laws requiring cable operators to provide leased access are the practical equivalent of making them common carriers, analogous in this respect to telephone companies: They are obliged to provide a conduit for the speech of others. The plurality resists any classification of leased access channels (as created in the 1984 Act) as a common-carrier provision, ante, at 739-740, although we described in just those *797terms the access (including leased'access) rules promulgated by the FCC in 1976:

“The access rules plainly impose common-carrier obligations on cable operators. Under the rules, cable systems are required to hold out dedicated channels on a first-come, nondiscriminatory basis. Operators are prohibited from determining or influencing the content of access programming. And the rules delimit what operators may charge for access and use of equipment.”Midwest Video, 440 U. S., at 701-702 (citations and footnotes omitted).

Indeed, we struck down the FCC’s rules as beyond the agency’s statutory authority at the time precisely because they made cable operators common carriers. Id,, at 702-709. The FCC characterizes §612 as a form of common-carrier requirement, App. to Pet. for Cert. 139a-140a, as does the Government, Brief for Federal Respondents 23.

Section 10(a) authorizes cable operators to ban indecent programming on leased access channels. We have held that a law precluding a common carrier from transmitting protected speech is subject to strict scrutiny, Sable Communications, 492 U. S., at 131 (striking down ban on indecent telephonic communications), but we have not had occasion to consider the standard for reviewing a law, such as § 10(a), permitting a carrier in its discretion to exclude specified speech.

Laws removing common-carriage protection from a single form of speech based on its content should be reviewed under the same standard as content-based restrictions on speech in a public forum. Making a cable operator a common carrier does not create a public forum in the sense of taking property from private control and dedicating it to public use; rather, regulations of a common carrier dictate the manner in which private control is exercised. A common-carriage

*798mandate, nonetheless, serves the same function as a public forum. It ensures open, nondiscriminatory access to the means of communication. This purpose is evident in the statute itself and in the committee findings supporting it. Congress described the leased access requirements as intended “to promote competition in the delivery of diverse sources- of video programming and to assure that the widest possible diversity of information sources are made available to the public from cable systems in a manner consistent with growth and development of cable systems.” 47 U. S. C. § 532(a). The House Committee reporting the 1984 cable bill acknowledged that, in general, market demand would prompt cable operators to provide diverse programming. It recognized, though, the incentives cable operators might have to exclude “programming which represents a social or political viewpoint that a cable operator does not wish to disseminate, or ... competes with a program service already being provided by that cable system.” H. R. Rep. No. 98-934, at 48. In its view, the leased access provisions were narrowly drawn structural regulations of private industry, cf. Associated Press v. United States, 326 U. S. 1 (1945), to enhance the free flow and diversity of information available to the public without governmental intrusion into decisions about program content. H. R. Rep. No. 98-934, supra, at 32-35. The functional equivalence of designating a public forum and mandating common carriage suggests the same scrutiny should be applied to attempts in either setting to impose content discrimination by law. Under our precedents, the scrutiny is strict.

“The Constitution forbids a State to enforce certain exclusions from a forum generally open to the public even if it was not required to create the forum in the first place. Widmar v. Vincent, 454 U. S. 263 (1981) (university meeting facilities); City of Madison Joint School District v. Wisconsin Employment Relations Comm’n, 429 U. S. 167 (1976) (school board meeting); Southeast*799ern Promotions, Ltd. v. Conrad, 420 U. S. 546 (1975) (municipal theater). Although a State is not required to indefinitely retain the open character of the facility, as long as it does so it is bound by the same standards as apply in a traditional public forum. Reasonable time, place, and manner regulations are permissible, and a content-based prohibition must be narrowly drawn to effectuate a compelling state interest.” Perry, 460 U. S., at 45-46 (footnote omitted).

In Police Dept. of Chicago v. Mosley, 408 U. S. 92 (1972), we made clear that selective exclusions from a public forum were unconstitutional. Invoking the First and Fourteenth Amendments to strike down a city ordinance allowing only labor picketing on any public way near schools, we held the “government may not grant the use of a forum to people whose views it finds acceptable, but deny use to those wishing to express less favored or more controversial views.” Id., at 96.

“Once a forum is opened up to assembly or speaking by some groups, government may not prohibit others from assembling or speaking on the basis of what they intend to say. Selective exclusions from a public forum may not be based on content alone, and may not be justified by reference to content alone.” Ibid.

Since the same standard applies to exclusions from limited or unlimited designated public fora as from traditional forums, Lee, 505 U. S., at 678, there is no reason the kind of selective exclusion we condemned in Mosley should be tolerated here.

The plurality acknowledges content-based exclusions from the right to use a common carrier could violate the First Amendment. It tells us, however, that it is wary of analogies to doctrines developed elsewhere, and so does not address this issue. Ante, at 749. This newfound aversion to analogical reasoning strikes at a process basic to legal analy*800sis. See E. Levi, An Introduction to Legal Reasoning 1-2 (1949). I am not suggesting the plurality should look far afield to other areas of law; these are settled First Amendment doctrines dealing with state action depriving certain speakers of protections afforded to all others.

In all events, the plurality’s unwillingness to consider our public-forum precedents does not relieve it of the burden of explaining why strict scrutiny should not apply. Except in instances involving well-settled categories of proscribable speech, see R. A. V., 505 U. S., at 382-390, strict scrutiny is the baseline rule for reviewing any content-based discrimination against speech. The purpose of forum analysis is to determine whether, because of the property or medium where speech takes place, there should be any dispensation from this rule. See Consolidated Edison Co. of N. Y v. Public Service Comm’n of N. Y, 447 U. S. 530, 538-539 (1980). In the context of government property, we have recognized an exception “[w]here the government is acting as a proprietor, managing its internal operations, rather than acting as lawmaker with the power to regulate or license,” and in those circumstances, we have said, regulations of speech need only be reasonable and viewpoint neutral. Lee, supra, at 678-679. Here, of course, the Government has not dedicated the cable operator’s property for leased access to serve some proprietary function of its own; it has done so to provide a forum for a vital class of programmers who otherwise would be excluded from cable television.

The question remains whether a dispensation from strict scrutiny might be appropriate because § 10(a) restores in part an editorial discretion once exercised by the cable operator over speech occurring on its property. This is where public-forum doctrine gives guidance. Common-carrier requirements of leased access are little different in function from designated public fora, and no different standard of review should apply. It is not that the functional equivalence of leased access channels to designated public fora *801compels strict scrutiny; rather, it simply militates against recognizing an exception to the normal rule.

Perhaps, as the plurality suggests, ante, at 749-750, § 10(a) should be treated as a limitation on a forum rather than an exclusion from it. This would not change the analysis, however. If Government has a freer hand to draw content-based distinctions in limiting a forum than in excluding someone from it, the First Amendment would be a dead letter in designated public fora; every exclusion could be recast as a limitation. See Post, Between Governance and Management: the History and Theory of the Public Forum, 34 UCLA L. Rev. 1713, 1753 (1987). We have allowed content-based limitations of public fora, but only when necessary to serve specific institutional ends. See Perry, 460 U. S., at 48 (school mailboxes, if considered designated public fora, could be limited to mailings from “organizations that engage in activities of interest and educational relevance to students”); Widmar v. Vincent, 454 U. S. 263, 267-268, n. 5 (1981) (recognizing a public university could limit the use of its facilities by reasonable regulations compatible with its mission of education); Madison Joint School Dist. No. 8 v. Wisconsin Employment Relations Comm’n, 429 U. S. 167, 175, n. 8 (1976) (in assessing a teacher’s right to speak at a school board meeting, considering it obvious that “public bodies may confine their meetings to specified subject matter”). The power to limit or redefine fora for a specific legitimate purpose, see Rosenberger, 515 U. S., at 829-830, does not allow the government to exclude certain speech or speakers from them for any reason at all.

Madison Joint School Dist., supra, illustrates the point. The Wisconsin Employment Relations Commission had ordered a school board to prohibit school employees other than union representatives from speaking at its meetings on matters subject to collective bargaining between the board and the union. Id., at 173. While recognizing the power of a State to limit school board meetings to certain subject mat*802ter, we held it could not confine the forum “to one category of interested individuals.” Id., at 175. The exclusion would skew the debate and deprive decisionmakers of the benefit of other voices. Id., at 175-176.

It is no answer to say Congress does not have to create access channels at all, so it may limit access as it pleases. Whether or not a government has any obligation to make railroads common carriers, under the Equal Protection Clause it could not define common carriage in ways that discriminate against suspect classes. See Bailey v. Patterson, 369 U. S. 31, 33 (1962) (per curiam) (States may not require railroads to segregate the races). For the same reason, even if Congress has no obligation to impose common-carriage rules on cable operators or retain them forever, it is not at liberty to exclude certain forms of speech from their protection on the suspect basis of content. See Perry, supra, at 45-46.

I do not foreclose the possibility that the Government could create a forum limited to certain topics or to serving the special needs of certain speakers or audiences without its actions being subject to strict scrutiny. This possibility seems to trouble the plurality, which wonders if a local government must “show a compelling state interest if it builds a band shell in the park and dedicates it solely to classical music (but not to jazz).” Ante, at 750. This is not the correct analogy. These cases are more akin to the Government’s creation of a band shell in which all types of music might be performed except for rap music. The provisions here are content-based discriminations in the strong sense of suppressing a certain form of expression that the Government dislikes or otherwise wishes to exclude on account of its effects, and there is no justification for anything but strict scrutiny here.

Giving government free rein to exclude speech it dislikes by delimiting public fora (or common-carriage provisions) would have pernicious effects in the modern age. Minds are *803not changed in streets and parks as they once were. To an increasing degree, the more significant interchanges of ideas and shaping of public consciousness occur in mass and electronic media. Cf. United States v. Kokinda, 497 U. S. 720, 737 (1990) (Kennedy, J., concurring in judgment). The extent of public entitlement to participate in those means of communication may be changed as technologies change; and in expanding those entitlements the Government has no greater right to discriminate on suspect grounds than it does when it effects a ban on speech against the backdrop of the entitlements to which we have been more accustomed. It contravenes the First Amendment to give Government a general license to single out some categories of speech for lesser protection so long as it stops short of viewpoint discrimination.

D

The Government advances a different argument for not applying strict scrutiny in these cases. The nature of access channels to one side, it argues the nature of the speech in question — indecent broadcast (or cablecast) — is subject to the lower standard of review it contends was applied in FCC v. Pacifica Foundation, 438 U. S. 726, 748 (1978) (upholding an FCC order declaring the radio broadcast of indecent speech during daytime hours to be sanctionable).

Pacifica did not purport, however, to apply a special standard for indecent broadcasting. Emphasizing the narrowness of its holding, the Court in Pacifica conducted a context-specific analysis of the FCC’s restriction on indecent programming during daytime hours. See id., at 750. See also Sable Communications, 492 U. S., at 127-128 (underscoring the narrowness of Pacifica). It relied on the general rule that “broadcasting . . . has received the most limited First Amendment protection.” 438 U. S., at 748. We already have rejected the application of this lower broadcast standard of review to infringements on the liberties of cable operators, even though they control an important communica*804tions medium. Turner Broadcasting, 512 U. S., at 637-641. There is even less cause for a lower standard here.

Pacifica did identify two important considerations relevant to the broadcast of objectionable material. First, indecent broadcasting “confronts the citizen, not only in public, but also in the privacy of the home, where the individual's right to be left alone plainly outweighs the First Amendment rights of an intruder.” 438 U. S., at 748. Second, “broadcasting is uniquely accessible to children, even those too young to read.” Id., at 749. Pacifica teaches that access channels, even if analogous to ordinary public fora from the standpoint of the programmer, must also be considered from the standpoint of the viewer. An access channel is not a forum confined to a discrete public space; it can bring indecent expression into the home of every cable subscriber, where children spend astounding amounts of time watching television, cf. ante, at 744-745 (citing studies). Though in Cohen we explained that people in public areas may have to avert their eyes from messages that offend them, 403 U. S., at 21, we further acknowledged that “government may properly act in many situations to prohibit intrusion into the privacy of the home of unwelcome views and ideas which cannot be totally banned from the public dialogue,” ibid. See Hess v. Indiana, 414 U. S. 105, 108 (1973) (per curiam); Rowan v. Post Office Dept., 397 U. S. 728, 736-738 (1970). This is more true when the interests of children are at stake. See id., at 738 (“[T]he householder [should not] have to risk that offensive material come into the hands of his children before it can be stopped”).

These concerns are weighty and will be relevant to whether the law passes strict scrutiny. They do not justify, however, a blanket rule of lesser protection for indecent speech. Other than the few categories of expression that can be proscribed, see R. A. V., 505 U. S., at 382-390, we have been reluctant to mark off new categories of speech for diminished constitutional protection. Our hesitancy reflects *805skepticism about the possibility of courts drawing principled distinctions to use in judging governmental restrictions on speech and ideas, Cohen, 403 U. S., at 25, a concern heightened here by the inextricability of indecency from expression. “[W]e cannot indulge the facile assumption that one can forbid particular words without also running a substantial risk of suppressing ideas in the process.” Id., at 26. The same is true of forbidding programs indecent in some respect. In artistic or political settings, indecency may have strong communicative content, protesting conventional norms or giving an edge to a work by conveying “otherwise inexpressible emotions.” Ibid. In scientific programs, the more graphic the depiction (even if to the point of offensiveness), the more accurate and comprehensive the portrayal of the truth may be. Indecency often is inseparable from the ideas and viewpoints conveyed, or separable only with loss of truth or expressive power. Under our traditional First Amendment jurisprudence, factors perhaps justifying some restriction on indecent cable programming may all be taken into account without derogating this category of protected speech as marginal.

IV

At a minimum, the proper standard for reviewing §§ 10(a) and (c) is strict scrutiny. The plurality gives no reason why it should be otherwise. I would hold these enactments unconstitutional because they are not narrowly tailored to serve a compelling interest.

The Government has no compelling interest in restoring a cable operator’s First Amendment right of editorial discretion. As to § 10(c), Congress has no interest at all, since under most franchises operators had no rights of editorial discretion over PEG access channels in the first place. As to § 10(a), any governmental interest in restoring operator discretion over indecent programming on leased access channels is too minimal to justify the law. First, the transmission of indecent programming over leased access channels *806is not forced speech of the operator. Turner Broadcasting, supra, at 655-656; PruneYard, 447 U. S., at 87. Second, the discretion conferred by the law is slight. The operator is not authorized to place programs of its own liking on the leased access channels, nor to remove other speech (racist or violent, for example) that might be offensive to it or to viewers. The operator is just given a veto over the one kind of lawful speech Congress disdains.

Congress does have, however, a compelling interest in protecting children from indecent speech. Sable Communications, 492 U. S., at 126; Ginsberg v. New York, 390 U. S. 629, 639-640 (1968). See also Pacifica, 438 U. S., at 749-750 (same). So long as society gives proper respect to parental choices, it may, under an appropriate standard, intervene to spare children exposure to material not suitable for minors. This interest is substantial enough to justify some regulation of indecent speech even under, I will assume, the indecency standard used here.

Sections 10(a) and (c) nonetheless are not narrowly tailored to protect children from indecent programs on access channels. First, to the extent some operators may allow indecent programming, children in localities those operators serve will be left unprotected. Partial service of a compelling interest is not narrow tailoring. FCC v. League of Women Voters of Cal., 468 U. S. 364, 396 (1984) (asserted interest in keeping noncommercial stations free from controversial or partisan opinions not served by ban on station editorials, if such opinions could be aired through other programming); Florida Star v. B. J. F., 491 U. S. 524, 540-541 (1989) (selective ban on publication of rape victim’s name in some media but not others not narrowly tailored). Cf. Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 73 (1983) (restriction that “provides only the most limited incremental support for the interest asserted” cannot pass muster under commercial-speech standards). Put another way, the *807interest in protecting children from indecency only at the caprice of the cable operator is not compelling. Perhaps Congress drafted the law this way to avoid the clear constitutional difficulties of banning indecent speech from access channels, but the First Amendment does not permit this sort of ill fit between a law restricting speech and the interest it is said to serve.

