Kinds and Classification: Why the Gun Control and Canine Profiling Analogy Breaks Down

Though this post is partly in response to comments on my previous post concerning breed specific legislation from  Mike Steiner, a fellow APT contributor, this is now, in effect, also a response to Yvevs Boisvert’s post for the Globe and Mail.

Now is a timely moment to discuss the analogy between pit bulls and guns given that Yves Boisvert’s post for the Globe and Mail was just released entitled, “Why Montreal’s ban on pit bulls is a reasonable restraint.”  What follows is a rather lengthy post.  It combines the sort of philosophical work I do with a matter of public safety that directly impacts my personal life and where I can live as a Canadian working in academia.  And so, I have been thinking about this for while.  If you might be so kind, let me take you through my thought process on this matter—sit back, pour a glass of wine or a cup of tea, and if you’re still with me at the end I promise it will be worth it.  Well maybe I can’t promise that. But I can hope that you’ll stay with me as a I bumble through the breakdown of the gun-pit bull analogy. And I will provide a moral to the story, as every good story deserves.

SIDEBAR: I was initially concerned by Boisvert’s following leading comment: The discovery concerning the dog who mauled a Montreal woman in June turned out to be registered as a boxer, was really a discovery to “mislead authorities,” he conjectured.  ‘Authorities’ in this case are people working for animal control—the very organization that boxer was apparently registered to, so if they’re trying to mislead themselves…well I’m not sure how that would help them with their aim to ban pit bulls, or their role in the city’s aim to administer the ban, given that as animal control workers they are working for the city. I can’t actually identify any arguments of note in his article. But I digress.  The claim in support of a gun-pit bull analogy is popular, so here it goes–I’ll make the argument for him.

Arguments used to justify canine profiling, which rest on an analogy between pit bull type breeds and restricted weapons, are worth some reflection. Let’s dig in.

Mike Steiner responded to my last post about breed specific legislation with a thoughtful comment concerning gun control and the alleged disproportional strength of pit bulls compared to other breeds.  The case to consider is this:

Imagine a German Shepherd is running towards you on the street. It is approaching fast. No owner in sight. You do not know this dog, where it comes from, and you’re with your kid.  Should you be less scared than if the dog were a pit bull?  That is, should you be even slightly relieved that the dog is a German Shepherd rather than a pit bull type dog?

Another way to put the question: What sorts of predictions about the dog’s behaviour should you make based on the breed that particular dog is classified as?

One response, that may be a driving factor behind Boisvert’s Opinion, is that based on what we know about pit bull type breeds–a certain “class” or “type” of dog–we should err on the side of caution.  This is similar to how we should err on the side of caution when in the presence of a Colt CM901 Military Rifle, which is a fully automatic weapon that unloads multiple bullets upon the squeeze of a trigger.  Fully automatic weapons are of a certain type of gun, which are prohibited in Canada (unless you are military personnel) to mitigate possible harm. Alternatively, we may not be so cautious in the presence a non-restricted, semi-automatic carbine rifle that unloads only one bullet upon a trigger squeeze.  The idea is that we can infer, based on what we know about weapons that belong in the category of automatic, that the increased rate of fire from fully automatic weapons decreases the chance of escape.  This is why we can say you should feel more threatened in the presence of automatic weapons than in the presence of semi-automatic weapons. Okay so far, so good.

Now picture yourself back on that street.  The German Shepherd is approaching quickly. If we follow the analogy with canine profiling through, then you should feel less threatened than if that dog was of the pit bull type.

The above analogy between canine profiling and gun control is an argument often used in favour of breed specific legislation.  It goes something like this: “We restrict certain guns because the potential danger is greater, therefore, we should restrict certain breeds because the potential for more serious bites (i.e. danger) is greater.”

My first reaction is always to point out how the analogy between guns and dogs breaks down by focusing on an obvious dissimilarity between dogs and guns–dogs are not inanimate objects.  They have individual personalities, internal mental states, and exhibit behaviour that is suggestive of a rich internal mental life generally.  This fact necessarily translates animal control issues to animal rights and welfare issues or at least it calls for balancing public safety against those matters.  This of course assumes that we all agree that animal lives matter, at least to some degree.  In this sense, guns which are restricted or prohibited experience no loss, but the dogs do.  Dogs often lose their lives over such restrictions.  And if dogs’ lives matter, then how such restrictions affect them should be taken into account by considering ways in which we may reduce unnecessary harm to them, while at the same time increasing public safety.  Therefore, although issues surrounding both guns and dogs can include public safety concerns, the similarities simply stop there. The focus of our concern in this case, i.e. dogs, should heavily inform how we devise policies, if we truly want our society to be of the character those famous Gandhi quotes recommend.  However, not everyone is convinced by an argument of this type. So let’s take a closer look at what it means to sort objects (be it dogs or guns) into different categories or classifications and go from there.

