For whatever this is worth, for those who need to care (or do, regardless of need).
These propositions are written for the Canadian lawyer whose knowledge of the relevant Canadian law is such that a Superior Court (or equivalent) judge would consider that lawyer competent to prosecute or defend an "ordinary" personal injury or property damage action. As such, they presume a certain level of knowledge.
Comments are on for a limited purpose. I will attempt to clarify any of these propositions if the manner in which I have stated the proposition is not sufficiently clear, bearing in mind what I have said is the premise upon which the propositions are written. That means I will not attempt, in this forum, to justify any of the statements. If you disagree, you disagree. Ultimately, none of us, here, have the final say – even in this life.
There is a certain amount of philosophy here. However it is not abstract metaphysical philosophy (whatever that means). It is "not-abstract, physical, philosophy" (my label). I have assumed, for the purposes of this discussion, that if abstract metaphysical philosophy has any valid meaning, even for law, and is a "bad thing", then its opposite – which I have labelled "not-abstract, physical, philosophy" (whatever that means) – is the type of philosophy you will find, here, and is a "good thing".
After all, Clements has acknowledged, in para. 7, that even lawyers and judges working in the area of tort have to know some philosophy.
 Recovery in negligence presupposes a relationship between the plaintiff and defendant based on the existence of a duty of care — a defendant who is at fault and a plaintiff who has been injured by that fault. If the defendant breaches this duty and thereby causes injury to the plaintiff, the law “corrects” the deficiency in the relationship by requiring the defendant to compensate the plaintiff for the injury suffered. This basis for recovery, sometimes referred to as “corrective justice”, assigns liability when the plaintiff and defendant are linked in a correlative relationship of doer and sufferer of the same harm: E. J. Weinrib, The Idea of Private Law (1995), at p. 156.
If the Supreme Court of Canada tells us that we have to understand a bit of philosophy that matters to our understanding of causation in tort law, then, that philosophy cannot be undesirable "abstract metaphysical philosophy".
I hope the propositions are helpful.
1. The but-for test is a factual inquiry.
Clements v. Clements, 2012 SCC 32 at para. 8
2. The but-for test remains the default test for determining factual causation in negligence.
Resurfice Corp. v. Hanke,  1 S.C.R. 333 at paras. 21, 29. 2007 SCC 7; Clements at para. 13.
3. The ordinary but-for test is the only approved method of establishing factual causation on the balance of probability, with two clear exceptions (based on existing case law) and two more which I suggest will have to be acknowledged. The rule in Cook v. Lewis,  S.C.R. 830, and the Walker Estate v. York Finch General Hospital,  1 S.C.R 647, sufficient-condition material contribution test still exist as tests for proof of factual causation on the balance of probability, in the limited circumstances to which these cases apply. "Ordinary" is important, here, because the rule in Cook v. Lewis is an instance of the but-for test.
The two more are this. First: The but-for test, because of how it is defined, cannot be validly applicable to instances of duplicative causation which are treated as instances of duplicative causation. Second: there are instances of factual causation where the negligence is neither necessary (that means it can't satisfy the but-for test) nor sufficient (that means it isn't an instance of duplicative causation) but the negligence is still is a factual cause.
Resurfice and Clements cannot be understood to have overruled the grounds upon which Cook v. Lewis and Walker Estate were decided.
Resurfice does not discuss ratio of Cook v. Lewis. While Clements, at paras. 18-19, and 28, discusses the basis of the majority ratio in Cook v. Lewis, it would be wrong to claim that discussion amounts to a disavowal of that ratio. The obvious reason is that the Court did not say so.
Similarly, Resurfice did not discuss the ratio of Walker Estate. The fact that Resurfice used the facts of Walker Estate as an example of a situation to which the material contribution doctrine might apply cannot be understood to be a disavowal of the ratio of Walker Estate. While Clements, at paras. 24-25, and 28, discusses the ratio of Walker Estate, it would also be wrong to claim this discussion amounts to a disavowal of that ratio. Once again, the Court did not say so.
The rule in Cook v. Lewis is limited to instances of “alternative” causation involving two tortfeasors where there is finding, on the balance of probability, that the injury was caused by the negligence of one or the other of the tortfeasors but not both: see Clements, at paras. 18-19. See Klar, Tort Law, for a detailed explanation of Cook v. Lewis. As such, the instances to which Cook v. Lewis applies cannot be instances to which Resurfice material contribution applies since that doctrine is premised on the possibility the injury might have been caused by both tortfeasors: see Clements, ibid. at paras. 13, 43.