Second, to the extent cable operators prohibit indecent programming on access channels, not only children but adults will be deprived of it. The Government may not “reduce the adult population ... to [viewing] only what is fit for children.” Butler v. Michigan, 352 U. S. 380, 383 (1957). It matters not that indecent programming might be available on the operator’s other channels. The Government has no legitimate interest in making access channels pristine. A block-and-segregate requirement similar to § 10(b), but without its constitutional infirmity of requiring persons to place themselves on a list to receive programming, see ante, at 756-757, protects children with far less intrusion on the liberties of programmers and adult viewers than allowing cable operators to ban indecent programming from access channels altogether. When applying strict scrutiny, we will not assume plausible alternatives will fail to protect compelling interests; there must be some basis in the record, in legislative findings or otherwise, establishing the law enacted as the least restrictive means. Sable Communications, supra, at 128-130. Cf. Turner Broadcasting, 512 U. S., at 664-668. There is none here.

Sections 10(a) and (c) present a classic case of discrimination against speech based on its content. There are legitimate reasons why the Government might wish to regulate or even restrict the speech at issue here, but §§ 10(a) and (c) are not drawn to address those reasons with the precision the First Amendment requires.

*808V

Not only does the plurality fail to apply strict scrutiny, but its reasoning is unpersuasive on its own terms.

The plurality declares § 10(c) unconstitutional because it interferes with local supervisory systems that “can set programming policy and approve or disapprove particular programming services.” Ante, at 762. Replacing these local schemes with a cable operator veto would, in the plurality’s view, “greatly increase the risk that certain categories of programming (say, borderline offensive programs) will not appear,” ante, at 766. Although the plurality terms these local schemes “public/nonprofit programming control systems,” ante, at 763, it does not contend (nor does the record suggest) that any local board or access center has the authority to exclude indecent programming, or to do anything that would cast doubt on the status of public access channels as public fora. Cf. Agosta 88 (New York state law forbids editorial control over public access programs by either the cable operator or the municipality); Comments of Hills-borough County Board of County Commissioners 2, FCC Record (explaining county’s inability to exclude indecent programming). Indeed, “[m]ost access centers surveyed do not prescreen at all, except, as in [two named localities], a high speed run-through for technical quality.” R Aufderheide, Public Access Cable Programming, Controversial Speech, and Free Expression (1992), reprinted in App. 61, 68. As the plurality acknowledges, the record indicates no response to indecent programming by local access centers (whether they prescreen or not) other than “requiring indemnification by programmers, certification of compliance with local standards, time segregation, [and] adult content advisories,” ante, at 762. Those are measures that, if challenged, would likely survive strict scrutiny as narrowly tailored to safeguard children. If those measures, in the words of the plurality, “normally avoid, minimize, or eliminate any child-related *809problems concerning ‘patently offensive’ programming” on public access channels, ante, at 763-764, one is left to wonder why the cable operator veto over leased access programming authorized in § 10(a) is constitutional even under the plurality’s First Amendment analysis. Although I concur in its judgment that § 10(c) is invalid, I cannot agree with the plurality’s reasoning.

In regard to § 10(a), the plurality’s analysis there undermines its claims of faithfulness to our First Amendment jurisprudence and close attention to context.

First, the plurality places some weight on there being “nothing to stop ‘adults who feel the need’ from finding [indecent] programming elsewhere, say, on tape or in theaters,” or on competitive services like direct broadcast television, ante, at 745. The availability of alternative channels of communication may be relevant when we are assessing content-neutral time, place, and manner restrictions, Ward v. Rock Against Racism, 491 U. S. 781, 791, 802 (1989), but the fact that speech can occur elsewhere cannot justify a content-based restriction, Southeastern Promotions, 420 U. S., at 556; Schneider v. State (Town of Irvington), 308 U. S. 147, 163 (1939).

Second, the plurality suggests the permissive nature of § 10(a) at least does not create the same risk of exclusion as a total ban on indecency. Ante, at 745-746. This states the obvious, but the possibility the Government could have imposed more draconian limitations on speech never has justified a lesser abridgment. Indeed, such an argument almost always is available; few of our First Amendment cases involve outright bans on speech. See, e. g., Forsyth County v. Nationalist Movement, 505 U. S. 123, 130-137 (1992) (broad discretion of county administrator to award parade permits and to adjust permit fee according to content of speech violates First Amendment); Bantam Books, Inc. v. Sullivan, 372 U. S. 58 (1963) (informal threats to recommend criminal prosecutions and other pressure tactics by state moral*810ity commission against book publishers violate the First Amendment).

Third, based on its own factual speculations, the plurality discounts the risks created by the law that operators will not run indecent programming on access channels. The plurality takes “a glance at the programming that cable operators allow on their own (nonaccess) channels,” and, espying some indecent programming there, supposes some cable operators may be willing to allow similar programs on leased access channels. Ante, at 746. This sort of surmise, giving the Government the benefit of the doubt when it restricts speech, is an unusual approach to the First Amendment, to put it mildly. Worse, it ignores evidence of industry structure that should cast doubt on the plurality’s sanguine view of the probable fate of programming considered “indecent” under § 10(a). The plurality fails to note that, aside from the indecency provisions of § 10 tacked on in a Senate floor amendment, the 1992 Act strengthened the regulation of leased access channels because it was feared cable operators would exercise their substantial market power to exclude disfavored programmers. The congressional findings in the statute and the conclusions of the Senate Committee on Commerce, Science, and Transportation after more than two years of hearings on the cable market, see S. Rep. No. 102-92, pp. 3-4 (1991), are instructive. Leased access channels had been underused since their inception in 1984, the Senate Committee determined. Id., at 30. Though it recognized the adverse economics of leased access for programmers may have been one reason for the underutilization, the Committee found the obstinacy of cable operators and their control over prices, terms, and conditions also were to blame. Id., at 31.

“The cable operator is almost certain to have interests that clash with that of the programmer seeking to use leased access channels. If their interests were similar, the operator would have been more than willing to carry *811the programmer on regular cable channels. The operator thus has already decided for any number of reasons not to carry the programmer. For example, the operator may believe that the programmer might compete with programming that the [operator] owns or controls. To permit the operator to establish the leased access rate thus makes little sense.” Ibid.

Perhaps some operators will choose to show the indecent programming they now may banish if they can command a better price than other access programmers are willing to pay. In the main, however, leased access programs are the ones the cable operator, for competitive reasons or otherwise, has no interest in showing. And because the cable operator may put to his own commercial use any leased access capacity not taken by unaffiliated programmers, 47 U. S. C. § 532(b)(4), operators have little incentive to allow indecent programming if they have excess capacity on leased access channels.

There is even less reason to think cable operators will choose to show indecent programs on public access channels. The operator is not paid, or paid much, for transmitting programs on these channels; public access programs may compete with the operator’s own programs; the operator will wish to avoid unwanted controversy; and here, as with leased access channels, the operator may reclaim unused PEG capacity for its own paid use, 47 U. S. C. § 531(d)(1).

In the 1992 Act, Congress recognized cable operators might want to exclude unaffiliated or otherwise disfavored programmers from their channels, but it granted operators discretion to do so in regard to but a single category of speech. The obvious consequence invited by the discretion is exclusion. I am not sure why the plurality would suppose otherwise, or contend the practical consequences of § 10(a) would be no worse for programmers than those flowing from the sort of time-segregation requirement approved in Pa-cifica. See ante, at 746-747. Despite its claim of making *812“a more contextual assessment” of these cases, ante, at 748, the plurality ignores a key difference of these cases from Pacifica. There, the broadcaster wanted to air the speech in question; here, the cable operator does not. So the safe harbor of late-night programming permitted by the FCC in Pacifica would likely promote speech, whereas suppression will follow from § 10(a).

VI

In agreement with the plurality’s analysis of § 10(b) of the Act, insofar as it applies strict scrutiny, I join Part III of its opinion. Its position there, however, cannot be reconciled with upholding § 10(a). In the plurality’s view, § 10(b), which standing alone would guarantee an indecent programmer some access to a cable audience, violates the First Amendment, but § 10(a), which authorizes exclusion of indecent programming from access channels altogether, does not. There is little to commend this logic or result. I dissent from the judgment of the Court insofar as it upholds the constitutionality of § 10(a).

joined by The Chief Justice and Justice Scalia, concurring in the judgment in part and dissenting in part.

I agree with the principal opinion’s conclusion that § 10(a) is constitutionally permissible, but I disagree with its conclusion that §§ 10(b) and (c) violate the First Amendment. For many years, we have failed to articulate how, and to what extent, the First Amendment protects cable operators, programmers, and viewers from state and federal regulation. I think it is time we did so, and I cannot go along with Justice Breyer’s assiduous attempts to avoid addressing that issue openly.

I

The text of the First Amendment makes no distinctions among print, broadcast, and cable media, but we have done so. In Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 *813(1969), we held that, in light of the scarcity of broadcasting frequencies, the Government may require a broadcast licensee “to share his frequency with others and to conduct himself as a proxy or fiduciary with obligations to present those views and voices which are representative of his community and which would otherwise, by necessity, be barred from the airwaves.” Id., at 389. We thus endowed the public with a right of access “to social, political, esthetic, moral, and other ideas and experiences.” Id., at 390. That public right left broadcasters with substantial, but not complete, First Amendment protection of their editorial discretion. See, e. g., Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U. S. 94, 117-118 (1973) (“A broadcast licensee has a large measure of journalistic freedom but not as large as that exercised by a newspaper”).

In contrast, we have not permitted that level of government interference in the context of the print media. In Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241 (1974), for instance, we invalidated a Florida statute that required newspapers to allow, free of charge, a right of reply to political candidates whose personal or professional character the paper assailed. We rejected the claim that the statute was constitutional because it fostered speech rather than restricted it, as well as a related claim that the newspaper could permissibly be made to serve as a public forum. Id., at 256-268. We also flatly rejected the argument that the newspaper’s alleged media monopoly could justify forcing the paper to speak in contravention of its own editorial discretion. Id., at 256.

Our First Amendment distinctions between media, dubious from their infancy,1 placed cable in a doctrinal wasteland in which regulators and cable operators alike could not be sure whether cable was entitled to the substantial First Amendment protections afforded the print media or was *814subject to the more onerous obligations shouldered by the broadcast media. See Los Angeles v. Preferred Communications, Inc., 476 U. S. 488, 496 (1986) (Blackmun, J., concurring) (“In assessing First Amendment claims concerning cable access, the Court must determine whether the characteristics of cable television make it sufficiently analogous to another medium to warrant application of an already existing standard or whether those characteristics require a new analysis”). Over time, however, we have drawn closer to recognizing that cable operators should enjoy the same First Amendment rights as the nonbroadcast media.

Our first ventures into the world of cable regulation involved no claims arising under the First Amendment, and we addressed only the regulatory authority of the Federal Communications Commission (FCC) over cable operators.2 Only in later cases did we begin to address the level of First Amendment protection applicable to cable operators. In Preferred Communications, for instance, when a cable operator challenged the city of Los Angeles’ auction process for a single cable franchise, we held that the cable operator had stated a First Amendment claim upon which relief could be granted. Id., at 493. We noted that cable operators communicate various topics “through original programming or by exercising editorial discretion over which stations or programs to include in [their] repertoire.” Id., at 494. Cf. FCC v. Midwest Video Corp., 440 U. S. 689, 707 (1979) (Midwest Video II) (“Cable operators now share with broadcasters a significant amount of editorial discretion regarding what their programming will include”). But we then lik*815ened the operators’ First Amendment interests to those of broadcasters subject to Red Lion’s right of access requirement. 476 U. S., at 494-495.

Five years later, in Leathers v. Medlock, 499 U. S. 439 (1991), we dropped any reference to the relaxed scrutiny permitted by Red Lion. Arkansas had subjected cable operators to the State’s general sales tax, while continuing to exempt newspapers, magazines, and scrambled satellite broadcast television. Cable operators, among others, challenged the tax on First Amendment grounds, arguing that the State could not discriminatorily apply the tax to some, but not all, members of the press. Though we ultimately upheld the tax scheme because it was not content based, we agreed with the operators that they enjoyed the protection of the First Amendment. We found that cable operators engage in speech by providing news, information, and entertainment to their subscribers and that they are “part of the ‘press.’ ” 499 U. S., at 444.

Two Terms ago, in Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622 (1994), we stated expressly what we had implied in Leathers: The Red Lion standard does not apply to cable television. 512 U. S., at 637 (“[T]he rationale for applying a less rigorous standard of First Amendment scrutiny to broadcast regulation ... does not apply in the context of cable regulation”); id., at 639 (“[Application of the more relaxed standard of scrutiny adopted in Red Lion and the other broadcast cases is inapt when determining the First Amendment validity of cable regulation”). While Members of the Court disagreed about whether the must-carry rules imposed by Congress were content based, and therefore subject to strict scrutiny, there was agreement that cable operators are generally entitled to much the same First Amendment protection as the print media. But see id., at 670 (Stevens, J., concurring in part and concurring in judgment) (“Cable operators’ control of essential facilities provides a *816basis for intrusive regulation that would be inappropriate and perhaps impermissible for other communicative media”).

In Turner, by adopting much of the print paradigm, and by rejecting Red Lion, we adopted with it a considerable body of precedent that governs the respective First Amendment rights of competing speakers. In Red Lion, we had legitimized consideration of the public interest and emphasized the rights of viewers, at least in the abstract. Under that view, “[i]t is the right of the viewers and listeners, not the right of the broadcasters, which is paramount.” 395 U. S., at 390. After Turner, however, that view can no longer be given any credence in the cable context. It is the operator’s right that is preeminent. If Tornillo and Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U. S. 1 (1986), are applicable, and I think they are, see Turner, supra, at 681-682 (O’Connor, J., concurring in part and dissenting in part), then, when there is a conflict, a programmer’s asserted right to transmit over an operator’s cable system must give way to the operator’s editorial discretion. Drawing an analogy to the print media, for example, the author of a book is protected in writing the book, but has no right to have the book sold in a particular bookstore without the store owner’s consent. Nor can government force the editor of a collection of essays to print other essays on the same subject.

The Court in Turner found that the FCC’s must-carry rules implicated the First Amendment rights of both cable operators and cable programmers. The rules interfered with the operators’ editorial discretion by forcing them to carry broadcast programming that they might not otherwise carry, and they interfered with the programmers’ ability to compete for space on the operators’ channels. 512 U. S., at 636-637; id., at 675-676 (O’Connor, J., concurring in part and dissenting in part). We implicitly recognized in Turner that the programmer’s right to compete for channel space

*817is derivative of, and subordinate to, the operator’s editorial discretion. Like a freelance writer seeking a paper in which to publish newspaper editorials, a programmer is protected in searching for an outlet for cable programming, but has no freestanding First Amendment right to have that programming transmitted. Cf. Miami Herald Publishing Co. v. Tornillo, 418 U. S., at 256-258. Likewise, the rights of would-be viewers are derivative of the speech rights of operators and programmers. Cf. Virginia Bd. of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U. S. 748, 756-757 (1976) (“Freedom of speech presupposes a willing speaker. But where a speaker exists, . . . the protection afforded is to the communication, to its source and to its recipients both”). Viewers have a general right to see what a willing operator transmits, but, under Tornillo and Pacific Gas, they certainly have no right to force an unwilling operator to speak.

By recognizing the general primacy of the cable operator’s editorial rights over the rights of programmers and viewers, Turner raises serious questions about the merits of petitioners’ claims. None of the petitioners in these cases are cable operators; they are all cable viewers or access programmers or their representative organizations. See Brief for Petitioners in No. 95-124, pp. 5-6; Brief for Petitioners New York Citizens Committee for Responsible Media et al. in No. 95-227, p. 3; Brief for Petitioners Alliance for Community Media et al. in No. 95-227, p. 3. It is not intuitively obvious that the First Amendment protects the interests petitioners assert, and neither petitioners nor the plurality have adequately explained the source or justification of those asserted rights.

