Essay Three

Leading textbooks and commentators are less than impressed at the principles behind the law of tort’s determination of causation, in one instance labelling them “quixotic”[1]. The problems which a lack of principles have caused are; an increasing pressure on the doctrine of duty of care; a dilution of logical legal reasoning in judicial judgements; and finally, the undermining of the doctrine of certainty in the rule of law, as characterised by writers such as Dicey[2] and Lord Bingham[3].  In this essay, the role of causation shall first be identified. Once this “normal way”[4], hereafter referred to as the ‘classic position’ has been identified, five fact patterns which lie outside this test will be highlighted. The third step is to analyse the alternative principles proposed by the courts, in finding causation in these five problem areas. It will be argued that these five areas, commonly regarded as fatal to the idea of a ‘but for’ test, arise not because of any problems with the causation doctrine but because they are scenarios in which the question of factual and legal causation are being examined simultaneously. This is wrong. It shall be argued that the distinction between “the evidential and the conceptual”, well identified by Ibbetson and Steel[5], should be taken further and is not merely an academic distinction. In synthesis; the ‘but for’ test retains its integrity, and problems arise from its misapplications to problems which are caught under a now-too generic umbrella labelled ‘causation’.

Firstly, the function of causation must be identified. Causation’s function is as part of the factors used to determine where tortious liability lies. These factors are four and come together such that tortious liability is found where there is: (i) a duty of care sourcing either from common law or from a statutory instrument, which (ii) is breached and (iii) causes (iv) a remediable harm. It is submitted that the seemingly problematic status of causation has resulted from an obfuscation between the second and the third question of tort. Causation in tort is a factual question. To prove causation, it must be shown that, on the balance of probabilities, A (a remediable harm) has happened because of B, which must be an act of C and not an Act of God, nor an inevitable act[6], nor a necessary act[7], nor the result of a reasonable defence of property[8], nor an act allowed by statutory authority[9].

How is ‘cause’ identified? The ‘but for’ doctrine states that an act can be said to have caused the harm, if it is “an action without which the harm or loss suffered could not have occurred”[10]. In Bonnigton Castings Ltd. v Wardlaw[11] this is simple to establish. But for the defendant’s failure to keep their factory’s grinders free from obstruction, the appellant would not have contracted pneumoconiosis. This will be referred to as the ‘classic position’ – a situation where it is simple to identify the cause and the harm.

However, nearly half a century later, Lord Nicholls[12] summarises the apparent difficulties that this test has run into by stating that the test is now considered “over-exclusionary” (at [73]. The second step therefore is to identify positions where the classic test will not function.

Lord Nicholls goes on to give the example of: (i) multiple wrongdoers – “the classic example is where two persons independently search for the source of a gas leak with the aid of lighted candles. According to the simple ‘but for’ test, neither would be liable for damage caused by the resultant explosion” (at [74]).

(ii) Causal indeterminacy will also not be caught under classic cases of the ‘but for’ test. Causal indeterminacy is the term given where A might have caused B, just as C might have caused B but either (a) only one of A and C is tortious, or (b) only one of A and C is attributable to the given defendant. The reason that this cannot be formulated into a valid ‘but for’ formula is that any such attempt would lead to an invalid position. Take position 1: ‘but for A, B would not have occurred’ is false because (a) it might be that A did not cause B and (b) it might be that B was caused by C – and thus that ‘but for C, B would not have occurred’.

(iii) Loss of a chance cannot be caught under the classic doctrine. Where A decreases the chance of B, this cannot be formulated into any proposition of the ‘but for’ test. Position 3: ‘but for A, B would occur’ does not accurately portray loss of a chance. Position 4 accurately describes the situation: ‘but for A, B would have a higher chance of occurring. However, this does not fit into the classical position because ‘B’ does not stand for the harm but rather the lack of harm (i.e. ‘but for A, not harm would have a higher chance of occurring’). Where one seeks to formulate a variant: ‘but for A, harm would have a lower chance of occurring’ this is (still) formulated with the language not of certainty but of probability.

