A more recent version of these Causation notes – written by Oxford students – is available here.
The following is a more accessble plain text extract of the PDF sample above, taken from our Tort Law Notes. Due to the challenges of extracting text from PDFs, it will have odd formatting:
4.1 CAUSATION (CAUSE IN FACT) Causation links actionable damage and breach of duty. It is predicated on a corrective justice model; D's interference with the life of C resulted in loss. Causation justifies the use of negligence at all --- otherwise we might as well use a social justice model.
? Honore: "to insist on a causal connection between conduct and harm ensures that in general we impose liability only on those who, by intervening in the world, have changed the course of events for the worse." It is based on the balance of probability, splitting the risk of error between C and D. Once D is proved as a 50%+ cause, this is treated as legal certainty (i.e. it does not make deductions in the damage recoverable because there was only a 51% likelihood that D's act caused C's harm) . Thus causation is invariably a question of probability, not certainty. BUT FOR CAUSATION The orthodox model is 'but for' causation --- but for the actions of D, the injury that C suffered would not have occurred.
? Barnett v Chelsea Hospital : Doctor negligently told C (poisoned by arsenic) to go home. C died, but D was not liable because C would have died anyway, even if D had examined him property, but for D's negligence, C would still have died. Bolitho gloss: D cannot argue that his negligence was not a 'but for' cause because, even if he had taken the correct course of action, he would have committed another subsequent breach.
? Bolitho  D (doctor) failed to respond to a call to attend to C. D argued that, even if he had attended, he would have committed a further breach (failing to intubate), so C's harm would have occurred anyway. HL: rejected his argument. He could be held liable. Note that D is not necessarily liable for all loss flowing from his actions if a 'but for' cause is established --- remoteness rules still apply (see SAMMCO). D CAUSED PART OF C'S INJURY This can only be applied where the injury can be divided between separate factors.
?????Not a part contribution to a single injury, but causing a separate injury (part of overall injury). The test here is still 'but for' --- need to establish, on the balance of probabilities, that injury would not have occurred but for the tort.
? Performance Cars v Abraham : C's car was involved in two collisions. First collision damaged the underside of the car and C had been unable recover. Second collision: D damaged the front wing. QB: D only liable for the damage he had caused, not for the damage sustained in the first collision --- D was not a 'but for' cause of the damage to the underside. To apply Performance Cars, need to work out if the injury/damage is divisible.
?Divisible: an injury which exists on a spectrum and can get worse over time (e.g. deafness
--- more exposure to noise can make C deafer; if D1 exposes C for 5 years and D2 for 10 years, each will be liable for their 'but for' contribution --- D1 1/3 and D2 2/3.) Indivisible: a binary injury --- e.g. death / coma; everyone who has the condition is the same. If there is a mild / serious form of injury, then it is divisible and Performance Cars applies.
CONSECUTIVE / SUPERVENING INJURIES These rules concern a contribution to a single indivisible injury. E.g. C suffers one injury from D1 (e.g. wound to a leg), but a supervening injury from D2 (e.g. loss of the same leg) occurs. Separate and unconnected torts: where C suffers two consecutive separate tortious injuries, and D1 is before the court but D2 is not, then D1 will be liable for the damages caused by him as if the second tort had not occurred --- his liability will not be reduced by the supervening harm of the second tort.
? Baker v Willoughby :C was hit by D1s car, leaving his left leg useless. D2 (a robber) shot C in the left leg, causing the leg to be amputated. C sued D1 (could not locate the robber). HL: D1 was liable to pay damages for the fact that C was unable to use his left leg for the rest of his life. D1's argument that he should only be liable up until the time C lost his leg was rejected. D1 should not receive a windfall merely because C was unlucky in sustaining further injury.
