The manufacturers of video games and their trade association seek to enjoin, as a violation of freedom of expression, the enforcement of an Indianapolis ordinance that seeks to limit the access of minors to video games that depict violence. Denial of a preliminary injunction has precipitated this appeal.

The ordinance defines the term "harmful to minors" to mean "an amusement machine that predominantly appeals to minors' morbid interest in violence or minors' prurient interest in sex, is patently offensive to prevailing standards in the adult community as a whole with respect to what is suitable material for persons under the age of eighteen (18) years, lacks serious literary, artistic, political or scientific value as a whole for persons under" that age, and contains either "graphic violence" or "strong sexual content." "Graphic violence," which is all that is involved in this case (so far as appears, the plaintiffs do not manufacture, at least for exhibition in game arcades and other public places, video games that have "strong sexual content"), is defined to mean "an amusement machine's visual depiction or representation of realistic serious injury to a human or human-like being where such serious injury includes amputation, decapitation, dismemberment, bloodshed, mutilation, maiming or disfiguration [disfigurement]."

The ordinance forbids any operator of five or more video-game machines in one place to allow a minor unaccompanied by a parent, guardian, or other custodian to use "an amusement machine that is harmful to minors," requires appropriate warning signs, and requires that such machines be separated by a partition from the other machines in the location and that their viewing areas be concealed from persons who are on the other side of the partition. Operators of fewer than five games in one location are subject to all but the partitioning restriction. Monetary penalties, as well as suspension and revocation of the right to operate the machines, are specified as remedies for violations of the ordinance.

The ordinance was enacted in 2000, but has not yet gone into effect, in part because we stayed it pending the decision of the appeal. The legislative history indicates that the City believes that participation in violent video games engenders violence on the part of the players, at least when they are minors. The City placed in evidence videotapes of several of the games that it believes violate the ordinance.

Although the district judge agreed with the plaintiffs that video games, possibly including some that would violate the ordinance, are "speech" within the meaning of the First Amendment and that children have rights under the free-speech clause, he held that the ordinance would violate the amendment only if the City lacked "a reasonable basis for believing the Ordinance would protect children from harm." He found a reasonable basis in a pair of empirical studies by psychologists which found that playing a violent video game tends to make young persons more aggressive in their attitudes and behavior, and also in a larger literature finding that violence in the media engenders aggressive feelings. The judge also ruled that the ordinance's tracking of the conventional standard for obscenity eliminated any concern that the ordinance might be excessively vague.

Having decided that the ordinance did not violate the plaintiffs' constitutional rights, the district judge did not consider the other criteria that might bear on the decision to grant or deny a preliminary injunction. In this appeal too, the parties argue only over whether the ordinance is legal, tempting us to treat this as if it were an appeal from a final judgment in favor of the defendants. We shall consider at the end of the opinion whether there is any occasion for further proceedings in the district court.

The ordinance brackets violence with sex, and the City asks us to squeeze the provision on violence into a familiar legal pigeonhole, that of obscenity, which is normally concerned with sex and is not protected by the First Amendment, while the plaintiffs insist that since their games are not obscene in the conventional sense they must receive the full protection of the First Amendment. Neither position is compelling. Violence and obscenity are distinct categories of objectionable depiction, Winters v. New York, 333 U.S. 507, 518-20 (1948); United States v. Thoma, 726 F.2d 1191, 1200 (7th Cir. 1984) ("depictions of torture and deformation are not inherently sexual and, absent some expert guidance as to how such violence appeals to the prurient interest of a deviant group, there is no basis upon which a trier of fact could deem such material obscene"); State v. Johnson, 343 So. 2d 705, 709-10 (La. 1977), and so the fact that obscenity is excluded from the protection of the principle that government may not regulate the content of expressive activity (as distinct from the time, place, or manner of the activity) neither compels nor forecloses a like exclusion of violent imagery. This would be obvious if a pornographer were to argue that because violence is "like" obscenity yet has not yet been placed on the list of expressive forms that can be regulated on the basis of their content, see, e.g., R.A.V. v. City of St. Paul, 505 U.S. 377, 382-84 (1992); DiMa Corp. v. Town of Hallie, 185 F.3d 823, 827 (7th Cir. 1999), obscenity should be struck from the list.