Justice Breyer’s detailed explanation of why he believes it is “unwise and unnecessary,” ante, at 742, to choose a standard against which to measure petitioners’ First Amendment claims largely disregards our recent attempt in Turner *818to define that standard.3 His attempt to distinguish Turner on the ground that it did not involve “the effects of television viewing on children,” ante, at 748, is meaningless because that factual distinction has no bearing on the existence and ordering of the free speech rights asserted in these cases.

In the process of deciding not to decide on a governing standard, Justice Breyer purports to discover in our cases an expansive, general principle permitting government to “directly regulate speech to address extraordinary problems, where its regulations are appropriately tailored to resolve those problems without imposing an unnecessarily great restriction on speech.” Ante, at 741. This heretofore unknown standard is facially subjective and openly invites balancing of asserted speech interests to a degree not ordinarily permitted. It is true that the standard I endorse lacks the “flexibility” inherent in the plurality’s balancing approach, ante, at 740, but that relative rigidity is required by our precedents and is not of my own making.

In any event, even if the plurality’s balancing test were an appropriate standard, it could only be applied to protect speech interests that, under the circumstances, are themselves protected by the First Amendment. But, by shifting the focus to the balancing of “complex” interests, ante, at 743, Justice Breyer never explains whether (and if so, how) a programmer’s ordinarily unprotected interest in affirmative transmission of its programming acquires constitutional significance on leased and public access channels. See *819ibid, (“interests of programmers in maintaining access channels”); ibid, (“interests served by the access requirements”). It is that question, left unanswered by the plurality, to which I now turn.

II

A

In 1984, Congress enacted 47 U. S. C. § 532(b), which generally requires cable operators to reserve approximately 10 to 15 percent of their available channels for commercial lease to “unaffiliated persons.” Operators were prohibited from “exercising] any editorial control” over these leased access channels. § 532(c)(2). In 1992, Congress withdrew part of its prohibition on the exercise of the cable operators’ editorial control and essentially permitted operators to censor privately programming that the “operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner.” § 532(h).

Since 1984, federal law has also permitted local franchise authorities to require cable operators , to set aside certain channels for “public, educational, or governmental use” (PEG channels),4 § 531(a), but unlike the leased access provisions, has not directly required operators to do so. As with leased access, Congress generally prohibited cable operators from exercising “any editorial control” over public access channels, but provided that operators could prohibit the transmission of obscene programming. § 531(e); see § 544(d). Section 10(e) of the 1992 Act broadened the operators’ editorial control and instructed the FCC to promulgate regulations enabling a cable operator to ban from its public access channels “any programming which contains obscene material, sexually explicit conduct, or material soliciting or promoting unlawful conduct.” Note following 47 U. S. C. § 531. The *820FCC subsequently promulgated regulations in its Second Report and Order, In re Implementation of Section 10 of the Cable Consumer Protection and Competition Act of 1992: Indecent Programming and Other Types of Materials on Cable Access Channels, 8 FCC Red 2638 (1993) (Second Report and Order). The FCC interpreted Congress’ reference to “sexually explicit conduct” to mean that the programming must be indecent, and its regulations therefore permit cable operators to ban indecent programming from their public access channels. Id., at 2640.

As I read these provisions, they provide leased and public access programmers with an expansive and federally enforced statutory right to transmit virtually any programming over access channels, limited only by the bounds of decency. It is no doubt true that once programmers have been given, rightly or wrongly, the ability to speak on access channels, the First Amendment continues to protect programmers from certain Government intrusions. Certainly, under our current jurisprudence, Congress could not impose a total ban on the transmission of indecent programming. See Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 127 (1989) (striking down total ban on indecent dial-a-porn messages). At the same time, however, the Court has not recognized, as entitled to full constitutional protection, statutorily created speech rights that directly conflict with the constitutionally protected private speech rights of another person or entity.5 We have not found a First Amendment violation in statutory schemes that substantially expand the speech opportunities of the person or entity challenging the scheme.

There is no getting around the fact that leased and public access are a type of forced speech. Though the constitutionality of leased and public access channels is not directly at *821issue in these cases,6 the position adopted by the Court in Turner ineluctably leads to the conclusion that the federal access requirements are subject to some form of heightened scrutiny. See Turner, 512 U. S., at 661-662 (citing Ward v. Rock Against Racism, 491 U. S. 781 (1989); United States v. O’Brien, 391 U. S. 367 (1968)). Under that view, content-neutral governmental impositions on an operator’s editorial discretion may be sustained only if they further an important governmental interest unrelated to the suppression of free speech and are no greater than is essential to further the asserted interest. See id., at 377. Of course, the analysis I joined in Turner would have required strict scrutiny. 512 U. S., at 680-682 (O’Connor, J., concurring in part and dissenting in part).

Petitioners must concede that cable access is not a constitutionally required entitlement and that the right they claim to leased and public access has, by definition, been govern-mentally created at the expense of cable operators’ editorial *822discretion. Just because the Court has apparently accepted, for now, the proposition that the Constitution permits some degree of forced speech in the cable context does not mean that the beneficiaries of a Government-imposed forced speech program enjoy additional First Amendment protections beyond those normally afforded to purely private speakers.

We have said that “[i]n the realm of private speech or expression, government regulation may not favor one speaker over another,” Rosenberger v. Rector and Visitors of Univ. of Va., 515 U. S. 819, 828 (1995), but this principle hardly supports petitioners’ claims, for, if they do anything, the leased and public access requirements favor access programmers over cable operators. I do not see §§ 10(a) and (c) as independent restrictions on programmers, but as intricate parts of the leased and public access restrictions imposed by Congress (and state and local governments) on cable operators. The question petitioners pose is whether §§ 10(a) and (c) are improper restrictions on their free speech rights, but Turner strongly suggests that the proper question is whether the leased and public access requirements (with §§ 10(a) and (c)) are improper restrictions on the operators’ free speech rights. In my view, the constitutional presumption properly runs in favor of the operators’ editorial discretion, and that discretion may not be burdened without a compelling reason for doing so. Petitioners’ view that the constitutional presumption favors their asserted right to speak on access channels is directly contrary to Turner and our established precedents.

It is one thing to compel an operator to carry leased and public access speech, in apparent violation of Tornillo, but it is another thing altogether to say that the First Amendment forbids Congress to give back part of the operators’ editorial discretion, which all recognize as fundamentally protected, in favor of a broader access right. It is no answer to say that leased and public access are content neutral and that *823§§ 10(a) and (c) are not, for that does not change the fundamental fact, which petitioners never address, that it is the operators’ journalistic freedom that is infringed, whether the challenged restrictions be content neutral or content based.

Because the access provisions are part of a scheme that restricts the free speech rights of cable operators and expands the speaking opportunities of access programmers, who have no underlying constitutional right to speak through the cable medium, I do not believe that access programmers can challenge the scheme, or a particular part of it, as an abridgment of their “freedom of speech.” Outside the public forum doctrine, discussed infra, at 826-831, Government intervention that grants access programmers an opportunity to speak that they would not otherwise enjoy— and which does not directly limit programmers’ underlying speech rights — cannot be an abridgment of the same programmers’ First Amendment rights, even if the new speaking opportunity is content based.

The permissive nature of §§ 10(a) and (c) is important in this regard. If Congress had forbidden cable operators to carry indecent programming on leased and public access channels, that law would have burdened the programmer’s right, recognized in Turner, supra, at 645, to compete for space on an operator’s system. The Court would undoubtedly strictly scrutinize such a law. See Sable, 492 U. S., at 126. But §§ 10(a) and (c) do not burden a programmer’s right to seek access for its indecent programming on an operator’s system. Rather, they merely restore part of the editorial discretion an operator would have absent Government regulation without burdening the programmer’s underlying speech rights.7

*824The First Amendment challenge, if one is to be made, must come from the party whose constitutionally protected freedom of speech has been burdened. Viewing the federal access requirements as a whole, it is the cable operator, not the access programmer,8 whose speech rights have been infringed. Consequently, it is the operator, and not the programmer, whose speech has arguably been infringed by these provisions. If Congress passed a law forcing bookstores to sell all books published on the subject of congressional politics, we would undoubtedly entertain a claim by bookstores that this law violated the First Amendment principles established in Tornillo and Pacific Gas. But I doubt that we would similarly find merit in a claim by publishers of gardening books that the law violated their First Amendment rights. If that is so, then petitioners in these cases cannot reasonably assert that the Court should strictly scrutinize the provisions at issue in a way that maximizes their ability to speak over leased and public access channels and, by necessity, minimizes the operators’ discretion.

B

It makes no difference that the leased access restrictions may take the form of common carrier obligations. See Midwest Video II, 440 U. S., at 701; see also Brief for Federal Respondents 23. But see 47 U. S. C. § 541(c) (“Any cable system shall not be subject to regulation as a common carrier or utility by reason of providing any cable service”). That the leased access provisions may be described in common carrier terms does not demonstrate that access programmers *825have obtained a First Amendment right to transmit programming over leased access channels. Labeling leased access a common carrier scheme has no real First Amendment consequences. It simply does not follow from common carrier status that cable operators may not, with Congress’ blessing, decline to carry indecent speech on their leased access channels. Common carriers are private entities and may, consistent with the First Amendment, exercise editorial discretion in the absence of a specific statutory prohibition. Concurring in Sable, Justice Scalia explained: “I note that while we hold the Constitution prevents Congress from banning indecent speech in this fashion, we do not hold that the Constitution requires public utilities to carry it.” 492 U. S., at 133. See also Information Providers’ Coalition for Defense of First Amendment v. FCC, 928 F. 2d 866, 877 (CA9 1991) (“[A] carrier is free under the Constitution to terminate service to dial-a-porn operators altogether”); Carlin Communications, Inc. v. Mountain States Telephone & Telegraph Co., 827 F. 2d 1291, 1297 (CA9 1987) (same), cert. denied, 485 U. S. 1029 (1988); Carlin Communication, Inc. v. Southern Bell Telephone & Telegraph Co., 802 F. 2d 1352, 1357 (CA11 1986) (same).

Nothing about common carrier status per se constitutional-izes the asserted interests of petitioners in these cases, and Justice Kennedy provides no authority for his assertion that common carrier regulations “should be reviewed under the same standard as content-based restrictions on speech in a public forum.” Ante, at 797. Whether viewed as the creation of a common carrier scheme or simply as a regulatory restriction on cable operators’ editorial discretion, the net effect is the same: operators’ speech rights are restricted to make room for access programmers. Consequently, the fact that the leased access provisions impose a form of common carrier obligation on cable operators does not alter my view that Congress’ leased access scheme burdens the constitutionally protected speech rights of cable operators in order *826to expand the speaking opportunities of access programmers, but does not independently burden the First Amendment rights of programmers or viewers.

C

Petitioners argue that public access channels are public forums in which they have First Amendment rights to speak and that § 10(c) is invalid because it imposes content-based burdens on those rights. Brief for Petitioners New York Citizens Committee for Responsible Media et al. in No. 95-227, pp. 8-23; Brief for Petitioners Alliance for Community Media et al. in No. 95-227, pp. 32-35. Though I agree that content-based prohibitions in a public forum “must be narrowly drawn to effectuate a compelling state interest,” Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 46 (1983), I do not agree with petitioners’ antecedent assertion that public access channels are public forums.

We have said that government may designate public property for use by the public as a place for expressive activity and that, so designated, that property becomes a public forum. Id., at 45. Petitioners argue that “[a] local government does exactly that by requiring as a condition of franchise approval that the cable operator set aside a public access channel for the free use of the general public on a first-come, first-served, nondiscriminatory basis.”9 *827Brief for Petitioners Alliance for Community Media et al. in No. 95-227, p. 33. I disagree.

Cable systems are not public property.10 Cable systems are privately owned and privately managed, and petitioners point to no case in which we have held that government may designate private property as a public forum. The public forum doctrine is a rule governing claims of “a right of access to public property,” Perry Ed. Assn., supra, at 44, and has never been thought to extend beyond property generally understood to belong to the government. See International Soc. for Krishna Consciousness, Inc. v. Lee, 505 U. S. 672, 681 (1992) (evidence of expressive activity at rail stations, bus stations, wharves, and Ellis Island was “irrelevant to public fora analysis, because sites such as bus and rail terminals traditionally have had private ownership” (emphasis in original)). See also id., at 678 (public forum is “government” or “public” property); Perry Ed. Assn., supra, at 45 (designated public forum “consists of public property”).

Petitioners point to dictum in Cornelius v. NAACP Legal Defense & Ed. Fund, 473 U. S. 788, 801 (1985), that a public forum may consist of “private property dedicated to public use,” but that statement has no applicability here. That statement properly refers to the common practice of formally dedicating land for streets and parks when subdividing real estate for developments. See 1A C. Antieau & J. Antieau, Antieau’s Local Government Law § 9.05 (1991); 11A E. McQuillin, Law of Municipal Corporations §33.03 (3d ed. 1991). Such dedications may or may not transfer title, but they at least create enforceable public easements in the dedicated land. 1A Antieau, supra, § 9.15; 11A McQuillin, supra, *828§ 33.68. To the extent that those easements create a property interest in the underlying land, it is that government-owned property interest that may be designated as a public forum.

It may be true, as petitioners argue, that title is not dis-positive of the public forum analysis, but the nature of the regulatory restrictions placed on cable operators by local franchising authorities is not consistent with the kinds of governmental property interests we have said may be formally dedicated as public forums. Our public forum cases have involved property in which the government has held at least some formal easement or other property interest permitting the government to treat the property as its own in designating the property as a public forum. See, e. g., Hague v. Committee for Industrial Organization, 307 U. S. 496, 515 (1939) (streets and parks); Police Dept. of Chicago v. Mosley, 408 U. S. 92, 96 (1972) (sidewalks adjoining public school); Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 555 (1975) (theater under long-term lease to city); Carey v. Brown, 447 U. S. 455, 460-462 (1980) (sidewalks in front of private residence); Widmar v. Vincent. 454 U. S. 263, 267-268 (1981) (university facilities that had been opened for student activities). That is simply not true in these cases. Pursuant to federal and state law, franchising authorities require cable operators to create public access channels, but nothing in the record suggests that local franchising authorities take any formal easement or other property interest in those channels that would permit the government to designate that property as a public forum.11

*829Similarly, assertion of government control over private property cannot justify designation of that property as a public forum. We have expressly stated that neither government ownership nor government control will guarantee public access to property. See Cornelius, supra, at 803; Postal Service v. Council of Greenburgh Civic Assns., 453 U. S. 114, 129 (1981). Government control over its own property or private property in which it has taken a cognizable property interest, like the theater in Southeastern Promotions, is consistent with designation of a public forum. But we have never even hinted that regulatory control, and particularly direct regulatory control over a private entity’s First Amendment speech rights, could justify creation of a public forum. Properly construed, our cases have limited the government’s ability to declare a public forum to property the government owns outright, or in which the government holds a significant property interest consistent with the communicative purpose of the forum to be designated.

