Login
Written by Rebecca Brown
Breed Specific Legislation
The UK currently imposes what’s called ‘Breed Specific Legislation’ in an effort to limit serious injuries due to dog attacks. The legislation was introduced in 1991 and made it illegal to own, sell, abandon, give away or breed dogs deemed to belong to one of four banned breeds. These are the Pit Bull Terrier, Japanese Tosa, Dogo Argentino and Fila Brasileiro. These breeds, having been selectively bred for purposes such as fighting, hunting and guarding, and are considered to have physical and behavioural attributes that mean they pose an unacceptable risk to the public. Dogs that meet the criteria for being a banned breed can be seized and either destroyed or permitted to remain with their owner under restrictive conditions. Breed specific legislation has been recently criticised in a number of organisations.
I do not intend to defend Breed Specific Legislation. It’s plausible that there are alternative, more effective and less damaging ways of reducing harm from dog attacks. However, many of the critiques of Breed Specific Legislation made by prominent animal charities and veterinary bodies are flawed. In pursuing what they no doubt see as a worthwhile end (the scrapping of Breed Specific Legislation), those publicly lobbying for change have made numerous confused and misleading arguments. Below, I outline why these arguments are misleading, implausible or weak, and how they fail to show that Breed Specific Legislation should be revoked.
Any dog can bite
One common claim is that “All dogs, whatever their breed type or size, are capable of showing aggression” and that “All dogs have the potential to be dangerous”. I don’t want to quibble about whether or not a Chihuahua is technically capable of inflicting a serious injury, but the relevant question is surely a) how likely a given dog is to bite a person, coupled with b) how damaging that bite is likely to be.
The Royal Society for Prevention of Cruelty to Animals (RSPCA) point to the more relevant claim that “recent studies found no difference observed between legislated and non-legislated breeds in the medical treatment required following a bite, or in the severity of bite and the type of dog that bit.” Unfortunately they don’t provide any citation or link to this evidence so it’s impossible to evaluate it. So we are left with the frankly implausible claim that if you get bitten by a toy poodle puppy you’re as likely to be harmed as if you are bitten by a breed of dog that was selectively bred to be able to fight a bull to the death.
Elsewhere on the RSPCA website is a document about dog aggression. Here they seem to acknowledge – albeit carefully – the fairly obvious point that some dog breeds are more likely to show aggression, and are more dangerous when aggressive, than others:
Although it might seem that some dogs are born to be aggressive, it is more accurate to say that they are born with inherited tendencies that might, if not controlled, make aggressive behaviour more likely… There are inherited ways of behaving that are particular to some breeds or types of dogs that make it more likely for individuals to grow up to use aggression where others would not… Aggression is not a single characteristic, however there are breeds of dogs that have historically been used for specific purposes, such as for fighting dogs or other animals, or for guarding. Whilst these breeds may not be any more likely to show aggression, because of their physical and temperamental attributes if they do show aggression it is likely to have more serious consequences. Persistence in attack coupled with strong jaws can cause serious injuries.
Lack of evidence
Notwithstanding the above quote from the RSPCA, another common claim is that there’s no ‘robust scientific evidence’ that some dogs (and types of dog) are able to inflict more damaging injuries than others. Again, this seems to be a case of absence of evidence rather than evidence of absence. It may be worth recalling Yeh et al’s (2018) article ‘Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial’ which usefully highlights how direct RCT evidence might not be necessary in order to conclude that interventions such as parachutes are probably a good idea when jumping out of aeroplanes. Thinking again of the toy poodle versus pit bull example, we can probably draw some fairly reliable conclusions based upon ‘mechanistic reasoning’, without a need to inflict different kinds of dogs bites upon participants as part of a randomised trial.
Disconfirmatory evidence
It is also claimed that there is direct evidence that banned breeds are no more dangerous than other breeds, and that Breed Specific Legislation is completely ineffective at reducing (serious) injuries from dogs.
The RSPCA tells us that “Between 1989 and 2017, 48 people died in dog-related incidents. Of the 62 dogs involved, 53 were dog breeds not on the prohibited list.” And that “Only 8% of dangerously out of control dog cases involved banned breeds”. They also reference evidence that “in the past 20 years (1999-2019), the number of hospital admissions for the treatment of dog bites has increased by 154%, despite the prohibition of certain types of dogs”.