Neither Clements and Resurfice explain what the test is to be for instances of duplicative causation treated as duplicative causation rather than (in effect) treated as an instance of multiple necessary causes (as in the indivisible injury cases). In instances of duplicative causation, there are multiple sufficient causes. The commonly used example is two fires negligently started by different people which merge before the reach a house, which is then destroyed. Each fire was sufficient to destroy the house even if the other had not merged with it. While this example might seem to literally fall within para. 39 of Clements, don't forget Resurfice material contribution is not a test for factual causation etc … (Clements para.14) so that applying material contribution duplicative causation – which is more common than many people realize – would amount to yet another questionable legal fiction.
Similarly, imagine 4 people each negligently put 1/3 of the required dose of a poison in X's teacup, and X drains the cup. X is poisoned and dies. Each person's 1/3 was not necessary. Each person's 1/3 was not sufficient. Again, this example literally falls within para. 39. Did the SCC intend to say there are no individual probable causes, just one collective cause?
Now let's look at why these two examples literally fall within para. 39:
 What then are the cases referring to when they say that it must be “impossible” to prove “but for” causation as a precondition to a material contribution to risk approach? The answer emerges from the facts of the cases that have adopted such an approach. Typically, there are a number of tortfeasors. All are at fault, and one or more has in fact caused the plaintiff’s injury. The plaintiff would not have been injured “but for” their negligence, viewed globally. However, because each can point the finger at the other, it is impossible for the plaintiff to show on a balance of probabilities that any one of them in fact caused her injury. This is the impossibility of which Cook and the multiple employer mesothelioma cases speak.
It is " impossible for the plaintiff to show on a balance of probabilities that any one of them in fact caused her injury, because each can point the finger at the other" only if the meaning of factual causation is defined to be necessity, which is the case so long as the but-for test (as it is currently defined) is the test for factual causation . (Round and round in circles.)
Perhaps there's a window for these exceptions in para. 13 of Clements: " To recap, the basic rule of recovery for negligence is that the plaintiff must establish on a balance of probabilities that the defendant caused the plaintiff’s injury on the 'but for' test." That's on the assumption that the "non-basic" rule the Court had in mind isn't just the Resurfice material contribution doctrine.
4. However, the but-for test is not conclusive on the question of the existence of factual causation.
Resurfice at para. 22.
Resurfice does not explain this proposition. Clements does not deal with it. It should not be understood as just a reference to instances in which there is negligence and factual causation but the evidence is such that that it does not produce a valid, on the balance of probability, answer that identifies the negligence of a particular person (or more than one person) as a probable cause.
The proposition may be understood as a reference to the issue identified by the 3rd and 4th exceptions mentioned in item 3.
The but-for test is not conclusive because it is a negative test which, because it is based on the concept of necessity, cannot identify factual causes which are not based on necessity – what causation philosophy calls “counterfactual dependence”. "Counterfactual dependence" is the causation philosophy label for what Clements called necessity.
So long as it is not understood to have been meant to refer to Resurfice material contribution, the “not conclusive statement” necessarily implies that there is at least one other method of validly identifying factual causation in instances where the but-for test does not provide a “conclusive” answer, so long as there are instances of factual causation where the reason the but-for test fails is not related to the lack of sufficient evidence to allow the valid application of the but-for test. As I've explained, these instances exist, even if they are not discussed in either Resurfice or Clements.
That does not mean that law has to accept any other method as valid for law's purposes. It just means that another method necessarily exists.
5. The general rule is the plaintiff has the burden of proof of establishing, on the balance of probability, the injury would not have occurred but-for the aspect of the conduct of the defendant which makes that conduct negligent.
Clements at para. 8; Resurfice at para. 21.
6. The but for test is based on the proposition that the negligence of a defendant is necessary for the occurrence of the injury.
Clements at para. 8; Athey v. Leonati,  3 S.C.R. 458 at para. 41.2.
7. The but-for test “applies to multi-cause injuries” whether there is a single tortfeasor or multiple tortfeasors.
Resurfice at para. 21; Clements at para. 43 (“ It is important to reaffirm that in the usual case of multiple agents or actors, the traditional ‘but for’ test still applies.”)
8. The but-for test applies to injury which caused by “a number of different negligent acts committed by different actors, each of which is a necessary or ‘but for’ cause of the injury.”
Clements at para. 12.
9. "A trial judge is to take a [robust, pragmatic, and common sense] approach to determining if a plaintiff has established that the defendant’s negligence caused her loss. Scientific proof of causation is not required.
Clements at paras. 9, 46(1). The quotation is from para. 46(1). I have added "common sense" in the first sentence, and changed the word order to make it match the equivalent statement in para. 9. The sentence in para. 46(1), as written, begins "A trial judge is to take a robust and pragmatic approach to determining if ".
The sentences in para. 9 are: "The 'but for' causation test must be applied in a robust common sense fashion. There is no need for scientific evidence of the precise contribution the defendant’s negligence made to the injury.”
