The two-year sentence imposed on notorious hockey coach and pedophile Graham James in March2012for his serial rapes of two young players has brought into stark relief the issue of leniency in sentencing by Canadian courts. Are Canadian courts lenient when sentencing convicted criminals? Consider the following cases:
The criminal identified only as “P.M.” raped his13-year old daughter approximately ten times over the course of a year. At some point, he began videotaping the assaults, some of the videos depicting his daughter begging him to stop – needless to say, her pleas did not end the vaginal and anal rapes he inflicted. P.M. was sentenced to “six years”, but given the peculiar judicial math which exists in Canada: the actual amount of time he was ordered to spend behind bars was a little over four years – though the amount of time that he will, in fact, spend behind bars will be even less as a result of the statutory release and parole processes. The sentence imposed on P.M. was lenient.
Paul Ferlisi attacked his ex-wife in her home, in front of their young children. In speaking of the crime, various judges describing the assault used words like “horrific” and described the wife as “terrorized”. The sentence imposed on Ferlisi was lenient.
When Justin Nolin was sentenced by Justice Melvyn Green of the Ontario Court of Justice, despite Nolin’s record of violent crime, he was given an extremely lenient sentence.
The Ontario Court of Appeal’s opinion
No need to take my word about the nature of the sentences imposed on the three criminals: their sentences were described as “lenient” (or, in Nolin’s case, “extremely lenient”) by no less than the Ontario Court of Appeal, all within the last eighteen months. Notwithstanding that leniency, however, the appellate court did not see fit to alter the lenient sentences imposed by the trial judges. We can state with certainty, then, that at least some sentences imposed by Canadian courts are lenient, as determined by the courts themselves.
So, next question: Why is that conclusion so contested among Canadian journalists?
Two approaches can be identified in how some Canadian journalists respond to the issue of leniency.
The first, promulgated by John Geddes of Maclean’s, expresses skepticism about whether leniency exists at all. Writing back in September 2011 on the Maclean’s website, Geddes stated he’s seen “no data … exposing a pattern of soft sentences, conditional or otherwise, handed down by Canadian judges for heinous crimes. For the odd case of coddling, there are always appeals.”
In other words, leniency doesn’t really exist, and if it does, the system will take care of it. We can describe Geddes position as systemic deference: trust in the system, says Geddes, because it cannot possibly generate lenient sentences.
But Geddes take on leniency is simply a logical non sequitur: just because there is an appeal tells us nothing about whether the appeal court’s own decision is, itself, lenient.
Even the appellate courts damn themselves with faint….sentences
First, as the cases of P.M., Ferlisi and Nolin confirm, even when appellate judges specifically cite a sentence as lenient, they do not always vary it. So we can state that as an empirical matter, on the justice system’s own terms, lenient sentences are handed down and allowed to stand.
Second, even when an appellate court varies a lower court’s sentence, the new sentence is often itself lenient.
As an example, let’s take the Ontario Court of Appeal’s sentence in the case of R. v Jeffrey B.. The case involved repeated assaults by a “sexual sadist” prone to “extreme acts of violence” against a five-week old baby boy and a six year-old girl. The victims were “horribly assaulted” by “extreme acts of brutality” which were “degrading” and “dehumanizing”. The sentence at trial? Six years. On appeal? Twelve years.
A journalist like Geddes might be tempted to cite this as proof of his contention that sentences subject to appeal by definition cannot be lenient: why, the sentence was doubled!
A more skeptical observer would query whether repeated “extreme acts of brutality” were really properly addressed by a sentence of twelve years (which, of course, would be subject to reduction due to the early release provisions of the Corrections and Conditional Release Act).
Jeffrey B might be out in less than eight years (assuming he is not paroled even earlier). Jeffrey B. was convicted of two counts of aggravated sexual assault and one count of aggravated assault – the aggravated assault conviction could have been punished by up to fourteen years imprisonment and each aggravated sexual assault conviction could have resulted in imprisonment for life.