What can we make of this? First of all, the population data: dog bites requiring hospital treatment do not appear to have dropped as a result of the introduction of Breed Specific Legislation. This may well be the case, but it’s not possible to tell based on the information provided. We don’t know how dog ownership has changed over the time period studied – perhaps there were lots more dogs, and this resulted in more bites. It has certainly been speculated that the enthusiasm for pets during the covid lockdown (the number of dogs registered with the UK Kennel Club increased by nearly 40% between 2020 and 2021) might have contributed to more dog bite injuries in recent years. The population data might be suggestive, but it is far from conclusive evidence that Breed Specific Legislation doesn’t or hasn’t prevented any serious injuries from dogs since it was introduced.
Second, the breed-linked data, supposedly showing that banned breeds are responsible for only a small minority of serious injuries, including deaths. But wait: banned breeds, of which there are only four (amongst hundreds of other dog breeds) were involved in 1/7 (14%) of the fatal attacks on people during the period mentioned; they were implicated in 8% of cases of ‘dangerously out of control’ dogs. We don’t know how many dogs belonging to banned breeds exist in the UK, but surely it’s less than 14%, meaning they’re at least overrepresented in these samples.
And surely we can go further than that. I’m pretty keen on dogs and had a childhood of obsessively recording and watching all the annual coverage of Crufts repeatedly. I can reliably tell the difference between an Italian Greyhound, a Whippet and a Greyhound, or describe to you what a Wirehaired Vizsla looks like. But I have never heard of three of the breeds on the banned list and possibly never seen them.
This Wikipedia article lists fatal dog attacks in the UK, including breed information where known. The breeds responsible for fatal attacks on humans since 1980 are largely unsurprising. In the below table I’ve summarised the data from the Wikipedia article. (Note that I’ve grouped some breeds / breed types together [Mastiffs; Bulldogs other than American Bulldogs]).
| Breed | Number of fatalities involved in since 1980 (including as part of cross-breed) |
| Staffordshire Bull Terrier | 12 |
| American Bulldog | 9 |
| American Bully XL | 7 |
| Mastiff / Bull Mastiff / Neapolitan Mastiff / Italian Mastiff | 7 |
| Alsatian / German Shepherd | 6 |
| American Pit Bull Terrier / Pit Bull type | 6 |
| Rottweiler | 6 |
| Bordeaux Bulldog / Bulldog type / Aylestone Bulldog / British Bulldog | 4 |
| Jack Russell Terrier | 3 |
| Bull Terrier | 2 |
| Cane Corso | 2 |
| Husky | 2 |
| Doberman | 1 |
| Alaskan Malamute | 1 |
| Lakeland Terrier | 1 |
| Presa Canario | 1 |
| Chow Chow | 1 |
Again, without baseline information about how common these breeds are, one must be cautious about extrapolating from this list to draw conclusions about the dangers of specific breeds. Yet what is clear is that, with the exception of Jack Russell Terriers and a single Lakeland Terrier (responsible for killing three newborn infants), small dogs do not tend to kill people. Dogs that end up inflicting injuries sufficient to kill people tend to be powerful types, which have been selectively bred for guarding, fighting or hunting.
These dogs don’t deserve to die
One unfortunate refrain used by critics including the RSPCA and British Veterinary Association is that dogs deemed to belong to a banned breed and which do not qualify for an exemption, despite showing no previous signs of aggressive or dangerous behaviour, do not deserve to die. Yet this seems to mischaracterise what the Breed Specific Legislation is intended to do. It is clearly not intended to hand out punishment to ‘bad’ dogs, but instead to act as a preventative measure, to stop injuries in the first place. Indeed, since dogs are not moral agents it is entirely inappropriate to punish them on the basis that they ‘deserve’ it. We should of course treat dogs kindly and not cause them unnecessary suffering. But to suggest that Breed Specific Legislation is a form of undeserved punishment of ‘innocent’ dogs is misleading. Even the destruction of dogs that have seriously injured or killed people should not be construed as punishment: the justification for such actions is to protect people from future attacks from a dog proven to be dangerous.
Banning breeds implies that all other dogs are safe
Commentators suggest that, by banning certain breeds, the government sends the message that only these breeds are dangerous, that dogs belonging to other breeds are never dangerous, and that it is breed alone (rather than circumstances, handling, training, etc.) that makes a dog dangerous. Yet I wonder if this is really the case. It hadn’t crossed my mind that only those banned breeds were (potentially) dangerous, and I would be surprised if other people concluded this from the legislation. Indeed, it seems equally plausible that banning specific breeds would raise awareness of general dangers posed by all dogs.
