Jon Bernthal has made a successful career out of portraying violent men. As Frank Castle in The Punisher, he has spent considerable screen time demonstrating what happens when aggression, physical power and determination are combined in one formidable package. One might therefore have expected him to appreciate an elementary proposition about risk: when something possesses both the capacity to inflict serious injury and characteristics associated with doing so, treating it cautiously is not necessarily irrational prejudice.
Bernthal apparently sees matters differently when it comes to pit bulls. The actor, who owns pit bulls and has long defended them, recently dismissed much of their dangerous reputation as "dog racism." His argument is essentially that pit bulls have been stereotyped, that irresponsible owners are substantially responsible for their behaviour, and that condemning an individual dog because of its breed resembles judging human beings according to group identity. He has also criticised people who acquire pit bulls precisely because they want a frightening or macho dog. There is an important truth buried inside that argument. Individual dogs differ enormously, ownership matters, training matters, socialisation matters, and it would be absurd to assume that every pit bull encountered in the street is about to attack somebody. Bernthal's mistake is to leap from those sensible observations to the conclusion that concern about breed characteristics is merely prejudice.
The problem begins with the word "breed." Dog breeds are not analogous to human racial categories in the sense Bernthal's rhetoric suggests. They are populations deliberately shaped by intense artificial selection for particular physical and behavioural characteristics. We take this proposition for granted whenever the characteristic happens to be useful or charming. Border collies have been selected for herding behaviour. Retrievers retrieve. Pointers point. Livestock guardian dogs possess characteristics useful for guarding livestock. Greyhounds have bodies magnificently adapted for speed. Nobody cries "dog racism" when somebody chooses a border collie for sheep work rather than a pug. The entire institution of dog breeding depends upon the proposition that inherited differences between breeds are real enough to matter.
That does not mean behaviour is genetically predetermined in every individual animal. Genes are not destiny, whether we are discussing dogs or anything else. A dog produced by generations of selection for particular characteristics can still be gentle, badly trained, unusually timid or otherwise different from the statistical tendencies of its population. Environment interacts with inheritance. Responsible ownership can make an enormous difference. Yet none of this makes the concept of breed meaningless. If selective breeding can reliably alter size, skull shape, musculature, coat, speed, scenting ability and herding behaviour, it would be peculiar to insist that behavioural propensities relevant to aggression somehow exist entirely outside biology.
This is where the empirical evidence becomes uncomfortable for the claim that concern about pit bulls is simply prejudice. A 2026 study in the Journal of Surgical Research examined 276 dog-bite patients treated at an American College of Surgeons-verified Level I trauma centre between 2016 and 2024. Dogs reported as pit bulls were responsible for nearly one-third of the injuries and were significantly more likely than other dogs to produce injuries requiring surgical intervention. Some 61.5 per cent of the pit-bull cases required surgery, compared with 47 per cent for other breeds, and pit-bull victims underwent a higher median number of operations.
That is not an internet meme or somebody frightened by the appearance of a muscular dog. It is clinical evidence about people injured badly enough to enter a trauma registry. Nor is it an isolated finding. A systematic review of American dog-bite research covering the literature from 1971 to 2018 concluded that pit bull-type dogs and German shepherds were consistently implicated in the most serious injuries across heterogeneous populations and multiple decades.
There are important qualifications, and anyone making the case against Bernthal's position should acknowledge them. "Pit bull" is itself an imprecise umbrella description encompassing several related types, and visual breed identification can be unreliable. The authors of the 2026 study specifically acknowledged that patient identification of breeds introduced possible bias. Population denominators are another problem. Knowing that a particular breed accounts for a certain proportion of serious attacks does not by itself establish an individual dog's probability of attacking unless we also know accurately how many such dogs exist in the relevant population and how they are kept. Ownership is a confounding factor as well. Powerful dogs may disproportionately attract precisely the sort of irresponsible owners Bernthal criticises.
Those limitations are reasons for careful interpretation. They are not reasons to pretend the evidence says nothing. The intellectually defensible position lies between two caricatures. It is unreasonable to say that every pit bull is a vicious animal waiting to attack, but it is equally unreasonable to say that breed carries no useful information about risk.
There is also a distinction between the probability of an attack and the consequences if an attack occurs. A small dog may be badly behaved and bite frequently while presenting comparatively little danger to a healthy adult. A powerful dog with substantial musculature can produce catastrophic injuries when an attack does occur. Risk is not simply a matter of asking, "Will this dog bite?" It also involves asking, "What happens if it does?"
