In this, since it is in reference to what would have happened If the defendant’s negligent action would not have occurred it is in part a counterfactual claim based on dependency (Lewis, 1973, pp. 556-558). It is not a claim about what happened in actuality but more so what would have happened in a different scenario where the negligent action did not occur.

Although the “but for” test if passed, means that the negligence was necessary in a legal context, it also allows for this to be proven on a balance of probabilities ( Clements v. Clements, 2012 , para 8). In this, a plaintiff must show based on a robust common sense approach to the “but for” test that more likely than not, if this negligent action did not occur then this injury would not have occurred (Clements v. Clements, 2012 , para 23). With this in mind, scientific proof does not need to be established to prove causation in a legal context (Clements v. Clements, 2012 , para 38).

The “but for” test is the primary rule for recovery; however, in exceptional cases the “but for” test can be substituted by a material contribution to risk assessment when it is “impossible” for causation to be proven using the “but for” test ( Clements v. Clements, 2012 , para 27).

The material contribution to risk of injury assessment is a test that, “imposes liability not because the evidence establishes that the defendant’s act caused the injury, but because the act contributed to the risk that injury would occur.” ( Clements v. Clements, 2012 , para 15).

In this, if a material contribution test is permitted, a plaintiff does not need to prove causation or counterfactual dependency, but rather merely needs to prove that the defendant’s negligence contributed to the risk of the injury to succeed. 

However, in this the Court has noted that moving from a “but for” test to a material contribution assessment is not merely moving to a different analysis of counterfactual causation, but an elimination of causation entirely, at least in a counterfactual understanding (Clements v. Clements, 2012 , para 16).

As such, since forgoing causation in a case of negligence is a step that goes against the foundational conventions of negligence and as such should be used rarely and only when necessary (Clements v. Clements, 2012 , para 16).

 

There has been some disputes regarding if it is more prudent to interpret this as, “material contribution to injury” or, “material contribution to risk” as a substitution for “but for”, or if there is even a practical difference at all, as seen in Snell v. Farrell (1990).

In Clements, the driver of a motorcycle was riding with his wife who was a passenger. In this, they were riding in wet conditions and unbeknownst to the driver a nail had punctured the rear tire. However, the bike was also overloaded by 100 pounds and the driver went past the speed limit, driving 120km/h in a 100km/h zone when attempting to overtake a car. Subsequently the nail fell out of the tire, causing the rear tire to deflate, making the bike unstable, leading to a crash. This crash lead to the passenger suffering injuries and suing the driver on the basis that his negligence caused her injury (Clements v. Clements, 2012 , para s 1-3).

The trial judge found that the “but for” test was unable to prove causation due to limitations of scientific reconstruction. As such, a material contribution assessment was used instead and found the driver, Joseph Clements liable on this basis (Clements v. Clements, 2012 , para 3).

Concerning the courts’ analysis of this concept, this entry will mainly focus on considerations made by the Supreme Court of Canada in Clements, including different types of impossibility and its analysis of other cases in relation to impossibility and “but for”.

In Clements, ideas of impossibility considered included logical impossibility, scientific impossibility, and impossibility due to multiple tortfeasors.

In Clements, the Court considered impossible to prove via logical impossibility or conceptual impossibility (Clements v. Clements, 2012 , para 36). This often refers to the notion that something is logically impossible if it either contains or entails a contradiction, or if it is inconceivable . However, that which is often “inconceivable” in this context is merely that which is contradictory.

However, the Court noted that implementing this as the understood definition of impossibility in this context could lead to further confusion. In many cases where it appeared that causation could not be proved using “but for”, the difficulty was not based around whether proving a specific plaintiff’s negligence was contradictory but that there was factual difficulty connecting a specific negligent action to the injury.

For example, In Cook v. Lewis, (1951) proving which of the tortfeasors’ negligent action was responsible for the injury suffered by the plaintiff is not inconceivable, nor is it contradictory. Nevertheless, based on the facts present at the time, finding out which of the defendants negligence was the one which caused the injury was highly implausible, especially since either one of the tortfeasors could “point the finger” at the other regarding responsibility. Furthermore, if logical impossibility became the standard definition of impossibility, and subsequently the marker of whether the “but for” test should be replaced, there could be cases where since finding proof is conceivable and not contradictory, “but for” should not be replaced; however, proving causation on a balance of probabilities would be so implausible that it it couldn’t be proven and as such no recovery, but in these cases it was clear that some negligence caused the injury and as such it would make it so that there was no recovery for a case of negligence which caused injury.