Classes or kinds of things are typically construed as sets of similar things.  There might be an essential property shared by all members of that set, such that all and only members of that set have that property.  For example, we can identify which lumps of stuff belong to a particular class of chemical elements, such as copper, by the properties that lump of stuff exhibits.  In order to be a piece of copper, that object must have, among other properties, an atomic number 29, a particular mass, a particular electron configuration, and so forth.  Pieces of copper are identified by essential properties, which make that class of chemical element the class that it is. Additionally, based on what we know about the class of copper, we can predict that this particular piece of copper will conduct electricity and melt at 1,085’C (are you still with me?  I promise there’s a point here).   And so, when we’re trying to sort objects into classes, such as types of dogs or types of guns, we do so by trying to identify properties they all share in order to group them together.

A reason why we might group objects together is so, upon coming across a particular member of that group, we can draw inferences about what that thing will do, how it might behave, impact our safety, etc.  In fact, up until the work of David Hull in the 1970s, it was common to believe that species are classes, that is, species are classes with organisms as their members. And in order for an organism to be a member of a certain species, that organism must have the essential properties that make that species what it is.  David Hull, famously, viewed this endeavor as primarily exclusionary—trying to find the essential properties that make an organism human, for example, almost always drew the boundaries far too narrow for comfort.  Today we pay special attention to the diversity of persons and aim to consider each individual according to their special histories and experiences.  We know all to well the dangers of painting certain groups with a broad brush.

Similarly, trying to identify essential characteristics of certain dog breeds, such as pit bull types, that justify the inference that those breeds are more dangerous, is just the same as trying to identifying essential characteristics of certain human races that allegedly make all members of the targeted race inherently more dangerous.  There is no doubt that singling out particular populations of a species in this manner is seriously worrisome from an ethical standpoint, if we indeed acknowledge the individual conscious experiences of humans and non-human animals alike.

But its seems we can probably all agree on the animal-lives-matter stuff.  Rather, the gun analogy is supposed to support identification of which dogs are more dangerous just as how we identify which particular guns are more dangerous by their membership to a particular class of guns.

However, classifying dogs into the kinds of dogs that one can infer a greater danger potential does NOT result in the clean classifications one might get when classifying guns with greater danger potential.  Let’s return to our automatic weapon example.  Some guns belong to the kinds of guns with mechanisms that permit multiple bullets to fly so long as the trigger is pressed.  These kinds of guns are fully automatic weapons.  Canada prohibits all guns in this class likely due to the threat on public safety they pose.  Kinds of guns which are semi-automatics, namely, all guns that require only one trigger-press for each bullet with minimal re-loading, can fall into either non-restricted or restricted categories.  Regardless, they are perhaps less dangerous than automatics, but some are still dangerous enough to call for control.  From my understanding, the kinds of guns that are not restricted (but still require licenses) are ones like long rifles and shot guns that call for more effort and time to re-load, and thus a targeted person has more time to get away from a shooter.  Semi-automatic and fully automatic classifications are distinguished by the differences in causal mechanisms responsible for how fast the bullets are released and the ways in which those bullets must be loaded.  In summary, the class of automatic weapons is defined by essential causal mechanisms that help us to identify whether a particular weapon belongs to that class.  We also devise policies concerning that class because of the dangers its members pose to public safety.

Is there a way to identify classes of dogs such that ‘pit bull’ denotes a kind of dog with some essential property, which makes it a member of that kind?  If we could do that, then we could predict the likelihood that this or that dog will bite and cause serious damage.  To do so, one would appeal to how a particular dog belongs to the kind(s) of breeds that bite and cause serious damage. But how do we discern those kinds of breeds?  And are we justified in claiming that those kinds of breeds are pit bulls?  No.  And here’s why.

One might say that what distinguishes pit bulls from non-pit bulls is their strength.  But this property is not a property of all and only pit bulls.  We know that ‘pit bull’ is a term that denotes multiple different breeds of dogs.  See the “find the pit bull” test as an example.  Instead of the German Shepherd, picture another dog running down the street towards you with short fur and a blocky stature—that dog may very well not be a true American Pit Bull Terrier.  Rottweilers and Mastiffs also share these properties.  Perhaps the property of concern here is size then? But many Staffordshire terriers (grouped under the ‘pit bull’ umbrella) are very small and referred to as “pocket pitties” weighing under 60lbs.  A 45lb staffy will not have the strength associated with the body mass of an 85 pound German Shepherd.  But perhaps by strong we mean ‘stronger bite’ and so more danger of serious damage.  So the property grouping together all dogs in the pit bull category is that they ‘have a stronger bite’ than dogs of other breeds.  Unfortunately, however, that claim lacks scientific evidence.