(iv) Fourthly, cumulative causes do not come under the first definition of the ‘but for’ test because they represent a fact situation where ‘but for A, B, etc. C would not have occurred’. This is clearly outside the classical position because no single factor is determinative.

(v) Finally, overtaken causes are the fifth form of fact patterns which cannot be caught under the classic form. This is less to do with the ‘causal’ element and more to do with the result. The fact situation here is: ‘but for A, B would not have occurred, and B would not have been worsened but for C’. Properly, this should therefore be separated into two parts: position 5 ‘but for A, B would not have occurred’ and position 6 ‘but for C, D would not have occurred, where proposition 5 has already occurred’.

The courts have not failed to find tortious liability for defendants in each of the five situations due to problems of causality. How, then, has the ‘causal problem’ been solved?

(i) Where there are multiple wrongdoers the fact pattern lies outside of the ‘but for’ test, as shown by the example of the two individuals searching for a gas leak with lighted candles. The case law instead provides that liability is attached to those defendants whose actions are “sufficiently close”[13] to the eventual harm. Lord Nicholls, in Kuwait Airways v Iraqi Airways (No. 6) [2002] goes on to say: “in so deciding the court is primarily making a value judgment on responsibility. In making this judgment the court will have regard to the purpose sought to be achieved by the relevant tort, as applied to the particular circumstances” (at [74]). In the cited case, His Lordship goes on to discuss the purpose of the tort of conversion.  This approach to dealing with multiple defendants seems commendable. The principle of sufficient closeness, to look for what contributed to the likelihood of the harm is clear because it can be broken down into clear questions, chiefly; (a) what is the principle of the tort? (b) did the alleged action increase the chance of the harm manifesting (c) does accepting this action as part of the increase compliment the nature of the tort?

(ii) Causal indeterminacy. Where scientific or medical knowledge cannot determine the cause of the harm, as shown in position 1. In McGhee v National Coal Board[14], the claimant contracted dermatitis as a result of two forms of exposures – one which was negligent (due to a lack of showers) and one which was lawful (as part of the job) –  yet which was the exposure which caused the disease could not be determined. Their Lordships, in order that the claimant might still recover, ruled that it was enough that the defendant has caused any “increase in risk”. Wilsher v Essex Area Health Authority[15] clearly states that the rule in McGhee is not a reversal of burden of proof. Thus, on the facts, where the negligence of a junior doctor had led to a baby losing sight in one eye, and being partially blind in the other, it was held that it must still be proven that the alleged negligent act caused the harm. The language of Fairchild v Glenhaven Funeral Services[16] suggests that, though the standard of proof was not reversed, it was lowered. In a limited number of cases (usually those surrounding mesothelioma) the test was indeed an “increase in risk”, which was material – a “material increase” or a “material contribution”. It is important to note that, as per Gregg v Scott[17], there is no new tort of ‘materially increasing a risk’. There must be a harm which has resulted from the increase in risk. Barker v Corus UK Ltd.[18] ruled that in these cases, damages apportioned based on the material increase in risk – although the Compensation Act 2006 overruled Barker for mesothelioma cases, providing instead that each defendant will be liable in full for the damage, and that it is left to the defendants to recover contribution from the others. What do the line of cases dealing with causal indeterminacy show about the ‘but for’ test? McGhee – Wilsher – Fairchild – Gregg v Scott – Barker essentially abandon the doctrine of looking for the one event without which the harm would not have occurred. Instead, as the language of Lord Hoffman in in Barker v Corus UK Ltd.[19] accepts, liability is attributed “according to the relative degree of contribution”. Essentially, the principle is to look for the defendant who contributed to the likelihood of the harm.