As such, implementing notions of logical impossibility in this context is often irrelevant or insufficient to the difficulties in proving causation. As such, the Court found that this should not be the definition of “impossibility” used when determining when the “but for” test should be replaced by a material contribution assessment.

Another form of impossibility that was considered was “scientific impossibility”, which in reference to the Courts interpretation in Clements was in regard to notions where based on the facts and current scientific developments, scientific proof regarding causation cannot be established.

However, the Courts re-stated that the law of negligence does not require scientific proof to determine causality via the “but for” test, where a robust common sense inference based on the facts can be sufficient.

As such, since scientific proof is not required to determine causation using the “but for” test, it appears difficult to posit that causation regarding a particular case being impossible to prove scientifically entails that this causation cannot be evaluated by the “but for” test.

In Clements, the Court found that after considering various definitions of impossibility, the circumstances in which it appears permissible to consider it “impossible” to prove using the “but for” test are those which have multiple tortfeasors. In this, multiple people committed negligent acts and one or more of them caused the injury to the plaintiff; however, based on the facts one cannot show on a balance of probabilities that any particular one of the tortfeasors negligence caused that injury. This is due to defendants being able to note that based on the facts their negligence is not more likely than not the cause of the injury than that of another defendant’s negligence. In this, if two tortfeasors committed negligent acts where one of them caused injury, it is hard to determine that one was 51% or more likely to have caused this injury than that of the other based on the facts, but rather at most 50% likely each. As such, since determining causation on a balance of probabilities requires a claim to be more than 50% true, a plaintiff cannot prove on a balance of probabilities that either plaintiff’s negligence caused the injury and recovery is denied. However, in this it is still clear that some negligence occurred which caused the injury, which goes against the goals of negligence for the Court. 

As such in cases like this, the Court determined that a material contribution assessment can replace the traditional “but for” test. However, it is clear in these cases that “but for” has still been proven globally in the sense that “but for” the negligent actions of the group the injury would not have occurred. As such, in cases where a material contribution assessment ought to be permitted by the court “but for” has at least been proven globally. 

Consequently, In Clements, the Court concluded that a plaintiff cannot succeed when causation cannot be proven by the “but for” test at least globally.

Now although interpreting logical impossibility in terms of contradiction and inconceivability much like that of the Court in Clements is correct, it is an ontological claim. In this, it is a statement about what circumstances could exist; however, as previously analyzed, this leads to unresolved difficulties in proving causation. Beyond what is considered logically impossible being too large in scope, what is being analyzed isn’t necessarily what exists, but rather what we can know and prove. As such, since what is being analyzed is if we could know if causation was possible using the “but for” test, this is perhaps better fit as a standard of epistemology. 

As such, this entry seeks to develop ideas of epistemology in regard to this condition of impossibility. Furthermore, to keep it in line with a structured and deductive approach similar to that of logical impossibility, this entry seeks to posit these in epistemic logic terms. 

Now, in regard to what one could know, if one considers logical impossibility, then this leads to the scope being too inclusive, where again the truth value of this is based on ontological impossibility. Rather, this entry notes how although this is too large, to simply move to a type of factual impossibility based on two or more tortfeasors feels ad hoc.

One of the big concerns regarding this conclusion is that it appears less so based on an abstract rule, but more so based on a particular example. In this, although it is clear that it is impossible, and to an extent why, it’s hard to clearly articulate in more abstract terms why as a rule this is true. What properties does this set of circumstances possess which makes it “impossible” 

Partially why this is ambiguous is due to it so far being a case of a particular type of factual impossibility; however, as noted earlier, not all instances of factual impossibility entail that it is impossible to find causation using the “but for” test in that instance. It should be noted that the Court in Clements did add onto this noting that it is impossible to find causation on a balance of probabilities and that global “but for” had been established. 