Nor am I convinced that a formal transfer of a property interest in public access channels would suffice to permit a local franchising authority to designate those channels as a public forum. In no other public forum that we have recognized does a private entity, owner or not, have the obligation not only to permit another to speak, but to actually help produce and then transmit the message on that person’s behalf. Cable operators regularly retain some level of managerial and operational control over their public access channels, subject only to the requirements of federal, state, and local law and the franchise agreement. In more traditional public forums, the government shoulders the burden of administering and enforcing the openness of the expressive forum, but it is frequently a private citizen, the operator, who shoulders that burden for public access channels. For instance, *830it is often the operator who must accept and schedule an access programmer’s request for time on a channel.12 And, in many places, the operator is actually obligated to provide production facilities and production assistance to persons seeking to produce access programming.13 Moreover, unlike a park picketer, an access programmer cannot transmit its own message. Instead, it is the operator who must transmit, or “speak,” the access programmer’s message. That the speech may be considered the operator’s is driven home by 47 U. S. C. § 559, which authorizes a fine of up to $10,000 and two years’ imprisonment for any person who “transmits over any cable system any matter which is obscene.” See also *831§558 (making operators immune for all public access programming, except that which is obscene).14

Thus, even were I inclined to view public access channels as public property, which I am not, the numerous additional obligations imposed on the cable operator in managing and operating the public access channels convince me that these channels share few, if any, of the basic characteristics of a public forum. As I have already indicated, public access requirements, in my view, are a regulatory restriction on the exercise of cable operators’ editorial discretion, not a transfer of a sufficient property interest in the channels to support a designation of that property as a public forum. Public access channels are not public forums, and, therefore, petitioners’ attempt to redistribute cable speech rights in their favor must fail. For this reason, and the other reasons articulated earlier, I would sustain both § 10(a) and § 10(c).

III

Most sexually oriented programming appears on premium or pay-per-view channels that are naturally blocked from nonpaying customers by market forces, see In re Implementation of Section 10 of the Consumer Protection and Competition Act of 1992: Indecent Programming and Other Types of Materials on Cable Access Channels, First Report and Order, 8 FCC Red 998, 1001, n. 20 (1993) (First Report and Order), and it is only governmental intervention in the first instance that requires access channels, on which indecent programming may appear, to be made part of the basic cable package. Section 10(b) does nothing more than adjust the nature of Government-imposed leased access requirements *832in order to emulate the market forces that keep indecent programming primarily on premium channels (without permitting the operator to charge subscribers for that programming).

Unlike §§ 10(a) and (c), § 10(b) clearly implicates petitioners’ free speech rights. Though § 10(b) by no means bans indecent speech, it clearly places content-based restrictions on the transmission of private speech by requiring cable operators to block and segregate indecent programming that the operator has agreed to carry. Consequently, § 10(b) must be subjected to strict scrutiny and can be upheld only if it furthers a compelling governmental interest by the least restrictive means available. See Sable, 492 U. S., at 126. The parties agree that Congress has a “compelling interest in protecting the physical and psychological well-being of minors” and that its interest “extends to shielding minors from the influence of [indecent speech] that is not obscene by adult standards.” Ibid. See Ginsberg v. New York, 390 U. S. 629, 639 (1968) (persons “who have th[e] primary responsibility for children’s well-being are entitled to the support of laws designed to aid discharge of that responsibility”). Because § 10(b) is narrowly tailored to achieve that well-established compelling interest, I would uphold it. I therefore dissent from the Court’s decision to the contrary.

Our precedents establish that government may support parental authority to direct the moral upbringing of their children by imposing a blocking requirement as a default position. For example, in Ginsberg, in which we upheld a State’s ability to prohibit the sale of indecent literature to minors, we pointed out that the State had simply imposed its own default choice by noting that “the prohibition against sales to minors does not bar parents who so desire from purchasing the magazines for their children.” Ibid. Likewise, in Sable we set aside a complete ban on indecent dial-a-porn messages in part because the FCC had previously imposed certain default rules intended to prevent access by minors, *833and there was no evidence that those rules were ineffective. 492 U. S., at 128-130.15

The Court strikes down § 10(b) by pointing to alternatives, such as reverse blocking and lockboxes, that it says are less restrictive than segregation and blocking. Though these methods attempt to place in parents’ hands the ability to permit their children to watch as little, or as much, indecent programming as the parents think proper, they do not effectively support parents’ authority to direct the moral upbringing of their children. See First Report and Order, 8 FCC Red, at 1000-1001.16 The FCC recognized that leased access programming comes “from a wide variety of independent sources, with no single editor controlling [its] selection and presentation.” Id., at 1000. Thus, indecent programming on leased access channels is “especially likely to be shown randomly or intermittently between non-indecent programs.” Ibid. Rather than being able to simply block out certain channels at certain times, a subscriber armed with only a lockbox must carefully monitor all leased access programming and constantly reprogram the lockbox *834to keep out undesired programming. Thus, even assuming that cable subscribers generally have the technical proficiency to properly operate a lockbox, by no means a given, this distinguishing characteristic of leased access channels makes lockboxes and reverse blocking largely ineffective.

Petitioners argue that §10(b)’s segregation and blocking scheme is not sufficiently narrowly tailored because it requires the viewer’s “written consent,” 47 CFR § 76.701(b) (1995); it permits the cable operator 30 days to respond to the written request for access, § 76.701(c); and it is impermis-sibly underinclusive because it reaches only leased access programming.

Relying on Lamont v. Postmaster General, 381 U. S. 301 (1965), petitioners argue that forcing customers to submit a written request for access will chill dissemination of speech. In Lamont, we struck down a statute barring the mail delivery of “ ‘communist political propaganda’ ” to persons who had not requested the Post Office in writing to deliver such propaganda. Id., at 307. The law required the Post Office to keep an official list of persons desiring to receive communist political propaganda, id., at 303, which, of course, was intended to chill demand for such materials. Here, however, petitioners’ allegations of an official list “of those who wish to watch the ‘patently offensive’ channel,” as the majority puts it, ante, at 754, are pure hyperbole. The FCC regulation implementing § 10(b)’s written request requirement, 47 CFR § 76.701(b) (1995), says nothing about the creation of a list, much less an official Government list. It requires only that the cable operator receive written consent. Other statutory provisions make clear that the cable operator may not share that, or any other, information with any other person, including the Government. Section 551 mandates that all personally identifiable information regarding a subscriber be kept strictly confidential and further requires cable operators to destroy any information that is no longer necessary for the purpose for which it was collected. 47 U. S. C. § 551. *835None of the circumstances that figured prominently in La-rnont exists here.

Though petitioners cannot reasonably fear the specter of an officially published list of leased access indecency viewers, it is true that the fact that a subscriber is unblocked is ascertainable, if only by the cable operator. I find no legally significant stigma in that fact. If a segregation and blocking scheme is generally permissible, then a subscriber’s access request must take some form, whether written or oral, and I see nothing nefarious in Congress’ choice of a written, rather than an oral, consent.17 Any request for access to blocked programming — by whatever method — ultimately will make the subscriber’s identity knowable.18 But this is hardly the kind of chilling effect that implicates the First Amendment.

Though making an oral request for access, perhaps by telephone, is slightly less bothersome than making a written request, it is also true that a written request is less subject to fraud “by a determined child.” Ante, at 759. Consequently, despite the fact that an oral request is slightly less restrictive in absolute terms, it is also less effective in supporting parents’ interest in denying enterprising, but paren-tally unauthorized, minors access to blocked programming.

The segregation and blocking requirement was not intended to be a replacement for lockboxes, V-chips, reverse blocking, or other subscriber-initiated measures. Rather, Congress enacted in § 10(b) a default setting under which a subscriber receives no blocked programming without a writ*836ten request. Thus, subscribers who do not want the blocked programming are protected, and subscribers who do want it may request access. Once a subscriber requests access to blocked programming, however, the subscriber remains free to use other methods, such as lockboxes, to regulate the kind of programming shown on those channels in that home.19 Thus, petitioners are wrong to portray § 10(b) as a highly ineffective method of screening individual programs, see Brief for Petitioners in No. 95-124, at 43, and the majority is similarly wrong to suggest that a person cannot “watch a single program . . . without letting the ‘patently offensive’ channel in its entirety invade his household for days, perhaps weeks, at a time,” ante, at 754; see ante, at 756. Given the limited scope of § 10(b) as a defaiüt setting, I see nothing constitutionally infirm about Congress’ decision to permit the cable operator 30 days to unblock or rebloek the segregated channel.

Petitioners also claim that § 10(b) and its implementing regulations are impermissibly underinclusive because they apply only to leased access programming. In R. A. V. v. St. Paul, 505 U. S. 377 (1992), we rejected the view that a content-based restriction is subject to a separate and-independent “underinelusiveness” evaluation. Id., at 387 (“In our view, the First Amendment imposes not an ‘underinclu-siveness’ limitation but a ‘content discrimination’ limitation upon a State’s prohibition of proscribable speech”). See also ante, at 757 (“Congress need not deal with every problem at once”). Also, petitioners’ claim is in tension with the constitutional principle that Congress may not impose a remedy that is more restrictive than necessary to satisfy its asserted compelling interest and with their own arguments pressing that very principle. Cf. R. A. V., supra, at 402 (White, J., concurring in judgment) (though the “overbreadth doctrine *837has the redeeming virtue of attempting to avoid the chilling of protected expression,” an underbreadth challenge “serves no desirable function”).

In arguing that Congress could not impose a blocking requirement without also imposing that requirement on public access and nonaccess channels, petitioners fail to allege, much less argue, that doing so would further Congress’ compelling interest. While it is true that indecent programming appears on nonaccess channels, that programming appears almost exclusively on “per-program or per channel services that subscribers must specifically request in advance, in the same manner as under the blocking approach mandated by section 10(b).” First Report and Order, 8 FCC Red, at 1001, n. 20.20 In contrast to these premium services, leased access channels are part of the basic cable package, and the segregation and blocking scheme Congress imposed does nothing more than convert sexually oriented leased access programming into a free “premium service.”21 Similarly, Congress’ failure to impose segregation and blocking requirements on public access channels may have been based on its judgment that those channels presented a less severe problem of unintended indecency — it appears that most of the anecdotal evidence before Congress involved leased access channels. Congress may also have simply de*838cided to permit the-States and local franchising authorities to address the issue of indecency on public access channels at a local level, in accordance with the local rule policies evinced in 47 U. S. C. § 531. In any event, if the segregation and blocking scheme established by Congress is narrowly tailored to achieve a compelling governmental interest, it does not become constitutionally suspect merely because Congress did not extend the same restriction to other channels on which there was less of a perceived problem (and perhaps no compelling interest).

The United States has carried its burden of demonstrating that § 10(b) and its implementing regulations are narrowly tailored to satisfy a compelling governmental interest. Accordingly, I would affirm the judgment of the Court of Appeals in its entirety. I therefore concur in the judgment upholding § 10(a) and respectfully dissent from that portion of the judgment striking down §§ 10(b) and (c).

concurring.

The difference between § 10(a) and § 10(c) is the difference between a permit and a prohibition. The former restores the freedom of cable operators to reject indecent programs; the latter requires local franchising authorities to reject such programs. While I join the Court’s opinion, I add these comments to emphasize the difference between the two provisions and to endorse the analysis in Part III-B of Justice Kennedy’s opinion even though I do not think it necessary to characterize the public access channels as public fora. Like Justice Souter, I am convinced that it would be unwise to take a categorical approach to the resolution of novel First Amendment questions arising in an industry as dynamic as this. Cf. R. A. V. v. St. Paul, 505 U. S. 377, 426-427 (1992) (Stevens, J., concurring in judgment).

► — I

Federal law requires cable system operators to reserve about 15 percent of their channels for commercial lease to unaffiliated programmers. See 47 U. S. C. § 532(b). On *769these channels, federal law generally prohibits the cable operator from exercising any control over program content, see § 532(c)(2), with one exception: Section 10(a) allows the operator to refuse to air “indecent” programs. In my view, that exception is permissible.

The Federal Government established the leased access requirements to ensure that certain programmers would have more channels available to them. Section 10(a) is therefore best understood as a limitation on the amount of speech that the Federal Government has spared from the censorial control of the cable operator, rather than a direct prohibition against the communication of speech that, in the absence of federal intervention, would flow freely.

I do not agree, however, that § 10(a) established a public forum. Unlike sidewalks and parks, the Federal Government created leased access channels in the course of its legitimate regulation of the communications industry. In so doing, it did not establish an entirely open forum, but rather restricted access to certain speakers, namely, unaffiliated programmers able to lease the air time. By facilitating certain speech that cable operators would not otherwise carry, the leased access channels operate like the must-carry rules, that we considered in Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 643-646 (1994), without reference to our public forum precedents.

When the Federal Government opens cable channels that would otherwise be left entirely in private hands, it deserves more deference than a rigid application of the public forum doctrine would allow. At this early stage in the regulation of this developing industry, Congress should not be put to an all or nothing-at-all choice in deciding whether to open certain cable channels to programmers who would otherwise lack the resources to participate in the marketplace of ideas.

Just as Congress may legitimately limit access to these channels to unaffiliated programmers, I believe it may also limit, within certain reasonable bounds, the extent of the ac*770cess that it confers upon those programmers.1 If the Government had a reasonable basis for concluding that there were already enough classical musical programs or cartoons being telecast — or, perhaps, even enough political debate — I would find no First Amendment objection to an open access requirement that was extended on an impartial basis to all but those particular subjects. A contrary conclusion would ill-serve First Amendment values by dissuading the Government from creating access rights altogether.2

Of course, the fact that the Federal Government may be entitled to some deference in regulating access for cable programmers does not mean that it may evade First Amendment constraints by selectively choosing which speech should be excepted from private control. If the Government spared all speech but that communicated by Republicans from the control of the cable operator, for example, the First Amendment violation would be plain. See Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 806 *771(1985). More subtle viewpoint-based limitations on access also may be prohibited by the First Amendment. See Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 564 (1975) (Douglas, J., dissenting in part and concurring in result in part).

Even though it is often difficult to determine whether a given access restriction impermissibly singles out certain ideas for repression, in these cases I find no basis for concluding that § 10(a) is a species of viewpoint discrimination. By returning control over indecent programming to the cable operator, § 10(a) treats indecent programming on access channels no differently from indecent programming on regular channels. The decision to permit the operator to determine whether to show indecent programming on access channels therefore cannot be said to reflect a governmental bias against the indecent programming that appears on access channels in particular.

Nor can it be argued that indecent programming has no outlet other than leased access channels, and thus that the exclusion of such speech from special protection is designed to prohibit its communication altogether. Petitioners impliedly concede this point when they contend that the indecency restrictions are arbitrarily underinclusive because they do not affect the similarly indecent programming that appears on regular channels.

Moreover, the criteria § 10(a) identifies for limiting access are fully consistent with the Government’s contention that the speech restrictions are not designed to suppress “a certain form of expression that the Government dislikes,” post, at 802 (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part), but rather to protect children from sexually explicit programming on a pervasive medium. In other cases, we have concluded that such a justification is both viewpoint neutral and legitimate. Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115 (1989); *772FCC v. Pacifica Foundation, 438 U. S. 726 (1978). There is no reason to conclude otherwise here.

Finally, § 10(a) cannot be assailed on the somewhat broader ground that it nevertheless reduces the programming available to the adult population to what is suitable for children. Butler v. Michigan, 352 U. S. 380, 383 (1957); post, at 807 (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part). Section 10(a) serves only to ensure that the newly created access right will not require operators to expose children to more unsuitable communications than would otherwise be the case. It is thus far different in both purpose and effect from the provision at issue in Butler, which criminalized the sale of certain books. 352 U. S., at 381.

In sum, § 10(a) constitutes a reasonable, viewpoint-neutral limitation on a federally created access right for certain cable programmers. Accordingly, I would affirm the judgment of the Court of Appeals as to this provision.

II

As both Justice Breyer and Justice Kennedy have explained, the public, educational, and governmental access channels that are regulated by § 10(c) are not creations of the Federal Government. They owe their existence to contracts forged between cable operators and local cable franchising authorities. Ante, at 734, 760-762 (opinion of Breyer, J.); post, at 788-790, 791-794 (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part).

As their name reflects, so-called PEG channels are subject to a variety of local governmental controls and regulations that — apart from any federal requirement — may result either in a prohibition or a requirement that certain types of programs be carried. Ante, at 761-763 (opinion of Breyer, J.) Presumably, as Justice Breyer explains, the local authorities seldom permit programming of the type described by § 10(c) to air. Ante, at 762-763.