(iii) Loss of a chance. In Allied Maples Group Ltd. v Simmons & Simmons[20] it was ruled that loss of a chance consequent on tortious action could be recoverable. In Hotson v East Berkshire Health Authority[21], a child fell from a tree and damaged his hip. This fall damaged blood vessels which led to an enduring disability of the limb – however enough blood vessels might have survived the fall that it was only due to occlusion of those surviving blood vessels as a result of the negligent misdiagnosis of the doctor that the disability manifested. The witness for the prosecution state that there was a 50% likelihood that a diagnosis would have prevented the harm because the occluded blood vessels could have been saved. Meanwhile, the defence’s witness said that the harm was not preventable because the blood vessels had already been killed in the fall. The trial judge averaged this (50%+0%/2) gave the claimant 25% of the damages the injury would have incurred. This is a modified version of the approach in McGhee; liability for a change likelihood of damage. At appeal, Lord Mackay applied the 25%, in deciding not the amount of compensation, but in finding whether the ‘past fact’ of where there were sufficient undamaged blood vessels for the harm to have ever been avoided. 25% fails the test of the balance of probabilities. The conclusion was therefore that there was no causation as the harm could not have been avoided. Miller[22] presents the ruling Hotson v East Berkshire Area Health Authority in similar terms[23]. However, the article argues that there are two principles at play here, for there to be recover. Firstly, negligence must occur before the chance of recovery was 0%. Secondly, the chance of recovery must exceed 50%, per the standard balance of probabilities before the negligence[24]. With respect, it must be questioned whether this was Lord Mackay’s intent. An alternative interpretation of the ruling is that: where the effect of the negligence is unknown, an objectively greater than 50% chance of recovery before the negligence, must be lost. This postulation explains the otherwise seemingly discordant ruling in in Gregg v Scott[25]. The negligent misdiagnosis of a patient by a doctor meant the chance of curing them from cancer fell from 42% to 25%. Unlike Allied Maples v Simmons & Simmons and Hotson v East Berkshire Health Authority there is no question of past fact which needs must be answered. There is a clear harm (the drop of 17%) which is the ‘harm’ which is being recovered, under the second limb of Gregg v Scott. Lord Mackay was correct, in his determination of Hotson v East Berkshire Area Health Authority to not adopt the approach of the Supreme Court of America of Herskovits v Group Health Cooperative of Puget Sound[26]. The facts of this American case are similar to what would be the facts of Gregg v Scott: there was a loss of chance in the curing of a cancer due to a negligent misdiagnosis. Miller must be wary of over-generalisation. The two-tiered test seems acceptable in cases where the loss of chance is not known; where the loss of chance is known, the principle is that the defendant is proportionately liable for that loss of chance.

(iv) Cumulative disease is governed by the rule laid out in Bonnington Castings v Wardlaw Ltd.[27] it was ruled that liability should be apportioned according to whether or not there was a “material contribution” to the disease. It is relevant to note that Bonnington Castings was modified in Holtby v Brigham & Cowan (hull) Ltd.[28] such that liability should be apportioned according to the contribution of only negligent actions; it would be manifestly wrong to punish a defendant for an unavoidable contribution, to which the claimant might even have agreed to face, or even to a contribution incurred by the claimant. In summary therefore, it may be suggested that the principle behind cumulative disease is that those who participate in the end harm are liable insofar as they participated because they increased the risk of the harm.

The distinction between a case of type (ii) and those of type (iv) is important because it governs which principle should be applied. Academics have created an approach to distinguish between them. Miller is instructive, in describing the distinction through the ‘necessary element in at least one sufficient set [to causing the harm]’ idea[29]. In situations where there is only one sufficient set (i.e. there is only one way in which the harm could have eventuated), if an action is necessary to it then it satisfies the basic ‘but for’ test. To take the example of Miller’s cited article, if ‘but for Lady Macbeth’s intoxicating of Duncan, he would not have died’, that makes her equally liable as Macbeth, if ‘but for Macbeth’s stabbing, Duncan would not have died’. This is seen in type (i) cases.  However to turn to the article’s second example, of the assassination of Julius Caesar, if four wounds would have killed him, then any combination of four of the 23 conspirators would have been enough. Thus, any four can be said to be necessary to a ‘set’. This, Miller helpfully labels as ‘cumulative’ and can be identified what is here labelled as (ii) cases. This is distinct from the situation where only one thrust of the four killed Caesar – yet the fact that there were four thrusts meant it was more likely that one would be a killing one. The distinction in the former case is that there is an increase of likelihood of the harm, whilst in this latter case there is an increase in the risk of the harm. These are the type (iv) cases.