The statement in the second sentence from para. 9 should be understood to mean what the second sentence from para. 46(1) means.
The Court could not have intended to assert that scientific proof of causation is not required in any case.
10. “Where “but for” causation is established by inference only, it is open to the defendant to argue or call evidence that the accident would have happened without the defendant’s negligence, i.e. that the negligence was not a necessary cause of the injury, which was, in any event, inevitable.”
Clements at para. 11.
11. When terms such as “substantial connection” or “material contribution” are used to describe conduct amounting, or not amounting, to a but-for cause, the phrases must be understood to relate to the sufficiency of the evidence used by the court to decide that the negligence was a necessary cause.
Clements at paras. 21, 23; Resurfice at para. 23
This time the numbering went to 11. Go figure.
I emphasize the point I made in part 1 of the "Clements Commandments" series. Clements means Resurfice material contribution is irrelevant in most cases. Use the but-for test. Argue it's all common sense (but don't forget to call the required evidence so that you can argue it's common sense and cross your fingers you're not asked to explain what you mean by common sense). Don't waste your client's money or the judge's time arguing Resurfice material contribution in ordinary cases.
If you need judicial authority for that, you've these statements in Clements.
[T] the basic rule of recovery for negligence is that the plaintiff must establish on a balance of probabilities that the defendant caused the plaintiff’s injury on the “but for” test. (para. 13)
The “but for” causation test must be applied in a robust common sense fashion. (para. 9)
A common sense inference of “but for” causation from proof of negligence usually flows without difficulty. (para. 10)
It is important to reaffirm that in the usual case of multiple agents or actors, the traditional “but for” test still applies. (para. 43.)
I happen to think that the common sense mantra is a "vacuous incantation" - or the judicial equivalent of "abracadabra – but I don't make the rules. Still, I'm waiting for the "fun" in Ontario as trial judges attempt to decide whether the evidence the plaintiff relies on satisfies these admonitions of the Ontario Court of Appeal in Aristorenas v. Comcare Health Services (2006) 83 OR (3d) 282 at paras. 60-64 on the use of robust, pragmatic, common sense approach.
 … the “robust and pragmatic” approach modifies the type of evidence as well as the factors that the court may consider. It does not modify the amount of proof required to establish causation.
 Snell and other cases have recognized that it is important to require that the plaintiff provide some sort of evidence (or other considerations) that indicates that the defendant was the cause of the harm suffered. … In Fairchild v. Glenhaven Funeral Services,  3 W.L.R. 89 (H.L.), Lord Rodger of Earlsferry provided this caution about applying the “robust and pragmatic” approach:
[E]ven though it is always for the judge rather than for the expert witness to determine matters of fact, the judge must do so on the basis of the evidence, including the expert evidence. The mere application of “common sense” cannot conjure up a proper basis for inferring that an injury must have been caused in one way rather than another … (para. 150).
 It is, therefore, a misapplication of the “robust and pragmatic” approach to make a finding or draw an inference of causation where no factors of the kind set out in Snell are present and the proper evidentiary foundation is absent.
The "factors of the kind set out Snell" are listed in para. 61. They list was not intended to be a exhaustive definition of the factors.
 The above distinction can be illustrated by applying the “robust and pragmatic” approach to the facts in Snell. In Snell, neither expert was able to express with certainty an opinion as to what caused the harm or when it occurred. As a result, the court used a combination of evidence and other considerations to support a finding of causation on a balance of probabilities:
- First, while experts were not able to testify as to causation on a standard of scientific precision, there was some evidence that the negligent operation had led to the injury.
- Second, the trial judge was satisfied that there was a greatly increased risk of injury because of the negligence.
- Third, there was a finding by the trial judge that “virtually rule[d] out natural causes”.
- Fourth, the defendant was in a better position to observe what occurred. He also was in a better position to interpret what he saw.
- Fifth, the negligent operation resulted in a situation where it was impossible for anyone else to detect the precise cause of the injury.
 Taking these factors into consideration, Sopinka J. found that it was open to infer causation. The approach is “robust and pragmatic” because this type of evidence might not otherwise satisfy the “but for” test.
(Internal footnotes omitted).
I omitted the second sentence of para. 63 of Aristorenas even though it's arguably relevant to the paragraph because I prefer to use it as part of my conclusion. It is:
In Snell, Sopinka J. at p. 299, disapproved of permitting the plaintiff to “simply prove that the defendant created a risk that the injury which occurred would occur."
Appellate courts judges below the SCC don't always have the last word either, even when the SCC doesn't grant leave to appeal.
I hope this helps.
Postscript Added July 22, 2012: What might help is a speech, from last November, 2011, by Professor Russell Brown of the University of Alberta Faculty of Law. To download an mp3 of the speech, go to this site and look for the link at the bottom of the page.