*773What is of critical importance to me, however, is that if left to their own devices, those authorities may choose to carry some programming that the Federal Government has decided to restrict. As I read § 10(c), the federal statute would disable local governments from making that choice. It would inject federally authorized private censors into fora from which they might otherwise be excluded, and it would therefore limit local fora that might otherwise be open to all constitutionally protected speech.3

Section 10(c) operates as a direct restriction on speech that, in the absence of federal intervention, might flow freely. The Federal Government is therefore not entitled to the same leeway that I believe it deserves when it enacts provisions, such as § 10(a), that define the limits of federally created access rights. See supra, at 769-770. The Federal Government has no more entitlement to restrict the power of a local authority to disseminate materials on channels of its own creation, than it has to restrict the power of cable operators to do so on channels that they own. In this respect, I agree entirely with Justice Kennedy, save for his designation of these channels as public fora.

.That is not to say that the Federal Government may not impose restrictions on the dissemination of indecent materials on cable television. Although indecent speech is protected by the First Amendment, the Government may have a compelling interest in protecting children from indecent speech on such a pervasive medium. Sable Communications of Cal., Inc. v. FCC, 492 U. S. 116 (1989); FCC v. Pacifica Foundation, 438 U. S. 726 (1978). When the Gov*774ernment acts to suppress directly the dissemination of such speech, however, it may not rely solely on speculation and conjecture. See Sable Communications of Cal., Inc. v. FCC, 492 U. S., at 129-131.

Justice Breyer persuasively demonstrates that the Government has made no effort to identify the harm caused by permitting local franchising authorities to determine the quantum of so-called “indecent” speech that may be aired in their communities. Ante, at 763-766. Nor has the Government attempted to determine whether the intervention of the discretionary censorial authority of a private cable operator constitutes an appropriately limited means of addressing that harm. Ibid. Given the direct nature of the restriction on speech that § 10(c) imposes, the Government has failed to carry its burden of justification. Accordingly, I agree that the judgment of the Court of Appeals with respect to § 10(c) should be reversed.

concurring.

Justice Kennedy’s separate opinion stresses the worthy point that First Amendment values generally are well served by categorizing speech protection according to the respective characters of the expression, its context, and the restriction at issue. Reviewing speech regulations under fairly strict categorical rules keeps the starch in the standards for those moments when the daily politics cries loudest for limiting what may be said.1 Justice Kennedy sees no warrant in these cases for anything but a categorical and rule-based approach applying a fixed level of scrutiny, the strictest, to judge the content-based provisions of §§ 10(a), (b), and (c), and he accordingly faults the principal opinion *775for declining to decide the precise doctrinal categories that should govern the issue at hand. The value of the categorical approach generally to First Amendment security prompts a word to explain why I join the Court’s unwillingness to announce a definitive categorical analysis in these cases.

Neither the speech nor the limitation at issue here may be categorized simply by content. Our prior case most nearly on point dealt not with a flat restriction covering a separate category of indecency at the First Amendment’s periphery, but with less than a total ban, directed to instances of indecent speech easily available to children through broadcasts readily received in the household and difficult or impossible to control without immediate supervision. See FCC v. Pacifica Foundation, 438 U. S. 726, 747 (1978) (plurality opinion) (“It is a characteristic of speech such as this that both its capacity to offend and its 'social value’ . . . vary with the circumstances”).2 It is not surprising that so contextually complex a category was not expressly assigned a standard level of scrutiny for reviewing the Government’s limitation at issue there.3

Nor does the fact that we deal in these cases with cable transmission necessarily suggest that a simple category sub*776ject to a standard level of scrutiny ought to be recognized at this point; while we have found cable television different from broadcast with respect to the factors justifying intrusive access requirements under the rule in Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969), see Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 638-639 (1994) (finding that Red Lion’s, spectrum scarcity rationale had no application to cable), today’s plurality opinion rightly observes that the characteristics of broadcast radio that rendered indecency particularly threatening in Pacifica, that is, its intrusion into the house and accessibility to children, are also present in the case of cable television, ante, at 744-745. It would seem, then, that the appropriate category for cable indecency should be as contextually detailed as the Pacifica example, and settling upon a definitive level-of-scrutiny rule of review for so complex a category would require a subtle judgment; but there is even more to be considered, enough more to demand a subtlety tantamount to prescience.

All of the relevant characteristics of cable are presently in a state of technological and regulatory flux. Recent and far-reaching legislation not only affects the technical feasibility of parental control over children’s access to undesirable material, see, e. g., Telecommunications Act of 1996, §551, 110 Stat. 139-142 (provision for “V-chip” to block sexually explicit or violent programs), but portends fundamental changes in the competitive structure of the industry and, therefore, the ability of individual entities to act as bottlenecks to the free flow of information, see Title III, id., at 114-128 (promoting competition in cable services). As cable and telephone companies begin their competition for control over the single wire that will carry both their services, we can hardly settle rules for review of regulation on the assumption that cable will remain a separable and useful category of First Amendment scrutiny. And as broadcast, cable, and the cybertechnology of the Internet and the World Wide Web approach the day of using a common receiver, we can *777hardly assume that standards for judging the regulation of one of them will not have immense, but now unknown and unknowable, effects on the others.4

Accordingly, in charting a course that will permit reasonable regulation in light of the values in competition, we have to accept the likelihood that the media of communication will become less categorical and more protean. Because we cannot be confident that for purposes of judging speech restrictions it will continue to make sense to distinguish cable from other technologies, and because we know that changes in these regulated technologies will enormously alter the structure of regulation itself, we should be shy about saying the final word today about what will be accepted as reasonable tomorrow. In my own ignorance I have to accept the real possibility that “if we had to decide today . . . just what the First Amendment should mean in cyberspace, ... we would get it fundamentally wrong.” Lessig, The Path of Cyberlaw, 104 Yale L. J. 1743, 1745 (1995).

The upshot of appreciating the fluidity of the subject that Congress must regulate is simply to accept the fact that not every nuance of our old standards will necessarily do for the new technology, and that a proper choice among existing doctrinal categories is not obvious. Rather than definitively settling the issue now, Justice Breyer wisely reasons by direct analogy rather than by rule, concluding that the speech and the restriction at issue in these cases may usefully be measured against the ones at issue in Pacifica.5 If *778that means it will take some time before reaching a final method of review for cases like these, there may be consolation in recalling that 16 years passed, from Roth v. United States, 354 U. S. 476 (1957), to Miller v. California, 413 U. S. 15 (1973), before the modern obscenity rule jelled; that it took over 40 years, from Hague v. Committee for Industrial Organization, 307 U. S. 496 (1939), to Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37 (1983), for the public forum category to settle out; and that a round half-century passed before the clear and present danger of Schenck v. United States, 249 U. S. 47 (1919), evolved into the modern incitement rule of Brandenburg v. Ohio, 395 U. S. 444 (1969) (per curiam).

I cannot guess how much time will go by until the technologies of communication before us today have matured and their relationships become known. But until a category of indecency can be defined both with reference to the new technology and with a prospect of durability, the job of the courts will be just what Justice Breyer does today: recognizing established First Amendment interests through a close analysis that constrains the Congress, without wholly incapacitating it in all matters of the significance apparent here, maintaining the high value of open communication, measuring the costs of regulation by exact attention to fact, and compiling a pedigree of experience with the changing subject. These are familiar judicial responsibilities in times when we know too little to risk the finality of precision, and attention to them will probably take us through the communications revolution. Maybe the judicial obligation to shoulder these responsibilities can itself be captured by a much older rule, familiar to every doctor of medicine: “First, do no harm.”

announced the judgment of the Court and delivered the opinion of the Court with respect to Part III, an opinion with respect to Parts I, II, and V, in which Justice Stevens, Justice O’Connor, and Justice Sou-TER join, and an opinion with respect to Parts IV and VI, in which Justice Stevens and Justice Souter join.

These cases present First Amendment challenges to three statutory provisions that seek to regulate the broadcasting of “patently offensive” sex-related material on cable television. Cable Television Consumer Protection and Competition Act of 1992 (1992 Act or Act), 106 Stat. 1486, §§ 10(a), 10(b), and 10(c), 47 U. S. C. §§ 532(h), 532(j), and note following § 531. The provisions apply to programs broadcast over cable on what are known as “leased access channels” and “public, educational, or governmental channels.” Two of the provisions essentially permit a cable system operator to prohibit the broadcasting of “programming” that the “operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner.” 1992 *733Act, § 10(a); see § 10(c). See also In re Implementation of Section 10 of the Cable Consumer Protection and Competition Act of 1992: Indecent Programming and Other Types of Materials on Cable Access Channels, First Report and Order, 8 FCC Red 998 (1993) (First Report and Order); In re Implementation of Section 10 of the Cable Consumer Protection and Competition Act of1992, Indecent Programming and Other Types of Materials on Cable Access Channels, Second Report and Order, 8 FCC Red 2638 (1993) (Second Report and Order). The remaining provision requires cable system operators to segregate certain “patently offensive” programming, to place it on a single channel, and to block that channel from viewer access unless the viewer requests access in advance and in writing. 1992 Act, § 10(b); 47 CFR § 76.701(g) (1995).

We conclude that the first provision — which permits the operator to decide whether or not to broadcast such programs on leased access channels — is consistent with the First Amendment. The second provision, which requires leased channel operators to segregate and to block that programming, and the third provision, applicable to public, educational, and governmental channels, violate the First Amendment, for they are not appropriately tailored to achieve the basic, legitimate objective of protecting children from exposure to “patently offensive” material.

I

Cable operators typically own a physical cable network used to convey programming over several dozen cable channels into subscribers’ houses. Program sources vary from channel to channel. Most channels carry programming produced by independent firms, including “many national and regional cable programming networks that have emerged in recent years,” Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 629 (1994), as well as some programming that the system operator itself (or an operator affili*734ate) may provide. Other channels may simply retransmit through cable the signals of over-the-air broadcast stations. Ibid. Certain special channels here at issue, called “leased channels” and “public, educational, or governmental channels,” carry programs provided by those to whom the law gives special cable system access rights.

A “leased channel” is a channel that federal law requires a cable system operator to reserve for commercial lease by unaffiliated third parties. About 10 to 15 percent of a cable system’s channels would typically fall into this category. See 47 U. S. C. § 532(b). “[P]ublic, educational, or governmental channels” (which we shall call “public access” channels) are channels that, over the years, local governments have required cable system operators to set aside for public, educational, or governmental purposes as part of the consideration an operator gives in return for permission to install cables under city streets and to use public rights-of-way. See § 531; see also H. R. Rep. No. 98-934, p. 30 (1984) (authorizing local authorities to require creation of public access channels). Between 1984 and 1992, federal law (as had much pre-1984 state law, in respect to public access channels) prohibited cable system operators from exercising any editorial control over the content of any program broadcast over either leased or public access channels. See 47 U. S. C. §§ 531(e) (public access), 532(c)(2) (leased access).

In 1992, in an effort to control sexually explicit programming conveyed over access channels, Congress enacted the three provisions before us. The first two provisions relate to leased channels. The first says:

“This subsection shall permit a cable operator to enforce prospectively a written and published policy of prohibiting programming that the cable operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner as measured by contemporary community standards.” 1992 Act, § 10(a)(2), 106 Stat. 1486.

*735The second provision, applicable only to leased channels, requires cable operators to segregate and to block similar programming if they decide to permit, rather than to prohibit, its broadcast. The provision tells the Federal Communications Commission (FCC or Commission) to promulgate regulations that will (a) require “programmers to inform cable operators if the program[ming] would be indecent as defined by Commission regulations”; (b) require “cable operators to place” such material “on a single channel”; and (e) require “cable operators to block such single channel unless the subscriber requests access to such channel in writing.” 1992 Act, § 10(b)(1). The Commission issued regulations defining the material at issue in terms virtually identical to those we have already set forth, namely, as descriptions or depictions of “sexual or excretory activities or organs in a patently offensive manner” as measured by the cable viewing community. First Report and Order ¶ ¶ 33-38, at 1003-1004. The regulations require the cable operators to place this material on a single channel and to block it (say, by scrambling). They also require the system operator to provide access to the blocked channel “within 30 days” of a subscriber’s written request for access and to reblock it within 30 days of a subscriber’s request to do so. 47 CFR § 76.701(c) (1995).

The third provision is similar to the first provision, but applies only to public access channels. The relevant statutory section instructs the FCC to promulgate regulations that will

“enable a cable operator of a cable system to prohibit the use, on such system, of any channel capacity of any public, educational, or governmental access facility for any programming which contains obscene material, sexually explicit conduct, or material soliciting or promoting unlawful conduct.” 1992 Act, § 10(c), 106 Stat. 1486.

*736The FCC, carrying out this statutory instruction, promulgated regulations defining “sexually explicit” in language almost identical to that in the statute’s leased channel provision, namely, as descriptions or depictions of “sexual or excretory activities or organs in a patently offensive manner” as measured by the cable viewing community. See 47 CFR §76.702 (1995) (incorporating definition from § 76.701(g)).

The upshot is, as we said at the beginning, that the federal law before us (the statute as implemented through regulations) now permits cable operators either to allow or to forbid the transmission of “patently offensive” sex-related materials over both leased and public access channels, and requires those operators, at a minimum, to segregate and to block transmission of that same material on leased channels.

Petitioners, claiming that the three statutory provisions, as implemented by the Commission regulations, violate the First Amendment, sought judicial review of the Commission’s First Report and Order and its Second Report and Order in the United States Court of Appeals for the District of Columbia Circuit. A panel of that Circuit agreed with petitioners that the provisions violated the First Amendment. Alliance for Community Media v. FCC, 10 F. 3d 812 (1993). The entire Court of Appeals, however, heard the case en banc and reached the opposite conclusion. It held that all three statutory provisions (as implemented) were consistent with the First Amendment. Alliance for Community Media v. FCC, 56 F. 3d 105 (1995). Four of the eleven en banc appeals court judges dissented. Two of the dissenting judges concluded that all three provisions violated the First Amendment. Two others thought that either one, or two, but not all three of the provisions, violated the First Amendment. We granted certiorari to review the en banc court’s First Amendment determinations.

*737II

We turn initially to the provision that 'permits cable system operators to prohibit “patently offensive” (or “indecent”) programming transmitted over leased access channels. 1992 Act, § 10(a). The Court of Appeals held that this provision did not violate the First Amendment because the First Amendment prohibits only “Congress” (and, through the Fourteenth Amendment, a “State”), not private individuals, from “abridging the freedom of speech.” Although the court said that it found no “state action,” 56 F. 3d, at 113, it could not have meant that phrase literally, for, of course, petitioners attack (as “abridging] . . . speech”) a congressional statute — which, by definition, is an Act of “Congress.” More likely, the court viewed this statute’s “permissive” provisions as not themselves restricting speech, but, rather, as simply reaffirming the authority to pick and choose programming that a private entity, say, a private broadcaster, would have had in the absence of intervention by any federal, or local, governmental entity.

We recognize that the First Amendment, the terms of which apply to governmental action, ordinarily does not itself throw into constitutional doubt the decisions of private citizens to permit, or to restrict, speech — and this is so ordinarily even where those decisions take place within the framework of a regulatory regime such as broadcasting. Were that not so, courts might have to face the difficult, and potentially restrictive, practical task of deciding which, among any number of private parties involved in providing a program (for example, networks, station owners, program editors, and program producers), is the “speaker” whose rights may not be abridged, and who is the speech-restricting “censor.” Furthermore, as this Court has held, the editorial function itself is an aspect of “speech,” see Turner, 512 U. S., at 636, and a court’s decision that a private party, say, the station owner, is a “censor,” could itself inter*738fere with that private “censor’s” freedom to speak as an editor. Thus, not surprisingly, this Court’s First Amendment broadcasting cases have dealt with governmental efforts to restrict, not governmental efforts to provide or to maintain, a broadcaster’s freedom to pick and to choose programming. Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U. S. 94 (1973) (striking restrictions on broadcaster’s ability to refuse to carry political advertising); Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969) (upholding restrictions on editorial authority); FCC v. League of Women Voters of Cal., 468 U. S. 364 (1984) (striking restrictions); cf. Consolidated Edison Co. of N. Y. v. Public Serv. Comm’n of N. Y., 447 U. S. 530 (1980) (striking ban on political speech by public utility using its billing envelopes as a broadcast medium); Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n of N. Y, 447 U. S. 557 (1980) (striking restriction on public utility advertising).