We shall discover some possible intersections between the concerns that animate obscenity laws and the concerns that animate the Indianapolis ordinance as we proceed, but in general the concerns are different. The main worry about obscenity, the main reason for its proscription, is not that it is harmful, which is the worry behind the Indianapolis ordinance, but that it is offensive. A work is classified as obscene not upon proof that it is likely to affect anyone's conduct, but upon proof that it violates community norms regarding the permissible scope of depictions of sexual or sex-related activity. Miller v. California, 413 U.S. 15, 24 (1973); United States v. Moore, 215 F.3d 681, 686 (7th Cir. 2000); United States v. Langford, 688 F.2d 1088, 1091 (7th Cir. 1982); United States v. Loy, 237 F.3d 251, 262 (3d Cir. 2001). Obscenity is to many people disgusting, embarrassing, degrading, disturbing, outrageous, and insulting, but it generally is not believed to inflict temporal (as distinct from spiritual) harm; or at least the evidence that it does is not generally considered as persuasive as the evidence that other speech that can be regulated on the basis of its content, such as threats of physical harm, conspiratorial communications, incitements, frauds, and libels and slanders, inflicts such harm. There are people who believe that some forms of graphically sexual expression, not necessarily obscene in the conventional legal sense, may incite men to commit rape, or to disvalue women in the workplace or elsewhere, see, e.g., Catharine A. MacKinnon, Only Words (1993); but that is not the basis on which obscenity has traditionally been punished. No proof that obscenity is harmful is required either to defend an obscenity statute against being invalidated on constitutional grounds or to uphold a prosecution for obscenity. Offensiveness is the offense.

One can imagine an ordinance directed at depictions of violence because they, too, were offensive. Maybe violent photographs of a person being drawn and quartered could be suppressed as disgusting, embarrassing, degrading, or disturbing without proof that they are likely to cause any of the viewers to commit a violent act. They might even be described as "obscene," in the same way that photographs of people defecating might be, and in many obscenity statutes are, included within the legal category of the obscene, Miller v. California, supra, 413 U.S. at 25; Pope v. Illinois, 481 U.S. 497, 501 n. 4 (1987); United States v. Langford, supra, 688 F.2d at 1091 n. 3, even if they have nothing to do with sex. In common speech, indeed, "obscene" is often just a synonym for repulsive, with no sexual overtones at all.

But offensiveness is not the basis on which Indianapolis seeks to regulate violent video games. Nor could the ordinance be defended on that basis. The most violent game in the record, "The House of the Dead," depicts zombies being killed flamboyantly, with much severing of limbs and effusion of blood; but so stylized and patently fictitious is the cartoon-like depiction that no one would suppose it "obscene" in the sense in which a photograph of a person being decapitated might be described as "obscene." It will not turn anyone's stomach. The basis of the ordinance, rather, is a belief that violent video games cause temporal harm by engendering aggressive attitudes and behavior, which might lead to violence.

This is a different concern from that which animates the obscenity laws, though it does not follow from this that government is helpless to respond to the concern by regulating such games. Protecting people from violence is at least as hallowed a role for government as protecting people from graphic sexual imagery. Chaplinsky v. New Hampshire, 315 U.S. 568, 572-73 (1942), permits punishment of "fighting words," that is, words likely to cause a breach of the peace-- violence. See also R.A.V. v. City of St. Paul, supra, 505 U.S. at 386, 391-92. Such punishment is permissible "content based" regulation, and in effect Indianapolis is arguing that violent video games incite youthful players to breaches of the peace. But this is to use the word "incitement" metaphorically. As we'll see, no showing has been made that games of the sort found in the record of this case have such an effect. Nor can such a showing be dispensed with on the ground that preventing violence is as canonical a role of government as shielding people from graphic sexual imagery. The issue in this case is not violence as such, or directly; it is violent images; and here the symmetry with obscenity breaks down. Classic literature and art, and not merely today's popular culture, are saturated with graphic scenes of violence, whether narrated or pictorial. The notion of forbidding not violence itself, but pictures of violence, is a novelty, whereas concern with pictures of graphic sexual conduct is of the essence of the traditional concern with obscenity.