? Lord Pearson: "The original accident... produced a general reduction of C's capacity to do things, to earn money and to enjoy life. For that devaluation D1 should be and remain responsible to the full extent" unless before damages are assessed, something happens which "diminishes the devaluation (e.g. if there is an unexpected recovery)." If the "supervening event is a tort, D2 should be responsible for the additional devaluation caused by him." L+O: if D1's argument was followed, there would be a gap in damages --- D1 would pay for the injury up until the time of the subsequent tort, but D2 would only be liable for the additional damage caused (in Baker this would not have been substantial, not much difference between no leg and a useless leg). I.e. C would get no compensation for the effects of the original injury after second tort. NB: Baker only applies where D2 is not before the court --- if D2 had been before the court, Performance Cars would have applied. D would only have been liable for the injury up to the robbery, the robber would have been liable for the rest. Second injury is not tortious: where the second injury is not tortious (e.g. a naturally occurring misfortune) then D will only be liable for the injury until it was overtaken by the supervening injury.
? Jobling v Associated Dairies : D's negligence caused C an injury to his back. Before trial, C independently developed another back injury (not connected to the injury). HL: D was only liable for the back injury up until the time of the supervening non-tortious harm. Reasoning is based on the 'vicissitudes principle': damages should not place C in a better position than she would have been in if D's tort had not occurred. When assessing damages, the
court will speculate on the effects of injury and make reductions based on the likelihood that an illness would eventually cause the same effect anyway (e.g. if injury prevents C working, but a medical condition would likely force her to stop in 10 years anyway, a reduction will be made). Where a vicissitude is known to the court before trial, there is no need to speculate --- the court can make an exact deduction. The vicissitudes principle only applies to innocent/non-tortious hypothetical or actual subsequent harms, hence the distinction between Jobling and Baker.1 PART CAUSING A SINGLE INJURY While Performance Cars is about causing part of a divisible injury, this is about a contribution to a single, indivisible injury. If an injury is necessarily indivisible and causes cannot be divided between spate factors because those factors operate cumulatively and interdependently, then apply Bonnington:
? Bonnington Castings v Wardlaw : employee was exposed to silicon dust from two sources, but only one source was attributable to D's negligence. C developed a lung condition that gradually developed over time. Lord Keith: exposure to 'innocent' and 'guilty' sources were simultaneous, so could not be divided into chronological chunks. The condition wouldn't have happened (or wouldn't have occurred when it did) 'but for' the exposure to guilty dust, so D was liable for C's injury.
Lord Phillips in Sienkiewicz on the difference between Bonnington and Performance Cars: "Where the disease is indivisible, such as lung cancer, D who has tortiously contributed to the cause of the disease will be liable in full. Where the disease is divisible, such as asbestosis, the tortfeasor will be liable in respect of the share of the disease for which he is responsible."
There is a question as to whether Bonnington is an orthodox application of causation principles. Green argues that it is, just used where there is a difficult factual situation --- i.e. in Bonnington, D is a 'but for' cause because without D the injury wouldn't have occurred (or wouldn't have occurred when it did). This was the approach taken by the Privy Council in Williams. Williams v Bermuda Hospitals Board 
? Facts: C was admitted to hospital with appendicitis. The hospital was negligent in failing to diagnose the appendicitis and took more than 10hrs to operate. As a result, pus leaking from the appendix caused sepsis to set in and injury to C's heart and lungs. Issue: some of the leakage was non-negligent (would have occurred anyway) and some was the result of the hospital's negligence (in not operating fast enough).
? Lord Toulson (PC): C's injuries were indivisible (not greater / smaller depending on the amount of leakage). As such, Bonnington applied: "where D has been found to have caused 1 Summary: Two independent torts and both Ds before the court Performance Cars (if injury is divisible); Two independent torts and only D1 before the court Baker; Two independent injuries one tortious and the other non-tortious: Jobling.
or contributed to an indivisible injury, she will be held fully liable for it, even though there may well have been other contributing causes." The hospital was liable for the full injury. McBride:
? Different rule: Lord Toulson followed Sarah Green's approach that Bonnington is a straightforward application of 'but for causation'. However, McBride disagrees --- it is a different rule because it does not ask if D's conduct is a but for cause, it asks 'did D's negligence make a material contribution to the state of affairs that resulted in C suffering an indivisible injury'?