A vet writing for the British Veterinary Association states “there has never been an onus on anyone, young or old, to behave sensibly and respectfully around dogs.” The problem is that many serious bites – particularly those resulting in fatalities – happen to children. 31 of the 64 fatalities in the UK since 1980 listed on the Wikipedia page were in children under 12. They include a number of babies, one of whom was 5 days old when she died. Children are particularly vulnerable to dog attacks: they are small (and thus injuries may be more severe and they are less able to escape or protect themselves if attacked); they are more likely to provoke dogs through excitement or rough treatment; they are more likely to miss the signs that a dog feels threatened and may attack. But there is a limit to what small children can be taught or the extent to which parents can control their behaviour, and we inevitably place them at risk by putting them in shared spaces with dogs. Of course there is an onus on people to treat dogs carefully and kindly, and to ensure those without the capacity to do this are protected. Yet focusing on whether or not children or their parents could (or should) have behaved differently once again seems to attend to questions of blame and desert, rather than the core matter: how best to prevent injuries in the first place.
Whilst it may well be a good idea to provide training and support to people regarding how to safely handle and act around dogs, this is not incompatible with Breed Specific Legislation. We need not accept the supposition that people are too stupid to simultaneously appreciate that poor training, irresponsible handling, stressful circumstances and the inherited predispositions and physical characteristics of a dog can all contribute to how likely it is to bite and injure someone.
Don’t judge a book by its cover
An interesting argument presented in criticism of Breed Specific Legislation is that the law is implemented based on the extent to which a dog fits the breed standard of a banned breed. This raises complaints from the RSPCA and British Veterinary Association for failing to use, for instance, genetic testing to determine whether or not a dog belongs to a particular (banned) breed.
But isn’t this disingenuous? First of all, if it is physical attributes that play a part in how severe a dog bite may turn out to be, then using physical attributes as a guide to which individuals should be restricted seems not unreasonable. Second, I do not think that, were the legislation to be enforced via genetic testing rather than breed standard, the RSPCA, British Veterinary Association or any of the other opposed groups would be any happier with it. Third, it’s not clear what additional information a genetic test will give you beyond the visible phenotypes that are used in the current system to determine breed. From my Crufts-watching days, the judges were looking at how well the dogs matched up against the breed standard. Of course, genes are partially responsible for what dogs end up looking like, but the relevant thing for dog breeds seems to be what they are actually like rather than what their genetic make-up is. The ‘you shouldn’t judge a book by its cover’ claim seems to be a piece of rhetoric rather than a genuine complaint about how the legislation is enforced.
A plea for better communications
As I said, I don’t want to defend Breed Specific Legislation. But it’s hard to evaluate the value of a policy on the basis of communications that are so clearly aimed at achieving a particular outcome (scrapping the legislation) rather than providing informative arguments. At the moment, the arguments made by the RSPCA, the British Veterinary Association and others could apply as equally to wolves as to any other dog breed. If I want to keep a wolf then, extrapolating from the above claims, I could defend myself against accusations I was risking others’ safety by saying “any dog can bite, and no one has shown me evidence that a wolf bite is worse than any other dog”. Indeed, none of the lethal attacks on humans by dogs in the UK in the last 50 years came from wolves. As long as the wolf hasn’t bitten anyone else yet, subjecting it to restrictions would be equivalent to ‘punishing the innocent’. Children just need to treat wolves respectfully and then they should be fine; there should be more of an emphasis on (potential) victims of wolf attacks to avoid putting themselves at risk, rather than restricting people’s freedom to keep wolves. Even though it looks and behaves exactly like a wolf, it’s not fair to say it’s a wolf unless this has been genetically proven.
Is this really the argument the RSPCA, British Veterinary Association and others want to make? It seems to me it would be vastly preferable to present the best evidence in support of their claims, and acknowledge where uncertainty lies so people can form a sensible and well-informed view on Breed Specific Legislation.
Written by Rebecca Brown
This post is based on two recently accepted articles: Brown and de Barra ‘A Taxonomy of Non-Honesty in Public Health Communication’, and de Barra and Brown ‘Public Health Communication Should be More Transparent’.
In a previous post, I discussed some of the requirements for public health institutions to count as ‘honest’. I now want to follow that up to discuss some of the ways in which public health communication seems to fall short of honesty.