There is a hint, though, that the City is also concerned with the welfare of the game-playing children themselves, and not just the welfare of their potential victims. This concern is implicit in the City's citation of Ginsberg v. New York, 390 U.S. 629, 639-43 (1968), which holds that potential harm to children's ethical and psychological development is a permissible ground for trying to shield them from forms of sexual expression that fall short of obscenity. See also FCC v. Pacifica Foundation, 438 U.S. 726, 749-50 (1978). Ginsberg upheld a statute that forbade any representation of nudity that "(i) predominantly appeals to the prurient, shameful or morbid interest of minors, and (ii) is patently offensive to prevailing standards in the adult community as a whole with respect to what is suitable material for minors, and (iii) is utterly without redeeming social importance for minors." 390 U.S. at 633. In the present setting, concern with the welfare of the child might take two forms. One is a concern with the potential psychological harm to children of being exposed to violent images, and would be unrelated to the broader societal concern with violence that was the primary motivation for the ordinance. Another, subtler concern would be with the consequences for the child incited or predisposed to commit violent acts by exposure to violent images. In Hoctor v. U.S. Dept. of Agriculture, 82 F.3d 165, 168 (7th Cir. 1996), we noted that the Animal Welfare Act requires secure containment of dangerous animals in part because if they escape and injure a human being they are likely to be killed. A child who is caught and punished for committing a violent act suffers, much as his victim does--indeed, one purpose of punishment is to inflict on the criminal suffering commensurate with that of his victims, either to deter him or others from committing such crimes or (in retributive theory) because it is considered just that he should suffer as his victims do. Obscenity statutes, too, might be thought concerned not just with offensiveness, or with third-party effects (the thrust of the Indianapolis pornography ordinance, a precursor of the present ordinance, invalidated in American Booksellers Association, Inc. v. Hudnut, 771 F.2d 323 (7th Cir. 1985), aff'd without opinion, 475 U.S. 1001 (1986)), but also with the potential harm to the consumer of obscenity, especially a child who might be disturbed by graphic sexual images or suffer psychological harm--and thus Ginsberg. See also Osborne v. Ohio, 495 U.S.103, 111 (1990).

If the community ceased to find obscenity offensive, yet sought to retain the prohibition of it on the ground that it incited its consumers to commit crimes or to engage in sexual discrimination, or that it interfered with the normal sexual development of its underage consumers, a state would have to present a compelling basis for believing that these were harms actually caused by obscenity and not pretexts for regulation on grounds not authorized by the First Amendment. The Court in Ginsberg was satisfied that New York had sufficient grounds for thinking that representations of nudity that would not constitute obscenity if the consumers were adults were harmful to children. We must consider whether the City of Indianapolis has equivalent grounds for thinking that violent video games cause harm either to the game players or (the point the City stresses) the public at large.

The grounds must be compelling and not merely plausible. Children have First Amendment rights. Erznoznik v. City of Jacksonville, 422 U.S. 205, 212-14 (1975); Tinker v. Des Moines Independent School District, 393 U.S. 503, 511-14 (1969). This is not merely a matter of pressing the First Amendment to a dryly logical extreme. The murderous fanaticism displayed by young German soldiers in World War II, alumni of the Hitler Jugend, illustrates the danger of allowing government to control the access of children to information and opinion. Now that eighteen-year- olds have the right to vote, it is obvious that they must be allowed the freedom to form their political views on the basis of uncensored speech before they turn eighteen, so that their minds are not a blank when they first exercise the franchise. And since an eighteen-year-old's right to vote is a right personal to him rather than a right to be exercised on his behalf by his parents, the right of parents to enlist the aid of the state to shield their children from ideas of which the parents disapprove cannot be plenary either. People are unlikely to become well- functioning, independent-minded adults and responsible citizens if they are raised in an intellectual bubble.

No doubt the City would concede this point if the question were whether to forbid children to read without the presence of an adult the Odyssey, with its graphic descriptions of Odysseus's grinding out the eye of Polyphemus with a heated, sharpened stake, killing the suitors, and hanging the treacherous maidservants; or The Divine Comedy with its graphic descriptions of the tortures of the damned; or War and Peace with its graphic descriptions of execution by firing squad, death in childbirth, and death from war wounds. Or if the question were whether to ban the stories of Edgar Allen Poe, or the famous horror movies made from the classic novels of Mary Wollstonecraft Shelley (Frankenstein) and Bram Stoker (Dracula). Violence has always been and remains a central interest of humankind and a recurrent, even obsessive theme of culture both high and low. It engages the interest of children from an early age, as anyone familiar with the classic fairy tales collected by Grimm, Andersen, and Perrault are aware. To shield children right up to the age of 18 from exposure to violent descriptions and images would not only be quixotic, but deforming; it would leave them unequipped to cope with the world as we know it.