That is why society already accepts countless forms of discrimination between dogs without considering them morally objectionable. We do not regard a Chihuahua, greyhound, border collie and mastiff as interchangeable biological units. Their physical capacities differ. Their historical functions differ. Insurance companies, landlords, police forces, farmers, breeders and ordinary owners make decisions based partly upon those differences. Sometimes those decisions will be unfair to an exceptionally placid individual animal, but risk management necessarily deals with probabilities rather than perfect foreknowledge.
Bernthal's comparison with racism therefore fails at a more fundamental level. Human racial discrimination raises profound moral questions because human beings are persons possessing individual rights and moral equality. A job applicant should not be rejected because somebody believes people who look like him possess certain average characteristics. Dogs are not participants in a liberal political order possessing a civil right to equal treatment irrespective of breed. We routinely select and regulate animals according to characteristics deliberately created through breeding. Calling that "racism" imports the moral language of human civil rights into a category where it does not comfortably belong.
Indeed, suppose for the sake of argument that we accept Bernthal's deliberately provocative expression "dog racism." It still would not settle the empirical question. If different dog breeds actually possess different physical capabilities and behavioural tendencies, then some forms of what he calls dog racism would simply amount to rational statistical discrimination between animals bred for different purposes. Nobody objects to "dog racism" when choosing a greyhound rather than a bulldog for a race. Nobody demands equal representation of dachshunds in police tracking units or Great Danes as lap dogs in nursing homes. Breed suddenly becomes an offensive consideration only when the characteristic under discussion is dangerousness.
The history of the pit bull matters here. Dogs in the broader pit bull family descend from bull-and-terrier types associated historically with blood sports, including bull-baiting and later dog fighting. That history does not prove that a modern family pet will attack anybody. Modern populations are heterogeneous, breeding practices vary enormously, and many pit bulls live their entire lives without seriously injuring a human being. Nevertheless, historical selection cannot simply be declared irrelevant while we simultaneously recognise inherited breed characteristics everywhere else in the canine world.
Bernthal is on stronger ground when he turns attention towards owners. A powerful dog in the hands of an irresponsible owner is plainly more dangerous than the same animal carefully socialised, trained, controlled and securely contained. People who deliberately cultivate aggression or acquire intimidating animals as status symbols create risks for everyone around them. The 2026 trauma study itself found that ownership status mattered: bites involving non-pet dogs were associated with greater healthcare utilisation, including more operations and higher readmission rates. Breed is therefore not the whole story.
But "not the whole story" is very different from "irrelevant." Road fatalities depend upon driver behaviour, vehicle design, speed, road conditions and chance. We do not respond by declaring that vehicle characteristics cannot influence crash severity because driver behaviour also matters. Risk usually has multiple causes. Dog attacks are no different.
The sensible position is consequently neither pit-bull hysteria nor pit-bull romanticism. Owners should not be treated as social pariahs merely because they possess one, and a well-controlled individual dog should be judged partly upon its actual behaviour. At the same time, communities are entitled to consider evidence about the frequency and severity of attacks when deciding what precautions are reasonable around powerful breeds. Parents in particular have every right to exercise greater caution when children are involved because the consequences of getting the calculation wrong can be devastating.
There is an additional irony in hearing this argument from the man famous for playing the Punisher. Frank Castle's entire fictional existence revolves around threat assessment. He does not wander into a room containing six heavily armed men and reassure himself that stereotyping people carrying assault rifles would be unfair. He notices characteristics relevant to danger and adjusts his behaviour accordingly. Outside the Marvel universe, ordinary people perform much milder versions of the same calculation every day.
None of this requires hatred of pit bulls. It requires only recognition that affection for animals should not prevent us from examining evidence about them. Dogs are wonderful companions precisely because human beings spent thousands of years selectively breeding them into animals with extraordinary and differing characteristics. Celebrating those inherited differences when dogs herd sheep, retrieve ducks or race around a track while denying that inheritance could have any relevance to aggression is scientifically inconsistent.
So perhaps Bernthal has accidentally given us a useful phrase. If taking breed into account when assessing a dog's potential risk is "dog racism," then some dog racism may simply be recognition that breeds are real and that selective breeding has consequences. The humane response is not to demonise every pit bull, but neither is it to ask people to ignore evidence concerning serious attacks because acknowledging breed differences has acquired an offensive label.
A pit bull does not become dangerous merely because somebody fears it, and it does not become harmless because somebody loves it. The relevant questions concern genetics, breeding history, individual temperament, training, ownership, physical capability and the empirical record of injuries. Those questions should be answered with evidence rather than slogans.
Even the Punisher cannot punish the statistics.
https://www.theblaze.com/shows/stu-and-dave-do-america/actors-pit-bull-defense-takes-an-insane-turn-they-were-bred-for-violence