Regarding how global but for could be established but for a specific negligent act is unclear, it’s unclear if this is a requirement for it to be “impossible” for the but for test. Nevertheless, on an ontological sense it seems implausible that being established on a global sense means that it is impossible to prove causation using “but for”, rather it appears to aid it. 

In this, although it doesn’t in itself serve to be an indicator of impossibility of proof, beyond the Court’s findings it appears this accomplishes two things at least:

1. By providing global “but for” some causation has been proven. In this it helps with the distinction between when causation cannot be proven and when causation cannot be proven using the “but for test”. Concerning this, the latter is more so an evaluation of the test rather than causation itself. As such, when global “but for” is proven it appears that it is more so a problem with the test than causation and as such it appears more plausible to allow for a different test.  

2. Additionally by making global “but for” a requirement before moving to a material contribution test, it avoids problems where a plaintiff simply cannot succeed in proving but for. Rather by enforcing global but for there are no cases where proof of causation is entirely forfeit, sustaining the fundamental principles of negligence.
   

The real questions appears to be what counterfactual circumstances make it impossible to determine guilt. A large problem with this is that it assumes that there are instances with the test of causation and causation itself where its possible to prove one and not the other. However, it appears that in this case causation is defined as a causal relationship of counterfactual dependency. In this, by definition it appears that if one is true than the other is false. As such it appears clear that if it is impossible to determine causation via the “but for” test, it appears impossible to determine causation. 

So in this, if causation is not only determined, but defined by the “but for” test, how can there be a situation where determining causation is possible but not by using the “but for” test?  

A big thing to note here is that when it is stated that “but for” is being replaced, due to “but for” also serving as the definition of causation when it is replaced there isn’t another test to prove causation but merely that causation has now been eliminated

Well as mentioned by the Court there are different things inquiries that can be made about causation 

1. What caused it
2. Did negligence cause it
3. What negligence caused it?

I suppose the big question is what part of two tortfeasors was the problem? Was it simply the fact that there were two? Was it the balance of probabilities? Was it the factual impossibility? Could there be another case very similar to it that doesn’t have multiple tortfeasors but is equally impossible? 

In regard to underdetermination as a concept, it more or less posits/means that based on the evidence available, it is insufficient to tell which conclusion to believe in. 

More specifically, In circumstances of undetermination, it requires that there exist multiple incompatible conclusions which are both based on all the same evidence and are equally supported based on this evidence. 

As such, since they are both based on all the same evidence, there is no evidence to believe one conclusion over the other, serving as an epistemic problem. 

Now, theories of underdetermination often revolve around how there isn’t enough to determine that a specific conclusion is true, which although serves as an obstacle to establishing scientific proof, is not necessarily an obstacle to proving “but for”. This is due to “but for” being carried out in a robust common sense manner based on a balance of probabilities. 

As such there are many potential cases that are undetermined; however, if using common sense and a balance of probabilities it is clear that one conclusion is more plausible than another. 

However, there are still cases where even if common sense is used, and it is set on a balance of probabilities, causation cannot still be determined under these standards due to underdetermination. 

Although the Court in Clements determined that in cases where “but for” could be replaced by a Material contribution to risk, this is when there are two or more tortfeasors. However, it appears unclear why specifically this condition of having two or more tortfeasors is required. Rather it appears to be that cases such as these have conclusions that are undetermined even under a common “but for” approach on a balance of probabilities. 

This is because it is not simply that there are two tortfeasors, as the Court noted that this condition is not sufficient, but rather than based on the evidence, the two rival conclusions that could be drawn were based on all the same evidence where there is no additional evidence to support believing one over the other. Additionally it is not the case that one of the conclusions is clearly more plausible than the other and as such a balance of probabilities. As such it appears plausible that cases where this is true are often those that have multiple tortfeasors, perhaps implementing notions of underdetermination rather than the notion of two tortfeasors would entail a more exhaustive account of the conditions required to replace the “but for” test.

Nevertheless, If it is not a better implementation of one of the conditions required for the replacement of the “but for” test, perhaps understanding this condition in terms of underdetermination could be benefits regarding further clarity as to why this serves as an obstacle to but for.