Maybe video games are different. They are, after all, interactive. But this point is superficial, in fact erroneous. All literature (here broadly defined to include movies, television, and the other photographic media, and popular as well as highbrow literature) is interactive; the better it is, the more interactive. Literature when it is successful draws the reader into the story, makes him identify with the characters, invites him to judge them and quarrel with them, to experience their joys and sufferings as the reader's own. Protests from readers caused Dickens to revise Great Expectations to give it a happy ending, and tourists visit sites in Dublin and its environs in which the fictitious events of Ulysses are imagined to have occurred. The cult of Sherlock Holmes is well known.

Most of the video games in the record of this case, games that the City believes violate its ordinances, are stories. Take once again "The House of the Dead." The player is armed with a gun--most fortunately, because he is being assailed by a seemingly unending succession of hideous axe-wielding zombies, the living dead conjured back to life by voodoo. The zombies have already knocked down and wounded several people, who are pleading pitiably for help; and one of the player's duties is to protect those unfortunates from renewed assaults by the zombies. His main task, however, is self-defense. Zombies are supernatural beings, therefore difficult to kill. Repeated shots are necessary to stop them as they rush headlong toward the player. He must not only be alert to the appearance of zombies from any quarter; he must be assiduous about reloading his gun periodically, lest he be overwhelmed by the rush of the zombies when his gun is empty.

Self-defense, protection of others, dread of the "undead," fighting against overwhelming odds-- these are all age-old themes of literature, and ones particularly appealing to the young. "The House of the Dead" is not distinguished literature. Neither, perhaps, is "The Night of the Living Dead," George A. Romero's famous zombie movie that was doubtless the inspiration for "The House of the Dead." Some games, such as "Dungeons and Dragons," have achieved cult status; although it seems unlikely, some of these games, perhaps including some that are as violent as those in the record, will become cultural icons. We are in the world of kids' popular culture. But it is not lightly to be suppressed.

Although violent video games appeal primarily to boys, the record contains, surprisingly, a feminist violent video game, "Ultimate Mortal Kombat 3." A man and a woman are dressed in vaguely medieval costumes, and wield huge swords. The woman is very tall, very fierce, and wields her sword effortlessly. The man and the woman duel, and the man is killed. Another man appears- -he is killed too. The woman wins all the duels. She is as strong as the men, she is more skillful, more determined, and she does not flinch at the sight of blood. Of course, her success depends on the player's skill, and the fact that the player, whether male or female, has chosen to be the female fighter. (The player chooses which fighter to be.) But the game is feminist in depicting a woman as fully capable of holding her own in violent combat with heavily armed men. It thus has a message, even an "ideology," just as books and movies do.

We are not persuaded by the City's argument that whatever contribution to the marketplace of ideas and expression the games in the record may have the potential to make is secured by the right of the parent (or guardian, or custodian--and does that include a babysitter?) to permit his or her child or ward to play these games. The right is to a considerable extent illusory. The parent is not permitted to give blanket consent, but must accompany the child to the game room. Many parents are too busy to accompany their child to a game room; most teenagers would be deterred from playing these games if they had to be accompanied by mom; even parents who think violent video games harmful or even edifying (some parents want their kids to develop a shooter's reflexes) may rather prevent their children from playing these games than incur the time and other costs of accompanying the children to the game room; and conditioning a minor's First Amendment rights on parental consent of this nature is a curtailment of those rights.