Nonetheless, petitioners, while conceding that this is ordinarily so, point to circumstances that, in their view, make the analogy with private broadcasters inapposite and make these cases special ones, warranting a different constitutional result. As a practical matter, they say, cable system operators have considerably more power to “censor” program viewing than do broadcasters, for individual communities typically have only one cable system, linking broadcasters and other program providers with each community’s many subscribers. See Turner, supra, at 633 (only one cable system in most communities; nationally more than 60% of homes subscribe to cable, which then becomes the primary or sole source of video programming in the overwhelming majority of these homes). Moreover, concern about system operators’ exercise of this considerable power originally led government— local and federal — to insist that operators provide leased and public access channels free of operator editorial control. H. R. Rep. No. 98-934, at 30-31. To permit system operators to supervise programming on leased access channels will *739create the very private-censorship risk that this anticensorship effort sought to avoid. At the same time, petitioners add, cable systems have two relevant special characteristics. They are unusually involved with government, for they depend upon government permission and government facilities (streets, rights-of-way) to string the cable necessary for their services. And in respect to leased channels, their speech interests are relatively weak because they act less like editors, such as newspapers or television broadcasters, than like common carriers, such as telephone companies.

Under these circumstances, petitioners conclude, Congress’ “permissive” law, in actuality, will “abridge” their free speech. And this Court should treat that law as a con-gressionally imposed, content-based, restriction unredeemed as a properly tailored effort to serve a “compelling interest.” See Simon & Schuster, Inc. v. Members of N. Y. State Crime Victims Bd., 502 U. S. 105, 118 (1991); Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 126 (1989). They further analogize the provisions to constitutionally forbidden content-based restrictions upon speech taking place in “public forums” such as public streets, parks, or buildings dedicated to open speech and communication. See Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 802 (1985); Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983); see also H. R. Rep. No. 98-934, supra, at 30 (identifying public access channels as the electronic equivalent of a “speaker’s soap box”). And, finally, petitioners say that the legal standard the law contains (the “patently offensive” standard) is unconstitutionally vague. See, e. g., Interstate Circuit, Inc. v. Dallas, 390 U. S. 676 (1968) (rejecting censorship ordinance as vague, even though it was intended to protect children).

Like petitioners, Justices Kennedy and Thomas would have us decide these cases simply by transferring and applying literally categorical standards this Court has developed in other contexts. For Justice Kennedy, leased access *740channels are like a common carrier, cablecast is a protected medium, strict scrutiny applies, § 10(a) fails this test, and, therefore, § 10(a) is invalid. Post, at 796-801,806-807. For Justice Thomas, the case is simple because the cable operator who owns the system over which access channels are broadcast, like a bookstore owner with respect to what it displays on the shelves, has a predominant First Amendment interest. Post, at 816-817, 822-824. Both categorical approaches suffer from the same flaws: They import law developed in very different contexts into a new and changing environment, and they lack the flexibility necessary to allow government to respond to very serious practical problems without sacrificing the free exchange of ideas the First Amendment is designed to protect.

The history of this Court’s First Amendment jurisprudence, however, is one of continual development, as the Constitution’s general command that “Congress shall make no law . . . abridging the freedom of speech, or of the press,” has been applied to new circumstances requiring different adaptations of prior principles and precedents. The essence of that protection is that Congress may not regulate speech except in cases of extraordinary need and with the exercise of a degree of care that we have not elsewhere required. See, e. g., Schenck v. United States, 249 U. S. 47, 51-52 (1919); Abrams v. United States, 250 U. S. 616, 627-628 (1919) (Holmes, J., dissenting); West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 639 (1943); Texas v. Johnson, 491 U. S. 397, 418-420 (1989). At the same time, our cases have not left Congress or the States powerless to address the most serious problems. See, e. g., Chaplinsky v. New Hampshire, 315 U. S. 568 (1942); Young v. American Mini Theatres, Inc., 427 U. S. 50 (1976); FCC v. Pacifica Foundation, 438 U. S. 726 (1978).

Over the years, this Court has restated and refined these basic First Amendment principles, adopting them more particularly to the balance of competing interests and the special *741circumstances of each field of application. See, e. g., New York Times Co. v. Sullivan, 376 U. S. 254 (1964) (allowing criticism of public officials to be regulated by civil libel only if the plaintiff shows actual malice); Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974) (allowing greater regulation of speech harming individuals who are not public officials, but still requiring a negligence standard); Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969) (employing highly flexible standard in response to the scarcity problem unique to over-the-air broadcast); Arkansas Writers’ Project, Inc. v. Ragland, 481 U. S. 221, 231-232 (1987) (requiring “compelling state interest” and a “narrowly drawn” means in context of differential taxation of media); Sable, supra, at 126, 131 (applying “compelling interest,” “least restrictive means,” and “narrowly tailored” requirements to indecent telephone communications); Turner, 512 U. S., at 641 (using “heightened scrutiny” to address content-neutral regulations of cable system broadcasts); Central Hudson Gas & Elec. Corp., 447 U. S., at 566 (restriction on commercial speech cannot be “more extensive than is necessary” to serve a “substantial” government interest).

This tradition teaches that the First Amendment embodies an overarching commitment to protect speech from government regulation through close judicial scrutiny, thereby enforcing the Constitution’s constraints, but without imposing judicial formulas so rigid that they become a strait jacket that disables government from responding to serious problems. This Court, in different contexts, has consistently held that government may directly regulate speech to address extraordinary problems, where its regulations are appropriately tailored to resolve those problems without imposing an unnecessarily great restriction on speech. Justices Kennedy and Thomas would have us further declare which, among the many applications of the general approach that this Court has developed over the years, we are applying here. But no definitive choice among competing *742analogies (broadcast, common carrier, bookstore) allows us to declare a rigid single standard, good for now and for all future media and purposes. That is not to say that we reject all the more specific formulations of the standard — they appropriately cover the vast majority of cases involving government regulation of speech. Rather, aware as we are of the changes taking place in the law, the technology, and the industrial structure related to telecommunications, see, e. g., Telecommunications Act of 1996, 110 Stat. 56; S. Rep. No. 104-23 (1995); H. R. Rep. No. 104-204 (1995), we believe it unwise and unnecessary definitively to pick one analogy or one specific set of words now. See Columbia Broadcasting, 412 U. S., at 102 (“The problems of regulation are rendered more difficult because the broadcast industry is dynamic in terms of technological change; solutions adequate a decade ago are not necessarily so now, and those acceptable today may well be outmoded 10 years hence”); Pacifica, supra, at 748 (“We have long recognized that each medium of expression presents special First Amendment problems”). We therefore think it premature to answer the broad questions that Justices Kennedy and Thomas raise in their efforts to find a definitive analogy, deciding, for example, the extent to which private property can be designated a public forum, compare post, at 791-793, 794 (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part), with post, at 826-829 (Thomas, J., concurring in judgment in part and dissenting in part); whether public access channels are a public forum, post, at 791-792 (opinion of Kennedy, J.); whether the Government’s viewpoint neutral decision to limit a public forum is subject to the same scrutiny as a selective exclusion from a pre-existing public forum, post, at 799-803 (opinion of Kennedy, J.); whether exclusion from common carriage must for all purposes be treated like exclusion from a public forum, post, at 797-798 (opinion of Kennedy, J.); and whether the interests of the owners of communica*743tions media always subordinate the interests of all other users of a medium, post, at 816-817 (opinion of Thomas, J.).

Rather than decide these issues, we can decide these cases more narrowly, by closely scrutinizing § 10(a) to assure that it properly addresses an extremely important problem, without imposing, in light of the relevant interests, an unnecessarily great restriction on speech. The importance of the interest at stake here — protecting children from exposure to patently offensive depictions of sex; the accommodation of the interests of programmers in maintaining access channels and of cable operators in editing the contents of their channels; the similarity of the problem and its solution to those at issue in Pacifica; and the flexibility inherent in an approach that permits private cable operators to make editorial decisions, lead us to conclude that § 10(a) is a sufficiently tailored response to an extraordinarily important problem.

First, the provision before us comes accompanied with an extremely important justification, one that this Court has often found compelling — the need to protect children from exposure to patently offensive sex-related material. Sable Communications, 492 U. S., at 126; Ginsberg v. New York, 390 U. S. 629, 639-640 (1968); New York v. Ferber, 458 U. S. 747, 756-757 (1982).

Second, the provision arises in a very particular context— congressional permission for cable operators to regulate programming that, but for a previous Act of Congress, would have had no path of access to cable channels free of an operator’s control. The First Amendment interests involved are therefore complex, and require a balance between those interests served by the access requirements themselves (increasing the availability of avenues of expression to programmers who otherwise would not have them), H. R. Rep. No. 98-934, at 31-36, and the disadvantage to the First Amendment interests of cable operators and other programmers (those to whom the cable operator would have assigned *744the channels devoted to access). See Turner, 512 U. S., at 635-637.

Third, the problem Congress addressed here is remarkably similar to the problem addressed by the FCC in Pacifica, and the balance Congress struck is commensurate with the balance we approved there. In Pacifica this Court considered a governmental ban of a radio broadcast of “indecent” materials, defined in part, like the provisions before us, to include

“ ‘language that describes, in terms patently offensive as measured by contemporary community standards for the broadcast medium, sexual or excretory activities and organs, at times of the day when there is a reasonable risk that children may be in the audience.’ ” 438 U. S., at 732 (quoting 56 F. C. C. 2d 94, 98 (1975)).

The Court found this ban constitutionally permissible primarily because “broadcasting is uniquely accessible to children” and children were likely listeners to the program there at issue — an afternoon radio broadcast. 438 U. S., at 749-750. In addition, the Court wrote, “the broadcast media have established a uniquely pervasive presence in the lives of all Americans,” id., at 748, “[p]atently offensive, indecent material... confronts the citizen, not only in public, but also in the privacy of the home,” generally without sufficient prior warning to allow the recipient to avert his or her eyes or ears, ibid.; and “[a]dults who feel the need may purchase tapes and records or go to theaters and nightclubs” to hear similar performances, id., at 750, n. 28.

All these factors are present here. Cable television broadcasting, including access channel broadcasting, is as “accessible to children” as over-the-air broadcasting, if not more so. See Heeter, Greenberg, Baldwin, Paugh, Srig-ley, & Atkin, Parental Influences on Viewing Style, in Cable-viewing 140 (C. Heeter & B. Greenberg eds. 1988) (children spend more time watching television and view more channels *745than do their parents, whether their household subscribes to cable or receives television over the air). Cable television systems, including access channels, “have established a uniquely pervasive presence in the lives of all Americans.” Pacifica, supra, at 748. See Jost, The Future of Television, 4 The CQ Researcher 1131, 1146 (Dec. 23, 1994) (63% of American homes subscribe to cable); Greenberg, Heeter, D’Alessio, & Sipes, Cable and Noncable Viewing Style Comparisons, in Cableviewing, supra, at 207 (cable households spend more of their day, on average, watching television, and will watch more channels, than households without cable service). “Patently offensive” material from these stations can “confron[t] the citizen” in the “privacy of the home,” Pa-cifica, supra, at 748, with little or no prior warning. Cable-viewing, supra, at 217-218 (while cable subscribers tend to use guides more than do broadcast viewers, there was no difference among these groups in the amount of viewing that was planned, and, in fact, cable subscribers tended to sample more channels before settling on a program, thereby making them more, not less, susceptible to random exposure to unwanted materials). There is nothing to stop “adults who feel the need” from finding similar programming elsewhere, say, on tape or in theaters. In fact, the power of cable systems to control home program viewing is not absolute. Over-the-air broadcasting and direct broadcast satellites already provide alternative ways for programmers to reach the home and are likely to do so to a greater extent in the near future. See generally Telecommunications Act of 1996, §201, 110 Stat. 107 (advanced television services), §205 (direct broadcast satellite), § 302 (video programming by telephone companies), and §304 (availability of navigation devices to enhance multichannel programming); L. Johnson, Toward Competition in Cable Television (1994).

Fourth, the permissive nature of § 10(a) means that it likely restricts speech less than, not more than, the ban at issue in Pacifica. The provision removes a restriction as to *746some speakers — namely, cable operators. See supra, at 743. Moreover, although the provision does create a risk that a program will not appear, that risk is not the same as the certainty that accompanies a governmental ban. In fact, a glance at the programming that cable operators allow on their own (nonaccess) channels suggests that this distinction is not theoretical, but real. See App. 393 (regular channel broadcast of Playboy and “Real Sex” programming). Finally, the provision’s permissive nature brings with it a flexibility that allows cable operators, for example, not to ban broadcasts, but, say, to rearrange broadcast times, better to fit the desires of adult audiences while lessening the risks of harm to children. See First Report and Order ¶ 31, at 1003 (interpreting the Act’s provisions to allow cable operators broad discretion over what to do with offensive materials). In all these respects, the permissive nature of the approach taken by Congress renders this measure appropriate as a means of achieving the underlying purpose of protecting children.

Of course, cable system operators may not always rearrange or reschedule patently offensive programming. Sometimes, as petitioners fear, they may ban the programming instead. But the same may be said of Pacifica’s ban. In practice, the FCC’s daytime broadcast ban could have become a total ban, depending upon how private operators (programmers, station owners, networks) responded to it. They would have had to decide whether to reschedule the daytime show for nighttime broadcast in light of comparative audience demand and a host of other practical factors that similarly would determine the practical outcomes of the provisions before us. The upshot, in both cases, must be uncertainty as to practical consequences — of the governmental ban in the one case and of the permission in the other. That common uncertainty makes it difficult to say the provision here is, in any respect, more restrictive than the order in *747Pacifica. At the same time, in the respects we discussed, the provision is significantly less restrictive.

The existence of this complex balance of interests persuades us that the permissive nature of the provision, coupled with its viewpoint-neutral application, is a constitutionally permissible way to protect children from the type of sexual material that concerned Congress, while accommodating both the First Amendment interests served by the access requirements and those served in restoring to cable operators a degree of the editorial control that Congress removed in 1984.

Our basic disagreement with Justice Kennedy is narrow. Like him, we believe that we must scrutinize § 10(a) with the greatest care. Like Justices Kennedy and Thomas, we believe that the interest of protecting children that § 10(a) purports to serve is compelling. But we part company with Justice Kennedy on two issues. First, Justice Kennedy’s focus on categorical analysis forces him to disregard the cable system operators’ interests. Post, at 805-806. We, on the other hand, recognize that in the context of cable broadcast that involves an access requirement (here, its partial removal), and unlike in most cases where we have explicitly required “narrow tailoring,” the expressive interests of cable operators do play a legitimate role. Cf. Turner, 512 U. S., at 636-637. While we cannot agree with Justice Thomas that everything turns on the rights of the cable owner, see post, at 823-824, we also cannot agree with Justice Kennedy that we must ignore the expressive interests of cable operators altogether. Second, Justice Kennedy’s application of a very strict “narrow tailoring” test depends upon an analogy with a category (“the public forum cases”), which has been distilled over time from the similarities of many cases. Rather than seeking an analogy to a category of cases, however, we have looked to the cases themselves. And, as we have said, we found that Pacifica provides the *748closest analogy and lends considerable support to our conclusion.