A study quoted by the AVMA finds no evidence that pit bull types breeds are disproportionately dangerous compared to other dogs.  That is, there are no (dangerous) properties of pit bulls that all and only pit bulls have, which includes bite potential.  This was a conclusion drawn from studies that controlled against statistical biases the public usually doesn’t notice.  It should be of no surprise that the most bites on record are from the most popular type of dog.  For example, the AKC noticed a distinct peak in registering Rottweilers that corresponding with Rottweilers making it to the top of the biting breeds list in the late 90s.  If we’re interesting in typing breeds most likely to bite, then we should be interested in something like ‘bites per breed’ similar to ‘murders per capita’—a smaller city may very well be more dangerous than a larger city even though the larger city has a greater number of murders.  The point is that the sheer numbers of bites from a certain breed are not statistically relevant if that breed makes up most of the dog population. However, one might worry about concerns with bite likelihood—many of the dogs in the AVMA study identified as most likely to bite are small breeds.  So this really does not get at the heart of the issue.  Perhaps we should be worried about bite force, i.e. how bad the bite can be if the dog bites, rather than bite likelihood.  The concern, then, is about bite-damage capacity.  For the measurement of bite force in terms of pounds per square inch (psi), as far as I can find, has only been addressed in in one recent 2008 case by Dr. Brady Barr from National Geographic.  This guy is a character.  He goes around measuring bite forces of different animals, such as wolves and crocs.  How he is still alive, I’m unsure. Conveniently, he measured the bite force of 3 dogs: a German Shepherd, a Rottweiler, and one of the terrier breeds referred to as a ‘pit bull’ (though it’s unclear whether it’s an Amstaff or and APBT). Check out this video: http://natgeotv.com/asia/webisodes/web-shows/dangerous-encounters

Notice that the terrier scored the least bite force with a rather pitiful (no pun intended!) first attempt.  Based on all of this, returning to that moment on the street with a dog charging towards you, I’m not sure that you should be less concerned that a German Shepherd is after you compared to other breeds. All domestic dogs generally have an average bite force of around 300 psi.  Singling out pit bulls by appeal to the shared property of bite force makes no sense when it is a property of (generally) all good-sized dogs. It gets even worse when we notice that the German Shepherd is representative for the breed German Shepherd (singular breed) and not for all shepherd type dogs (multiple breeds), whereas the pit bull is of an unidentifiable breed and is representative of the numerous breeds included in the pit bull category!

So how do we classify dogs such that we gain predictive power and thus control over decreasing dog bites?  It seems like the breeds do not sort into clean kinds—there are no essential properties of pit bulls that make them more dangerous than all other dogs.  And so, identifying groups of dogs as pit bulls does not help us gain predictive control for bite-prevention policies.  Maybe identifying a rough class of dogs, such as restricting ownership (or as I prefer, “guardianship”) for a class of all powerful breeds, could get the job done.  This could prevent backyard breeding practices in addition to other irresponsible people from owning dogs. Such practice is already underway: The Calgary Humane Society requires educational classes and training included in the adoption fees for large, strong dogs to ensure the owner is set up for success.

And so, while you are alone on that street wondering whether you are safe in the presence of that fast approaching dog, the problem is NOT its breed, but the fact that the dog is at large, not under control, and without supervision.  Education and preventative measures in screening potential adopters (e.g. adoption fees, enforcing existing by-laws  for dangerous dogs, increasing the number of dogs registered with the city to keep track, etc) is the only way to decrease dog bites without committing serious offenses against animal welfare, such as killing all members of one breed when no offending act has been committed.  It’s also the only way not to commit serious offenses against the reach of our inferences and predictions that can justifiably be drawn from certain categories of objects.

So that’s it.  If you’re still with me, then I should probably at least provide the moral to this story that I promised, especially if that glass of wine you poured turned into a few more. What I think is an incredible outcome of the gun analogy is this: we should be controlling who gets the privilege and responsibility of caring for dogs by screening potential adopters in terms of the knowledge and experience needed to successfully be a responsible dog guardian for dogs larger than a Chihuahua (okay but also for Chihuahuas, let’s be real here).  With more power comes more responsibility, which is a motto that should certainly not be ignored.  But this concession in no way singles out a particular breed of dog.

-Alison