? Bonnington not solid authority: the PC note pneumoconiosis is a divisible disease, so Bonnington may just be a Performance Cars case where the HL went wrong in holding D liable for the whole injury, rather than just a part (i.e. he was a part cause on straightforward 'but for' causation). It may not be solid authority for a subversive rule now embedded in English law. o The PC should have been more cautious in light of Fairchild ---"the courts depart from the 'but for' test for causation at their peril" it is therefore surprising that the PC were "so happy to depart from the 'but for' test in such a casual manner." Stapleton and Steel. critical of the result in Williams, but endorse the rule in Bonnington
? Endorsement of Bonnington: "to the extent that the rule is asserted as a general principle for cases of indivisible injury it is correct and must be understood as employing a causal concept which is broader than but-for causation" E.g. where D1, D2, and D3 all independently poison C's tea --- two drops are sufficient to kill, one drop is not --- C drinks the tea and dies. 'Is the law to say that since no one was a necessary, but-for, cause of the death no one is responsible?'
? Criticism: they argue that where there is an indivisible injury, there should be two stages: (i) did D make a material contribution to an indivisible injury; (ii) "if the mechanism for an indivisible injury would have been complete absent the contribution due to wrongful conduct, no compensatory liability should be imposed even though the contribution should be recognised as a cause." The PC only ask the first question, with the result that the D in Williams is liable, but if they asked the second question he might be absolved if he could show that prompt treatment would not have prevented sepsis. They're trying to refine the scope of the exception. A good example of damage falling into this category is mental illness --- it is considered an 'indivisible injury', so once C shows D is a 'but for' material cause, D is liable for the whole injury:
? Dickins v O2 plc  C worked for O2 and suffered from stress on the job. She repeatedly expressed her concerns to her employer. C eventually suffered a breakdown. CA: O2 was liable for stress related personal injury. o Foreseeability: Previous case Walker v Northumberland CC  which stated that an employer would not be liable for an employee's first breakdown (as it was not foreseeable) was overruled. Here, it was enough that C had previously complained about stress on the job and had communicated symptoms she was suffering. As such it was foreseeable.
o Breach: even though O2 referred C to internal counselling, this was not enough to discharge their duty. Because D was complaining of severe stress, O2 should have used 'managerial intervention' --- i.e. to investigate C's case. o Causation: there were a number of factors contributing to C's breakdown --- many of which were not related to the employer (e.g. stress in her home life). Here, the court found O2 made a "material contribution" to C's breakdown and "tipped her over the edge." I.e. this is an application of Bonnington --- mental illness is an indivisible injury and because D was a material 'but for' cause, D was liable for the whole injury. MATERIAL INCREASE IN RISK This is an exceptional test and is not actually about a material contribution to risk, but about increasing risk of injury which then eventuated. It can only operate where there is an evidentiary gap Need to prove:
? Indivisible injury: single risk which has eventuated in an injury (Rothwell --- cannot claim where it has not eventuated), but different sources of risk.
? Single agent
? Evidentiary gap: 'but for' causation is impossible to prove in principle --- not where theoretically possible to prove, but where the evidence is not available. Know that one D caused the injury, not sure which one. There must be a "rock of uncertainty." The Fairchild principle allows C to recover for her illness despite the evidentiary gap on the basis that each D materially increased the risk of C suffering the illness. C only needs to prove that D materially increased the risk she would suffer the illness, not that D was a 'but for' cause. Principle first laid out in McGhee v National Coal Board : C contracted dermatitis as a result of negligent and non-negligent exposure to brick dust by D; due to evidentiary gap it was impossible for C to show if the negligent dust was a 'but for' cause, but he was allowed to claim by the HL: o Lord Wilberforce's Justification: "if one asks which of the parties, the workman or the employers, should suffer from this inherent evidential difficulty, the answer as a matter of policy or justice should be that it is the creator of the risk ... who should bear its consequences". Rational behind Fairchild: as a matter of policy, the disadvantage caused by the medical evidentiary gap should be borne by the negligent D, rather than innocent C. Further, if Ds were not liable it would allow employers to ignore some of their health and safety obligations with impunity. Fairchild :
? Facts: C contracted mesothelioma as a consequence of exposure to asbestos due to various employers' negligence. C's meso must have been caused by at least one D, but it was
Buy the full version of these notes or essay plans and more in our Tort Law Notes.