I’ll follow Christian Miller’s definition of the term ‘honesty’ to refer to a character trait that involves being disposed, centrally and reliably, to not intentionally distort the facts as one sees them. As discussed previously, it is pretty hard to get at the intentions of agents. We can rarely – if ever – say with certainty that a particular agent intended some outcome, since that information is available only to the agent herself. Instead, we must rely on extrapolating from whatever evidence is available to us.
This can make it hard to prove a failure of honesty, since we must often rely upon a ‘best guess’ as to the agent’s motivation when they acted in a particular way. For the kinds of cases I’m interested in – communications from public health organisations – it is even trickier, as it is not clear what agent is responsible for the communicative act in question (it could be an individual within an organisation, the organisation itself, or some subset of the organisation/individuals within it). But assuming certainty isn’t required, we might still judge that an agent (group or individual) probably intended to mislead or otherwise distort facts if the way they presented information could reliably be expected to create false beliefs in the recipient, and if we would reasonably expect the communicator to have known this. With that in mind, we might speculate about the honesty or otherwise of public health communication.
Public health communication frequently engages in a number of practices that look like they might not meet the demands of honesty. Some examples are:
Magnitude neglect where public health communication doesn’t provide an indication of the expected effect size of the benefit / harm discussed. This could include, for instance, saying that screening reduces your chance of dying from cervical cancer, without telling you by how much your risk of death is reduced.
Harm neglect involves providing information about the benefits of recommended behaviour changes / interventions without mentioning any possible harms.
Relative over absolute risk presentation. Whilst communicators rarely provide quantified effect sizes of the benefits associated with a recommended behaviour, when they do provide such estimates they often use relative risks rather than absolute risks. This will tend to inflate people’s estimate of the size of the benefit they can expect. For example, an intervention that reduces your risk of developing diabetes from 0.001% to 0.0005% cuts your risk in half – that is, reduces it by 50%. The first way of presenting this (using absolute risks) makes the change seem quite small; the second way (in relative risks) makes it seem much bigger. Public health communicators will often use relative risks to describe health benefits, which tend to make them seem much more impressive.
Mismatched reporting takes the above tendency, and pairs it with a tendency to report the harms of recommended behaviours in terms of absolute risks. If relative risks inflate people’s expectations, absolute risks deflate them. Pairing benefits described in relative risks with harms describes in absolute risks seems geared to encourage people to overestimate benefits whilst underestimating harms.
Causation laundering occurs when the content of public health communication implies a causal link between phenomena when it is difficult to know to what extent the relationship is causal. For instance, the health benefits of exercise are widely proclaimed (the NHS describes it as a “miracle cure”. Yet it is very difficult to extrapolate what health benefits associated with exercise are a result of exercise making people healthier, and what results from healthier people doing more exercise. Public health communicators sometimes gloss over such ambiguity, and imply that the causal relationship acts only in one direction – that is, exercise causes good health, rather than the other way around.
These ways of communicating health information, amongst others, seem likely to encourage people to form overoptimistic beliefs about the benefits of various health behaviour changes. Moreover, this seems predictably likely and – we might speculate – intentional. If this is indeed the case – if public health communicators intentionally frame their messages so as to encourage people to form overly optimistic (and inaccurate) beliefs about the likely benefits of health behaviour change – then this is inconsistent with communicating honestly.
This leaves out the question of whether or not there is anything wrong with failing to communicate honestly. Whilst there will almost certainly be cases where honesty is not all-things-considered desirable (e.g. it is better to tell a lie to protect an innocent victim from a murderer than to tell the truth and condemn them to death, contra Kant) it is not clear that public health communication is typically one such case. Indeed, it seems unlikely, given the emphasis on honesty in clinical communication, and the importance of allowing people to make well-informed decisions about how to lead their lives and take care of their health.
Most of the behaviours considered here – relating to diet, exercise, cancer screening and other preventative behaviours – require people to forego pleasures in the pursuit of longer term health. Whilst public health professionals might emphasise that there is “no safe” drinking level, individuals might reasonable decide that the benefits of alcohol consumption outweigh the risks. But people can only make these decisions sensibly if they are accurately informed about the harms and the benefits of these behaviours. And this requires public health communication to give a non-distorted, honest, picture of the available evidence.
References
Brown and de Barra (in press) ‘A Taxonomy of Non-Honesty in Public Health Communication’ Public Health Ethics
de Barra and Brown (in press) ‘Public Health Communication Should be More Transparent’ Nature Human Behaviour
Miller (2021) Honesty: The Philosophy and Psychology of a Neglected Virtue, Oxford University Press
FreshRSS