Petitioners and Justice Kennedy, see post, at 797-798, 803-804, argue that the opposite result is required by two other cases: Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115 (1989), a case in which this Court found unconstitutional a statute that banned “indecent” telephone messages, and Turner, in which this Court stated that cable broadcast receives full First Amendment protection. See 512 U. S., at 637-641. The ban at issue in Sable, however, was not only a total governmentally imposed ban on a category of communications, but also involved a communications medium, telephone service, that was significantly less likely to expose children to the banned material, was less intrusive, and allowed for significantly more control over what comes into the home than either broadcasting or the cable transmission system before us. See 492 U. S., at 128. The Court’s distinction in Turner, furthermore, between cable and broadcast television, relied on the inapplicability of the spectrum scarcity problem to cable. See 512 U. S., at 637-641. While that distinction was relevant in Turner to the justification for structural regulations at issue there (the “must carry” rules), it has little to do with a case that involves the effects of television viewing on children. Those effects are the result of how parents and children view television programming, and how pervasive and intrusive that programming is. In that respect, cable and broadcast television differ little, if at all. See supra, at 744-745. Justice Kennedy would have us decide that all common carriage exclusions are subject to the highest scrutiny, see post, at 796-799, and then decide these cases on the basis of categories that provide imprecise analogies rather than on the basis of a more contextual assessment, consistent with our First Amendment tradition, of assessing whether Congress carefully and appropriately addressed a serious problem.

*749Petitioners also rely on this Court’s “public forum” cases. They point to Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S., at 45, a case in which this Court said that “public forums” are “places” that the government “has opened for use by the public as a place for expressive activity,” or which “by long tradition . . . have been devoted to assembly and debate.” Ibid. See also Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S., at 801 (assuming public forums may include “private property dedicated to public use”). They add that the Government cannot “enforce a content-based exclusion” from a public forum unless “necessary to serve a compelling state interest” and “narrowly drawn.” Perry, supra, at 45. They further argue that the statute’s permissive provisions unjustifiably exclude material, on the basis of content, from the “public forum” that the Government has created in the form of access channels. Justice Kennedy adds by analogy that the decision to exclude certain content from common carriage is similarly subject to strict scrutiny, and here does not satisfy that standard of review. See post, at 796-799, 805-807.

For three reasons, however, it is unnecessary, indeed, unwise, for us definitively to decide whether or how to apply the public forum doctrine to leased access channels. First, while it may be that content-based exclusions from the right to use common carriers could violate the First Amendment, see post, at 796-800 (opinion of Kennedy, J.), it is not at all clear that the public forum doctrine should be imported wholesale into the area of common carriage regulation. As discussed above, we are wary of the notion that a partial analogy in one context, for which we have developed doctrines, can compel a full range of decisions in such a new and changing area. See supra, at 739-743. Second, it is plain from this Court’s cases that a public forum “may be created for a limited purpose.” Perry, supra, at 46, n. 7; see also Cornelius, supra, at 802 (“[T]he government ‘is not required to indefinitely retain the open character of the facility’”)

*750(quoting Perry, supra, at 46). Our cases have not yet determined, however, that government’s decision to dedicate a public forum to one type of content or another is necessarily subject, to the highest level of scrutiny. Must a local government, for example, show a compelling state interest if it builds a band shell in the park and dedicates it solely to classical music (but not to jazz)? The answer is not obvious. Cf. Perry, supra, at 46, n. 7. But, at a minimum, these cases do not require us to answer it. Finally, and most important, the effects of Congress’ decision on the interests of programmers, viewers, cable operators, and children are the same, whether we characterize Congress’ decision as one that limits access to a public forum, discriminates in common carriage, or constrains speech because of its content. If we consider this particular limitation of indecent television programming acceptable as a constraint on speech, we must no less accept the limitation it places on access to the claimed public forum or on use of a common carrier.

Consequently, if one wishes to view the permissive provisions before us through a “public forum” lens, one should view those provisions as limiting the otherwise totally open nature of the forum that leased access channels provide for communication of other than patently offensive sexual material — taking account of the fact that the limitation was imposed in light of experience gained from maintaining a totally open “forum.” One must still ask whether the First Amendment forbids the limitation. But unless a label alone were to make a critical First Amendment difference (and we think here it does not), the features of these cases that we have already discussed — the Government’s interest in protecting children, the “permissive” aspect of the statute, and the nature of the medium — sufficiently justify the “limitation” on the availability of this forum.

Finally, petitioners argue that the definition of the materials subject to the challenged provisions is too vague, thereby granting cable system operators too broad a program-*751screening authority. Cf. Hoffman Estates v. Flipside, Hoffman Estates, Inc., 455 U. S. 489, 498 (1982) (citing Grayned v. City of Rockford, 408 U. S. 104, 108-109 (1972)) (vague laws may lead to arbitrary enforcement); Dombrowski v. Pfister, 380 U. S. 479, 486-487 (1965) (uncertainty may perniciously chill speech). That definition, however, uses language similar to language previously used by this Court for roughly similar purposes.

The provisions, as augmented by FCC regulations, permit cable system operators to prohibit

“programming that the cable operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner as measured by contemporary community standards.” 1992 Act, § 10(a), 106 Stat. 1486.

See also 47 CFR § 76.702 (1995) (reading approximately the same definition into § 10(c)). This language is similar to language adopted by this Court in Miller v. California, 413 U. S. 15, 24 (1973), as a “guideline]” for identifying materials that States may constitutionally regulate as obscene. In Miller, the Court defined obscene sexual material (material that lacks First Amendment protection) in terms of

“(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.” Ibid, (emphasis added; internal quotation marks omitted).

The language, while vague, attempts to identify the category of materials that Justice Stewart thought could be described only in terms of “I know it when I see it.” Jacobellis v. Ohio, 378 U. S. 184, 197 (1964) (concurring opinion). In *752§ 10(a) and the FCC regulations, without Miller’s qualifiers, the language would seem to refer to material that would be offensive enough to fall within that category but for the fact that the material also has “serious literary, artistic, political or scientific value” or nonprurient purposes.

This history suggests that the statute’s language aims at the kind of programming to which its sponsors referred— pictures of oral sex, bestiality, and rape, see 138 Cong. Rec. 981, 985 (1992) (statement of Sen. Helms) — and not at scientific or educational programs (at least unless done with a highly unusual lack of concern for viewer reaction). Moreover, as this Court pointed out in Pacifica, what is “patently offensive” depends on context (the kind of program on which it appears), degree (not “an occasional expletive”), and time of broadcast (a “pig” is offensive in “the parlor” but not the “barnyard”). 438 U. S., at 748, 750. Programming at 2 o’clock in the morning is seen by a basically adult audience and the “patently offensive” must be defined with that fact in mind.

Further, the statute protects against overly broad application of its standards insofar as it permits cable system operators to screen programs only pursuant to a “written and published policy.” 1992 Act, § 10(a), 106 Stat. 1486. A cable system operator would find it difficult to show that a leased access program prohibition reflects a rational “policy” if the operator permits similarly “offensive” programming to run elsewhere on its system at comparable times or in comparable ways. We concede that the statute’s protection against overly broad application is somewhat diminished by the fact that it permits a cable operator to ban programming that the operator “reasonably believes” is patently offensive. Ibid. (emphasis added). But the “reasonable] belie[f]” qualifier here, as elsewhere in the law, seems designed not to expand the category at which the law aims, but, rather, to provide a legal excuse, for (at least) one honest mistake, from liability that might otherwise attach. Cf. Waters v. Churchill, 511 *753U. S. 661, 682 (1994) (Souter, J., concurring) (public employer’s reasonable belief that employee engaged in unprotected speech excuses liability); United States v. United States Gypsum Co., 438 U. S. 422, 453-455, and n. 29 (1978) (“‘meeting competition’ ” defense in antitrust based on reasonable belief in the necessity to meet competition); Pierson v. Ray, 386 U. S. 547, 555-557 (1967) (police officer has defense to constitutional claim, as did officers of the peace at common law in actions for false arrest, when the officer reasonably believed the statute whose violation precipitated the arrest was valid). And the contours of the shield — reasonableness— constrain the discretion of the cable operator as much as they protect it. If, for example, a court had already found substantially similar programming to be beyond the pale of “patently offensive” material, or if a local authority overseeing the local public, governmental, or educational channels had indicated that materials of the type that the cable operator decides to ban were not “patently offensive” in that community, then the cable operator would be hard pressed to claim that the exclusion of the material was “reasonable.” We conclude that the statute is not impermissibly vague.

For the reasons discussed, we conclude that § 10(a) is consistent with the First Amendment.

III

The statute s second provision significantly differs from the first, for it does not simply permit, but rather requires, cable system operators to restrict speech — by segregating and blocking “patently offensive” sex-related material appearing on leased channels (but not on other channels). 1992 Act, § 10(b). In particular, as previously mentioned, see supra, at 735, this provision and its implementing regulations require cable system operators to place “patently offensive” leased channel programming on a separate channel; to block that channel; to unblock the channel within 30 days of a subscriber’s written request for access; and to *754reblock the channel within 30 days of a subscriber’s request for reblocking. 1992 Act, § 10(b); 47 CFR §§ 76.701(b), (c), (g) (1995). Also, leased channel programmers must notify cable operators of an intended “patently offensive” broadcast up to 30 days before its scheduled broadcast date. §§ 76.701(d), (g).

These requirements have obvious restrictive effects. The several up-to-30-day delays, along with single channel segregation, mean that a subscriber cannot decide to watch a single program without considerable advance planning and without letting the “patently offensive” channel in its entirety invade his household for days, perhaps weeks, at a time. These restrictions will prevent programmers from broadcasting to viewers who select programs day by day (or, through “surfing,” minute by minute); to viewers who would like occasionally to watch a few, but not many, of the programs on the “patently offensive” channel; and to viewers who simply tend to judge a program’s value through channel reputation, i. e., by the company it keeps. Moreover, the “written notice” requirement will further restrict viewing by subscribers who fear for their reputations should the operator, advertently or inadvertently, disclose the list of those who wish to watch the “patently offensive” channel. Cf. Lamont v. Postmaster General, 381 U. S. 301, 307 (1965) (finding unconstitutional a requirement that recipients of Communist literature notify the Post Office that they wish to receive it). Further, the added costs and burdens that these requirements impose upon a cable system operator may encourage that operator to ban programming that the operator would otherwise permit to run, even if only late at night.

The Government argues that, despite these adverse consequences, the “segregate and block” requirements are lawful because they are “the least restrictive means of realizing” a “‘compelling interest,’” namely, ‘“protecting the physical and psychological well-being of minors.’” See Brief for Federal Respondents 11 (quoting Sable, 492 U. S., at 126). *755It adds that, in any event, the First Amendment, as applied in Pacifica, “does not require that regulations of indecency on television be subject to the strictest” First Amendment “standard of review.” Brief for Federal Respondents 11.

We agree with the Government that protection of children is a “compelling interest.” See supra, at 743. But we do not agree that the “segregate and block” requirements properly accommodate the speech restrictions they impose and the legitimate objective they seek to attain. Nor need we here determine whether, or the extent to which, Pacifica does, or does not, impose some lesser standard of review where indecent speech is at issue, compare 438 U. S., at 745-748 (opinion of Stevens, J.) (indecent materials enjoy lesser First Amendment protection), with id., at 761-762 (Powell, J., concurring in part and concurring in judgment) (refusing to accept a lesser standard for nonobscene, indecent material). That is because once one examines this governmental restriction, it becomes apparent that, not only is it not a “least restrictive alternative” and is not “narrowly tailored” to meet its legitimate objective, it also seems considerably “more extensive than necessary.” That is to say, it fails to satisfy this Court’s formulations of the First Amendment’s “strictest,” as well as its somewhat less “strict,” requirements. See, e. g., Sable, 492 U. S., at 126 (“compelling interest” and “least restrictive means” requirements applied to indecent telephone communications); id., at 131 (requiring “narrowly tailored” law); Turner, 512 U. S., at 641 (using “heightened scrutiny” to address content-neutral structural regulations of cable systems); id., at 662 (quoting “‘no greater than . . . essential’ ” language from United States v. O’Brien, 391 U. S. 367, 377 (1968), as an example of “heightened,” less-than-strictest, First Amendment scrutiny); Central Hudson, 447 U. S., at 566 (restriction on commercial speech cannot be “more extensive than is necessary”); Florida Bar v. Went For It, Inc., 515 U. S. 618, 624 (1995) (restriction must be “narrowly drawn”); id., at 632 (there must be a *756“reasonable” “fit” with the objective that legitimates speech restriction). The provision before us does not reveal the caution and care that the standards underlying these various verbal formulas impose upon laws that seek to reconcile the critically important interest in protecting free speech with very important, or even compelling, interests that sometimes warrant restrictions.

Several circumstances lead us to this conclusion. For one thing, the law, as recently amended, uses other means to protect children from similar “patently offensive” material broadcast on mleased cable channels, i. e., broadcast over any of a system’s numerous ordinary, or public access, channels. The law, as recently amended, requires cable operators to “scramble or . . . block” such programming on any (unleased) channel “primarily dedicated to sexually-oriented programming.” Telecommunications Act of 1996, §505, 110 Stat. 136 (emphasis added). In addition, cable operators must honor a subscriber’s request to block any, or all, programs on any channel to which he or she does not wish to subscribe. §504, ibid. And manufacturers, in the future, will have to make television sets with a so-called “V-chip”— a device that will be able automatically to identify and block sexually explicit or violent programs. § 551, id., at 139-142.

Although we cannot, and do not, decide whether the new provisions are themselves lawful (a matter not before us), we note that they are significantly less restrictive than the provision here at issue. They do not force the viewer to receive (for days or weeks at a time) all “patently offensive” programming or none; they will not lead the viewer automatically to judge the few by the reputation of the many; and they will not automatically place the occasional viewer’s name on a special list. They therefore inevitably lead us to ask why, if they adequately protect children from “patently offensive” material broadcast on ordinary channels, they would not offer adequate protection from similar leased channel broadcasts as well? Alternatively, if these provi*757sions do not adequately protect children from “patently offensive” material broadcast on ordinary channels, how could one justify more severe leased channel restrictions when (given ordinary channel programming) they would yield so little additional protection for children?

The record does not answer these questions. It does not explain why, under the new Act, blocking alone — without written access requests — adequately protects children from exposure to regular sex-dedicated channels, but cannot adequately protect those children from programming on similarly sex-dedicated channels that are leased. It does not explain why a simple subscriber blocking request system, perhaps a phone-call-based system, would adequately protect children from “patently offensive” material broadcast on ordinary non-sex-dedicated channels (i. e., almost all channels) but a far more restrictive segregate/block/written-access system is needed to protect children from similar broadcasts on what (in the absence of the segregation requirement) would be non-sex-dedicated channels that are leased. Nor is there any indication Congress thought the new ordinary channel protections less than adequate.

The answers to the questions are not obvious. We have no empirical reason to believe, for example, that sex-dedicated channels are all (or mostly) leased channels, or that “patently offensive” programming on non-sex-dedicated channels is found only (or mostly) on leased channels. To the contrary, the parties’ briefs (and major city television guides) provide examples of what seems likely to be such programming broadcast over both kinds of channels.

We recognize, as the Government properly points out, that Congress need not deal with every problem at once. Cf. Semler v. Oregon Bd. of Dental Examiners, 294 U. S. 608, 610 (1935) (the legislature need not “strike at all evils at the same time”); and Congress also must have a degree of leeway in tailoring means to ends. Columbia Broadcasting, 412 U. S., at 102-103. But in light of the 1996 statute, it seems *758fair to say that Congress now has tried to deal with most of the problem. At this point, we can take Congress’ different, and significantly less restrictive, treatment of a highly similar problem at least as some indication that more restrictive means are not “essential” (or will not prove very helpful). Cf. Boos v. Barry, 485 U. S. 312, 329 (1988) (existence of a less restrictive statute suggested that a challenged ordinance, aimed at the same problem, was overly restrictive).

The record’s description and discussion of a different alternative — the “lockbox” — leads, through a different route, to a similar conclusion. The Cable Communications Policy Act of 1984 required cable operators to provide

“upon the request of a subscriber, a device by which the subscriber can prohibit viewing of a particular cable service during periods selected by the subscriber.” 47 U.S. C. § 544(d)(2).