The City rightly does not rest on "what everyone knows" about the harm inflicted by violent video games. These games with their cartoon characters and stylized mayhem are continuous with an age- old children's literature on violent themes. The exposure of children to the "girlie" magazines involved in the Ginsberg case was not. It seemed obvious to the Supreme Court that these magazines were an adult invasion of children's culture and parental prerogatives. No such argument is available here. The City instead appeals to social science to establish that games such as "The House of the Dead" and "Ultimate Mortal Kombat 3," games culturally isomorphic with (and often derivative from) movies aimed at the same under-18 crowd, are dangerous to public safety. The social science evidence on which the City relies consists primarily of the pair of psychological studies that we mentioned earlier, which are reported in Craig A. Anderson & Karen E. Dill, "Personality Processes and Individual Differences--Video Games and Aggressive Thoughts, Feelings, and Behavior in the Laboratory and in Life," 78 J. Personality & Soc. Psych. 772 (2000). Those studies do not support the ordinance. There is no indication that the games used in the studies are similar to those in the record of this case or to other games likely to be marketed in game arcades in Indianapolis. The studies do not find that video games have ever caused anyone to commit a violent act, as opposed to feeling aggressive, or have caused the average level of violence to increase anywhere. And they do not suggest that it is the interactive character of the games, as opposed to the violence of the images in them, that is the cause of the aggressive feelings. The studies thus are not evidence that violent video games are any more harmful to the consumer or to the public safety than violent movies or other violent, but passive, entertainments. It is highly unlikely that they are more harmful, because "passive" entertainment aspires to be interactive too and often succeeds. When Dirty Harry or some other avenging hero kills off a string of villains, the audience is expected to identify with him, to revel in his success, to feel their own finger on the trigger. It is conceivable that pushing a button or manipulating a toggle stick engenders an even deeper surge of aggressive joy, but of that there is no evidence at all.

We can imagine the City's arguing that it would like to ban violent movies too, but that either this is infeasible or the City has to start somewhere and should not be discouraged from experimenting. Experimentation should indeed not be discouraged. But the City makes neither argument. Its only expressed concern is with video games, in fact only video games in game arcades, movie-theater lobbies, and hotel game rooms. It doesn't even argue that the addition of violent video games to violent movies and television in the cultural menu of Indianapolis youth significantly increases whatever dangers media depictions of violence pose to healthy character formation or peaceable, law-abiding behavior. Violent video games played in public places are a tiny fraction of the media violence to which modern American children are exposed. Tiny--and judging from the record of this case not very violent compared to what is available to children on television and in movie theaters today. The characters in the video games in the record are cartoon characters, that is, animated drawings. No one would mistake them for photographs of real people--another difference between this case and Ginsberg. The idea that a child's interest in such fantasy mayhem is "morbid"--that any kid who enjoys playing "The House of the Dead" or "Ultimate Mortal Kombat 3" should be dragged off to a psychiatrist--gains no support from anything that has been cited to us in defense of the ordinance.

Ginsberg did not insist on social scientific evidence that quasi-obscene images are harmful to children. The Court, as we have noted, thought this a matter of common sense. It was in 1968; it may not be today; but that is not our case. We are not concerned with the part of the Indianapolis ordinance that concerns sexually graphic expression. The video games at issue in this case do not involve sex, but instead a children's world of violent adventures. Common sense says that the City's claim of harm to its citizens from these games is implausible, at best wildly speculative. Common sense is sometimes another word for prejudice, and the common sense reaction to the Indianapolis ordinance could be overcome by social scientific evidence, but has not been. The ordinance curtails freedom of expression significantly and, on this record, without any offsetting justification, "compelling" or otherwise.

It is conceivable though unlikely that in a plenary trial the City can establish the legality of the ordinance. We need not speculate on what evidence might be offered, or, if none is offered (in which event a permanent injunction should promptly be entered), what amendments might bring the ordinance into conformity with First Amendment principles. We have emphasized the "literary" character of the games in the record and the unrealistic appearance of their "graphic" violence. If the games used actors and simulated real death and mutilation convincingly, or if the games lacked any story line and were merely animated shooting galleries (as several of the games in the record appear to be), a more narrowly drawn ordinance might survive a constitutional challenge.

That we need not decide today. The plaintiffs are entitled to a preliminary injunction. Not only have they shown a strong likelihood of ultimate victory should the City persist with the case; they will suffer irreparable harm if the ordinance is permitted to go into effect, because compliance with it will impose costs on them of altering their facilities and will also cause them to lose revenue. And given the entirely conjectural nature of the benefits of the ordinance to the people of Indianapolis, the harm of a preliminary injunction to the City must be reckoned slight, and outweighed by the harm that denying the injunction would impose on the plaintiffs. The judgment is therefore reversed, and the case remanded with instructions to enter a preliminary injunction.

Reversed and Remanded, with Instructions.