Or take the case of Cody Paul Lemay: convicted of multiple rapes of a baby (leaving the baby with “the most severe genital injuries” ever seen by the examining doctor, including an “open, gaping laceration” to the baby’s anus),Lemaywas sentenced to five years imprisonment. On appeal, the British Columbia Court of Appeal increased the sentence to… seven years. Our skeptical observer might begin to query whether the appellate courts are all that reliable as a check on lenient sentencing decisions. Again, add in early release provisions, and seven years might be something rather less than that.
The absence of complete data doesn’t negate the obvious observation
A different response to the issue of leniency, advanced by Dan Gardner of the Ottawa Citizen adopts what we might call epistemological skepticism: we don’t have enough data to know whether sentences are lenient.
In response to the Graham James sentence,Gardnerallowed that the sentence might be unjust (he “hesitate[d] to form an opinion without knowing a lot more about the case”), but argued that it was impossible to conclude that the James sentence said something about the system overall. “For the James sentence to say something about criminal justice in general”, arguesGardner, “it can’t be an aberration. It has to be typical. And we don’t know if it’s typical.” Gardnerargues that we don’t know if it is typical becauseCanadadoes not have a comprehensive database of sentencing decisions from which we can draw empirically-substantiated comparisons.
WhileGardner’s point seems to be cogent on first reading, on further inspection it begins to crumble.
Because of the unique nature of legal reasoning, sentencing decisions are, by their nature, indicative of other sentencing decisions. Legal reasoning proceeds under the guidance of a principle known as stare decisis, meaning that judges are obliged to consider and conform their judgments to prior decisions.
Thus, sets of judicial reasons generally make reference to a litany of prior decisions which provide support for the conclusion reached. As an example, Judge Catherine Carlson, in sentencing Graham James, cited at least fifteen other decisions in coming to her conclusions.
When Judge Carlson examined prior cases with some similarities to Graham James’, she found that some courts imposed comparatively lengthier sentences and some imposed comparatively lenient sentences. Taking the previous decisions together as a matrix of “appropriate sentences”, the sentence imposed on James falls comfortably within the pattern established by the prior cases. That in itself tells us something about the pattern of sentencing decisions made by courts: what seems to be a shockingly lenient sentence is enmeshed in a broader pattern of sentencing decisions.
Someone, like myself, advancing the argument that the courts are lenient, provides concrete examples of cases (such as Jeffrey B., or P.M., or Graham James) might hear in response this from Gardner, something like: “Well, yes, the results in those cases might be lenient, but we have no way of knowing whether they resemble other cases out there”. The implication of Gardner’s response is that there may be out there some cases which do not impose lenient sentences.
Okay, we respond, but the court decisions in each of the cited cases themselves make reference to other cases which imposed similar sentences – in other words, similar lenient sentences. Still, responds Gardner, you’re not proving to me that leniency is some kind of systemic problem – you’re just coming up with examples, or indulging in argumentation by “anecdote and impression”.
Lots of leniency—and the missing “harsh” sentences
But then things start looking a little strange: examples of leniency pile up on our side of the table, butGardneradvances no, or few, counter-examples. Gardnerseems to be saying that unless we have information about every single sentence handed down, there might still be some unseen reservoir of harsh sentences which would disprove the assertion of leniency.
It’s the rough equivalent of saying that I can’t conclude whether it’s raining outside my front door, despite seeing ripples in puddles and feeling drops of water on my exposed skin, unless I get a certified report from Environment Canada confirming that a precipitation event has occurred.
Worse, it unjustifiably shifts the onus in the argument: the moving party suddenly bears all the burden of identifying cases of non-leniency, while Gardner has to demonstrate nothing.
Further, it is unclear at what point Gardner will allow us to make the transitive argument. He identifies no sufficient condition after which he will finally concede that a series of lenient decisions indicates a systemic problem. Ten decisions? One hundred? One thousand? Ten thousand? And what number of “harsh” sentences will he use to “disprove” the assertion of leniency? Will one harsh sentence countermand ten lenient ones? In any event, where are all these harsh sentences that Gardner needs in order to make his case?
At this point, someone might question the meaning of “leniency.” Is it not just an entirely relative and subjective notion, with one person’s harshness constituting another person’s leniency?
While there might be some fuzziness around the concept, it also has at least some content: it can’t simply mean “I don’t like the sentence” when it is used by critics of the system just as it is used by judges themselves.