This device — the “lockbox” — would help protect children by permitting their parents to “lock out” those programs or channels that they did not want their children to see. See FCC 85-179, ¶ 132, 50 Fed. Reg. 18637, 18655 (1985) (“[T]he provision for lockboxes largely disposes of issues involving the Commission’s standard for indecency”). The FCC, in upholding the “segregate and block” provisions, said that lockboxes protected children (including, say, children with inattentive parents) less effectively than those provisions. See First Report and Order ¶ ¶ 14-15, 8 FCC Red, at 1000. But it is important to understand why that is so.

The Government sets forth the reasons as follows:

“In the case of lockboxes, parents would have to discover that such devices exist; find out that their cable operators offer them for sale; spend the time and money to buy one; learn how to program the lockbox to block undesired programs; and, finally, exercise sufficient vigilance to ensure that they have, indeed, locked out *759whatever indecent programming they do not wish their children to view.” Brief for Federal Respondents 37.

We assume the accuracy of this statement. But the reasons do not show need for a provision as restrictive as the one before us. Rather, they suggest a set of provisions very much like those that Congress placed in the 1996 Act.

No provision, we concede, short of an absolute ban, can offer certain protection against assault by a determined child. We have not, however, generally allowed this fact alone to justify “ ‘ “reducing] the adult population . . . to .. . only what is fit for children.”’” Sable, 492 U. S., at 128 (quoting Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 73 (1983), in turn quoting Butler v. Michigan, 352 U. S. 380, 383 (1957)); see Sable, supra, at 130, and n. 10. But, leaving that problem aside, the Government’s list of practical difficulties would seem to call, not for “segregate and block” requirements, but, rather, for informational requirements, for a simple coding system, for readily available blocking equipment (perhaps accessible by telephone), for imposing cost burdens upon system operators (who may spread them through subscription fees); or perhaps even for a system that requires lockbox defaults to be set to block certain channels (say, sex-dedicated channels). These kinds of requirements resemble those that Congress has recently imposed upon all but leased channels. For that reason, the “lockbox” description and the discussion of its frailties reinforces our conclusion that the leased channel provision is overly restrictive when measured against the benefits it is likely to achieve. (We add that the record’s discussion of the “lockbox” does not explain why the law now treats leased channels more restrictively than ordinary channels.)

There may, of course, be other explanations. Congress may simply not have bothered to change the leased channel provisions when it introduced a new system for other channels. But responses of this sort, like guesses about the comparative seriousness of the problem, are hot legally adequate. *760In other cases, where,.as here, the record before Congress or before an agency provides no convincing explanation, this Court has not been willing to stretch the limits of the plausible, to create hypothetical nonobvious explanations in order to justify laws that impose significant restrictions upon speech. See, e. g., Sable, supra, at 130 (“[T]he 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”); Simon & Schuster, 502 U. S., at 120; Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U. S. 575, 585-586 (1983); Arkansas Writers’ Project, 481 U. S., at 231-232.

Consequently, we cannot find that the “segregate and block” restrictions on speech are a narrowly, or reasonably, tailored effort to protect children. Rather, they are overly restrictive, “sacrificing]” important First Amendment interests for too “speculative a gain.” Columbia Broadcasting, 412 U. S., at 127; see League of Women Voters, 468 U. S., at 397. For that reason they are not consistent with the First Amendment.

IV

The statute’s third provision, as implemented by FCC regulation, is similar to its first provision, in that it too permits a cable operator to prevent transmission of “patently offensive” programming, in this case on public access channels. 1992 Act, § 10(c); 47 CFR § 76.702 (1995). But there are four important differences.

The first is the historical background. As Justice Kennedy points out, see post, at 788-790, cable operators have traditionally agreed to reserve channel capacity for public, governmental, and educational channels as part of the consideration they give municipalities that award them cable franchises. See H. R; Rep. No. 98-934, at 30. In the terms preferred by Justice Thomas, see post, at 827-828, the requirement to reserve capacity for public access channels is similar to the reservation of a public easement, or a dedica*761tion of land for streets and parks, as part of a municipality’s approval of a subdivision of land. Cf. post, at 793-794 (opinion of Kennedy, J.). Significantly, these are channels over which cable operators have not historically exercised editorial control. H. R. Rep. No. 98-934, supra, at 30. Unlike § 10(a) therefore, § 10(c) does not restore to cable operators editorial rights that they once had, and the countervailing First Amendment interest is nonexistent, or at least much diminished. See also post, at 792-793 (opinion of Kennedy, J.).

The second difference is the institutional background that has developed as a result of the historical difference. When a “leased channel” is made available by the operator to a private lessee, the lessee has total control of programming during the leased time slot. See 47 U. S. C. § 532(c)(2). Public access channels, on the other hand, are normally subject to complex supervisory systems of various sorts, often with both public and private elements. See § 531(b) (franchising authorities “may require rules and procedures for the use of the [public access] channel capacity”). Municipalities generally provide in their cable franchising agreements for an access channel manager, who is most commonly a nonprofit organization, but may also be the municipality, or, in some instances, the cable system owner. See D. Brenner, M. Price, & M. Myerson, Cable Television and Other Non-broadcast Video ¶ 6.04[7] (1993); P. Aufderheide, Public Access Cable Programming, Controversial Speech, and Free Expression (1992) (hereinafter Aufderheide), reprinted in App. 61, 63 (surveying 61 communities; the access manager was: a nonprofit organization in 41, a local government official in 12, the cable operator in 5, and an unidentified entity in 3); D. Agosta, C. Rogoff, & A. Norman, The Participate Report: A Case Study of Public Access Cable Television in New York State 28 (1990) (hereinafter Agosta), attached as Exh. K to Joint Comments for the Alliance for Community Media et al., filed with the FCC under MM Docket No. 92-*762258 (materials so filed hereinafter FCC Record) (“In 88% [of New York public access systems] access channels were programmed jointly between the cable operator and another institution such as a university, library, or non-profit access organization”); id., at 28-32, FCC Record; Comments of National Cable Television Association Inc., at 14, FCC Record (“Operators often have no involvement in PEG channels that are run by local access organizations”). Access channel activity and management are partly financed with public funds — through franchise fees or other payments pursuant to the franchise agreement, or from general municipal funds, see Brenner, Price, & Myerson, supra, ¶ 6.04[3][c]; Aufder-heide, App. 59-60 — and are commonly subject to supervision by a local supervisory board. See, e. g., D. C. Code Ann. § 43-1829 (1990 and Supp. 1996); Lynchburg City Code § 12.1-44(d)(2) (1988).

This system of public, private, and mixed nonprofit elements, through its supervising boards and nonprofit or governmental access managers, can set programming policy and approve or disapprove particular programming services. And this system can police that policy by, for example,, requiring indemnification by programmers, certification of compliance with local standards, time segregation, adult content advisories, or even by prescreening individual programs. See Second Report and Order ¶ 26, 8 FCC Red, at 2642 (“[F]rom the comments received, it appears that a number of access organizations already have in place procedures that require certification statements [of compliance with local standards], or their equivalent, from access programmers”); Comments of Boston Community Access and Programming Foundation, App. 163-164; Aufderheide, id., at 69-71; Comments of Metropolitan Area Communications Commission 2, FCC Record; Reply Comments of Waycross Community Television 4-6, FCC Record; Reply Comments of Columbus Community Cable Access, Inc., App. 329; Reply Comments of City of St. Paul, id., at 318, 325; Reply Comments of Erik *763Mollberg, Public Access Coordinator, Ft. Wayne, Ind., 3, FCC Record; Comments of Defiance Community Television 3, FCC Record; Comments of Nutmeg Public Access Television, Inc., 3-4, FCC Record. Whether these locally accountable bodies prescreen programming, promulgate rules for the use of public access channels, or are merely available to respond when problems arise, the upshot is the same: There is a locally accountable body capable of addressing the problem, should it arise, of patently offensive programming broadcast to children, making it unlikely that many children will in fact be exposed to programming considered patently offensive in that community. See 56 F. 3d, at 127-128; Second Report and Order ¶ 26, 8 FCC Red 2642.

Third, the existence of a system aimed at encouraging and securing programming that the community considers valuable strongly suggests that a “cable operator’s veto” is less likely necessary to achieve the statute’s basic objective, protecting children, than a similar veto in the context of leased channels. Of course, the system of access managers and supervising boards can make mistakes, which the operator might in some cases correct with its veto power. Balanced against this potential benefit, however, is the risk that the veto itself may be mistaken; and its use, or threatened use, could prevent the presentation of programming, that, though borderline, is not “patently offensive” to its targeted audience. See Aufderheide, App. 64-66 (describing the programs that were considered borderline by access managers, including sex education, health education, broadcasts of politically marginal groups, and various artistic experiments). And this latter threat must bulk large within a system that already has publicly accountable systems for maintaining responsible programs.

Finally, our examination of the legislative history and the record before us is consistent with what common sense suggests, namely, that the public/nonprofit programming control systems now in place would normally avoid, minimize, or *764eliminate any child-related problems concerning “patently offensive” programming. We have found anecdotal references to what seem isolated instances of potentially indecent programming, some of which may well have occurred on leased, not public access, channels. See 138 Cong. Rec. 984, 990 (1992) (statement of Sen. Wirth) (mentioning “abuses” on Time Warner’s New York City channel); but see Comments of Manhattan Neighborhood Network, App. 235, 238 (New York access manager noting that leased, not public access, channels regularly carry sexually explicit programming in New York, and that no commercial programs or advertising are allowed on public access channels); Brief for Time Warner Cable as Amicus Curiae 2-3 (indicating that relevant “abuses” likely occurred on leased channels). See also 138 Cong. Rec., at 989 (statement of Sen. Fowler) (describing solicitation of prostitution); id., at 985 (statement of Sen. Helms) (identifying newspaper headline referring to mayor’s protest of a “strip act”); 56 F. 3d, at 117-118 (recounting comments submitted to the FCC describing three complaints of offensive programming); Letter from Mayor of Rancho Palos Verdes, FCC Record; Resolution of San Antonio City Council, No. 92-49-40, FCC Record.

But these few examples do not necessarily indicate a significant nationwide pattern. See 56 F. 3d, at 127-128 (public access channels “did not pose dangers on the order of magnitude of those identified on leased access channels,” and “local franchising authorities could respond” to such problems “by issuing Tules and procedures’ or other ‘requirements’”). The Commission itself did not report any examples of “indecent” programs on public access channels. See Second Report and Order, 8 FCC Red, at 2638; see also Comments of Boston Community Access and Programming Foundation, App. 162-163 (noting that the FCC’s Notice of Proposed Rulemaking, 7 FCC Red 7709 (1992), did not identify any “inappropriate” programming that actually exists on public *765access channels). Moreover, comments submitted to the FCC undermine any suggestion that prior to 1992 there were significant problems of indecent programming on public access channels. See Agosta 10,28, FCC Record (surveying 76 public access systems in New York over two years, and finding “only two examples of controversial programming, and both had been settled by the producers and the access channel”); Reply Comments of Staten Island Community Television 2, FCC Record (“Our access channels have been on the air since 1986 without a single incident which would be covered by Section 10 of the new law”); Reply Comments of Waycross Community Television, at 2, FCC Record (“[I]n-decent and obscene programs ... [have] never been cablecast through Waycross Community Television during our entire ten year programming history”); Reply Comments of Cambridge Community Television, App. 314 (“In Cambridge less than one hour out of 15,000 hours of programming CCTV has run in the past five year[s] may have been affected by the Act”); ibid. (“CCTV feels that there simply is not a problem which needs to be fixed”); Reply Comments of Columbus Community Cable Access, Inc., id., at 329 (“ACTV is unaware of any actions taken by the cable operators under [a local law authorizing them to prohibit “legally obscene matter”] within the last 10 years”); Reply Comments of Cincinnati Community Video, Inc., id., at 316 (“[I]n 10 years of access operations with over 30,000 access programs cablecast not a single obscenity violation has ever occurred”); Comments of Defiance Community Television, at 2-3, FCC Record (in eight years of operation, “there has never been a serious problem with the content of programming on the channel”).

At most, we have found borderline examples as to which people’s judgment may differ, perhaps acceptable in some communities but not others, of the type that petitioners fear the law might prohibit. See, e. g., Aufderheide, App. 64-66; Brief for Petitioners in No. 95-124, p. 7 (describing depiction *766of a self-help gynecological examination); Comments of Time Warner Entertainment Co., App. 252 (describing an Austin, Tex., program that included “nude scenes from a movie,” and an Indianapolis, Ind., “ ‘safe sex’ ” program). It is difficult to see how such borderline examples could show a compelling need, nationally, to protect children from significantly harmful materials. Compare 138 Cong. Rec., at 985 (statement of Sen. Helms) (justifying regulation of leased access channels in terms of programming that depicts “bestiality” and “rape”). In the absence of a factual basis substantiating the harm and the efficacy of its proposed cure, we cannot assume that the harm exists or that the regulation redresses it. See Turner, 512 U. S., at 664-665.

The upshot, in respect to the public access channels, is a law that could radically change present programming-related relationships among local community and nonprofit supervising boards and access managers, which relationships are established through municipal law, regulation, and contract. In doing so, it would not significantly restore editorial rights of cable operators, but would greatly increase the risk that certain categories of programming (say, borderline offensive programs) will not appear. At the same time, given present supervisory mechanisms, the need for this particular provision, aimed directly at public access channels, is not obvious. Having carefully reviewed the legislative history of the Act, the proceedings before the FCC, the record below, and the submissions of the parties and amici here, we conclude that the Government cannot sustain its burden of showing that § 10(c) is necessary to protect children or that it is appropriately tailored to secure that end. See, e.g., Columbia Broadcasting, 412 U. S., at 127; League of Women Voters, 468 U. S., at 398-399; Sable, 492 U. S., at 126. Consequently, we find that this third provision violates the First Amendment.

*767V

Finally, we must ask whether § 10(a) is severable from the two other provisions. The question is one of legislative intent: Would Congress still “have passed” § 10(a) “had it known” that the remaining “provision^ were] invalid”? Brockett v. Spokane Arcades, Inc., 472 U. S. 491, 506 (1985). If so, we need not invalidate all three provisions' New York v. Ferber, 458 U. S., at 769, n. 24 (citing United States v. Thirty-seven Photographs, 402 U. S. 363 (1971)).

Although the 1992 Act contains no express “severability clause,” we can find the Act’s “severability” intention in its structure and purpose. It seems fairly obvious Congress would have intended its permissive “leased access” channels provision, § 10(a), to stand irrespective of § 10(c)’s legal fate. That is because the latter provision concerns only public, educational, and governmental channels. Its presence had little, if any, effect upon “leased access” channels; hence its absence in respect to those channels could not make a significant difference.

The “segregate and block” requirement’s invalidity does make a difference, however, to the effectiveness of the permissive “leased access” provision, § 10(a). Together they told the cable system operator: “Either ban a ‘patently offensive’ program or ‘segregate and block’ it.” Without the “segregate and block” provision, cable operators are afforded broad discretion over what to do with a patently offensive program, and because they will no longer bear the costs of segregation and blocking if they refuse to ban such programs, cable operators may choose to ban fewer programs.

Nonetheless, this difference does not make the two provisions unseverable. Without the “segregate and block” provision, the law simply treats leased channels (in respect to patently offensive programming) just as it treats all other channels. And judging by the absence of similar segregate and block provisions in the context of these other channels, Congress would probably have thought that § 10(a), standing *768alone, was an effective (though, perhaps, not the most effective) means of pursuing its objective. Moreover, we can find no reason why, in light of Congress’ basic objective (the protection of children), Congress would have preferred no provisions at all to the permissive provision standing by itself. That provision, capable of functioning on its own, still helps to achieve that basic objective. Consequently, we believe the valid provision is severable from the others.

VI

For these reasons, the judgment of the Court of Appeals is affirmed insofar as it upheld § 10(a); the judgment of the Court of Appeals is reversed insofar as it upheld § 10(b) and § 10(c).

It is so ordered.