Dear colleagues:

 

It is an especially sunny day in Canada’s sunniest city, so it seems a shame to have to be troubled about factual causation at such a time. But Jason’s question is important, as are the responses from those who have been following and have written about these questions for quite some time. In light of their responses (which seem to have ended up in two email threads), I’ll try to keep my comments to a few bullet points:

 

• Whatever might be the merits of hanging Sindell-style liability (aka “market share liability” or “proportionate liability”) on Clements, plaintiffs don’t need to. They can at least withstand a pleadings attack based on Gariepy v Shell Oil, 51 OR (3d) 181, 2000 CanLII 22706. This was a classic “fungible goods” case, in which several manufacturers made polybutylene resin pellets that other manufacturers in turn used to make allegedly defective polybutylene plumbing. At para 11, Cumming J, citing Sindell, accepted that “where the specific manufacturer of a defective product used by a particular class member is unknown but the product by different manufacturers is the same, then liability may be able to be determined on a market share theory.” See also, and more recently, Cuillerier v André’s Furnace, 2011 ONSC 5310 at para 29.

 

• One of the consistent things that the SCC has done in this area—perhaps the one consistent thing—is to focus on the case before it, and to some extent the preceding ones. And generally to the exclusion of theory or with an eye to the kinds of cases that are yet to come. This is evident not only from the excerpt that David has pointed us to—para 46 of Clements—but to its natural parent, Resurfice Corp. v Hanke, [2007] 1 SCR 333, 2007 SCC 7. I’m thinking especially of paras 24–28 of that decision, in which we find the Court relying on Cook v Lewis and Walker Estate v York Finch General Hospital for the two branches of the “impossibility of proof” rationale that David has cited. Constructing a general theory from two cases is problematic; unsurprisingly, it resulted in some backpedalling in Clements.

 

• As usual, I find I can turn to my students for evidence of the problems. On final exams, I usually include a scenario that has either a Sindell-style component or a variation on Cook v Lewis. Sometimes, the Sindell-style liability will be confounded with other causes, whether tortious or non-tortious. The most sophisticated answer I’ve ever had argued for the Sindell-style tortfeasors being severally liable as among themselves, but jointly and severally liable as a group with the other “ordinary” tortfeasors. (!) And plenty of unsophisticated answers too, but things seemed reasonably OK until Clements came along. Now I’m getting answers that suggest Clements has opened up all sorts of possibilities to find defendants jointly and severally liable. Consider a short-answer question, one of several that I might pose with a request to say whether you agree, disagree, or partly agree/disagree with the proposition:

 

Dave, Dana and Darin shoot in Paula’s direction, with one of them—we don’t know who—injuring her. Dana, who has the most money, could have to pay for all of Paula’s damages.

 

• Typically, students will say that this is like Cook v Lewis, and that the 3Ds would be jointly and severally liable unless they could prove lack of intent or negligence; a passing answer should also explain that “jointly and severally liable” means that Dana could indeed have to pay the whole judgment, but would be able to seek contribution for any amount that she pays over her one-third several share. A more sophisticated answer would pick up on a class discussion about the fairness of this approach to defendants; a student might explain how, with two defendants, the onus reversal didn’t really deviate that much from a “balance of probabilities,” since there is a 50% likelihood that one of the defendants wasn’t the shooter; once you started adding defendants, though, things shifted: with three defendants, there’s a 1/3 chance of being correct, with 10 defendants, a 10% chance, and so on. Such answers would at least raise a concern about joint and several liability in the circumstances and argue that there might be a case for several-only liability instead. The point I want to make is that since Clements, such discussions have entirely evaporated: students who recognize the difference in the fact pattern point to the “two or more tortfeasors” language in the case and flatly say something like: “Since Clements this has not posed a problem.” This I find rather alarming and unthinking, and drives me to support the kind of arguments that David and Richard are making.