Leniency is not in the subjective eye of the beholder
But what objective set of criteria can be referred to in order to assess leniency? Helpfully, the Criminal Code itself provides an accessible metric: the Code sets out the range of appropriate sentences for many crimes.
By comparing a given sentence to the possible range, we can assess the “leniency” (or lack thereof) of a sentence. (We must set aside for the moment the question of whether the ranges themselves are appropriate.)
Examining the Graham James sentence in this light gives an indication of just how lenient his punishment was. As sentencing Judge Catherine Carlson noted, the range of punishments for sexual assault range from a non-custodial sentence to ten years imprisonment. The two years that James received can thus be seen in its proper light: at the lower end of the range for the crime. Roughly speaking, he got 20% of what he could have been given.
But even that understates just how lenient his sentence was: after all, the Criminal Code allows a maximum ten year sentence per sexual assault offence committed. James was convicted of assaulting not one but two different victims. So, in the harshest possible world, he could have gotten two ten year sentences to be served consecutively—20 years. Instead, he got two years.
Even that, though, does not fully apprehend the situation: after all, James was convicted of dozens of separate assaults against each victim (the prosecution alleged hundreds of separate assaults occurred, and even the defence conceded that the number was in the dozens).
Add up the full accounting of what James could have gotten, and compare it to what he received, and the magnitude of the leniency comes into clearer focus. (Among the many perversities of the Canadian criminal justice system, perhaps none is so grotesque as the bulk discount afforded to those convicted of raping children.)
Absurd laws on the books are beside the point
Ah, Gardner responds, but the Criminal Code actually does not provide a useful standard against which to measure leniency: the ranges in the Criminal Code are a mish-mash of incoherency and contradiction. The offence of selling “crime comics” (a holdover from a brief moral panic in the 1950s inspired by a series of violent crimes committed by young people and supposedly the result of them reading the increasingly gory and explicit comic books of that decade) has a maximum sentence of two years imprisonment. So something which should not even be a crime is potentially punishable by an absurdly long sentence.
Or consider how both sexual assault with a weapon and possessing counterfeit currency can be punished by the same maximum term of fourteen years; the latter crime, while harmful, is wholly unlike the former, so something seems incongruent in the potential punishments.
As Gardner notes, the punishment for breaking and entering ranges from effectively nothing to life in prison (except courts never hand down sentences at the top end of that particular range). But he paints with too broad a brush.
Certainly some offences need to be cleaned out of the Code, and the sentencing ranges for particular offences might benefit from a reassessment. But that says nothing about the particular range for a particular offence.
If the range for sexual assault is inappropriate, then that case needs to be made on its own merits; it certainly is not proven by showing that the range for B&E is so open-ended that it could theoretically result in a sentence of life imprisonment. So when making comparisons of sentences handed down to ranges in the Code, we should definitely interrogate the particular ranges for particular offences – they may prove to be so inappropriate that the exercise becomes meaningless – but it is not axiomatic that the comparison process is itself not illuminating.
While Gardner says something valuable about the need for data collection and coherency of the Criminal Code, he is not making a comment pertinent to the issue of leniency per se.
The “I-cannot-measure-therefore-leniency-cannot-be-proven” defence
As alluded to above, his skeptical critique collapses into something approaching empirical nihilism: we cannot determine whether sentences are lenient unless and until we have a comprehensive database of all sentences. Even then, we will not be able to assess leniency until the Criminal Code is rationalized and amended from top to bottom to ensure that the sentencing ranges it provides are harmonized for each offence. Leniency might exist, in other words, but we won’t be in a position to say for an awfully long time to come.
We can conclude then, that leniency in sentencing does exist – even judges admit it. That conclusion accords with the intuitive and common sense reactions that many Canadians have to notorious sentences (such as that given to Graham James).
So the arguments against the existence of leniency do not withstand scrutiny. But that leaves at least two questions unanswered: why does leniency exist, and what, if anything, can be done about it? More in my next column on both those queries.
Bob Tarantino is an entertainment lawyer in Toronto.
R. v. P.M., 2012 ONCA 162
R. v. Ferlisi, 2011 ONCA 14
R. v. Nolin, 2012 ONCA 183