(v) Overtaken causes. Where a negligent act has caused a harm, and further harm is caused by a second negligent harm, the rule in Performance Cars v Abraham[30] states that the second tortfeasor is liable only where they have extended the harm past its original severity in some way. Where the second event causing greater damage is natural, then the first tortfeasor is not liable for the extra harm, per Baker v Willoughby[31] and Jobling v Associated Diaries[32]. Strictly, this should not be considered a problem of causation. Instead, this should be seen as an exercise of attribution of liability, so that the claimant recovers the extent of the injury and the aggravated form of the injury he has suffered, without the defendant having to be liable for more than he has contributed. For this reason, although the classical ‘but for’ test is not sufficient, a double formulation (per position 5 and 6) would allow such situations would be covered. As a result, overtaken causes are to be seen as to espouse no different a principle than the classical ‘but for’ causation.

Liability, consequently is found, even in those cases which lie beyond the pale of the accepted causation doctrine. This is not satisfying, although the results of the cases might be. Soar convincingly argues[33] that cases are moving away from traumatic injuries, and towards the non-traumatic disease harm which results from industrialised processes with chemical products – notably, thalidomide and mesothelioma[34]. This means that there will be more cases casting doubts on causality. The courts openly recognise that this is a policy question: the late Lord Bingham of Cornhill, proposes that causation may be extended if “special circumstances of a case such a case, principle, authority, or policy requires or justifies [own emphasis] a modified approach to proof of causation”[35]. However, the courts cannot be left, unprincipled to deal with a policy question. This is not their role. Thus, there are two alternatives. Either the government legislates – which, on a topic as wide-ranging as causality, would probably do more harm than good – or a principle can be identified to guide causality which functions even in the cases highlighted as outside the ‘classic position’.

If two principle reservations can be dealt with then the extensions in causation can be labelled justifiable. Once this is done, then a doctrine to ensure that they are caught too can be proposed. First then are two principle reservations: (i) are the new classes of defendants covered in the alternative principles, ‘deserving’, and (ii) does the new analysis still come within the bounds of probabilistic causation, which per Saur, is the heritage of the English legal system? Since the answer to both queries is in the affirmative, it is concluded that extension of causation is justifiable, and all that remains is to find a principle to guide its expansion.

In the case of the Irish Bank Resolution Corporation Ltd. (in special Liquidation and others v Quinn and Others[36] at [63] the judgement of O’Donnell J in Minister for Justice v Devine, unreported, Supreme Court 2012 was cited: a principle of English law[37] is that “the deserving Plaintiff with the good claim” must not be left without a remedy. This is one of the most recent cases to make such a claim; there are many others. However, who a ‘deserving claimant’ is, is not well defined in legal literature. The prospective researcher is warned that there is no database which is “fully comprehensive”[38]. BAILII: Case Law – the “most comprehensive set of British and Irish primary legal materials freely available online”[39] – returns no results of any articles discussing the definition of a “deserving claimant”[40]. In the hope that a pay-access database provides a better answers, let us turn to LexisLibrary which claims to offer “unrivalled coverage of secondary legal context”[41] as well as “extensive legal commentary”[42]. Once again, despite many hits of the term, there is not one which offers a definition[43].

Why this concern about the meaning of deserving claimant? Fairchild v Glenhaven Funeral Services Ltd. is recognised and justified because it is a policy decision based on the idea of giving a “highly deserving claimant” a remedy. In Ellis v Environment Agency[44] a pertinent paragraph is dedicated to this idea (at [22]): “I [Lord Justice May] take Fairchild […because it shows…] that departure from the normal rule is exceptional and to be approached with caution; and that a modified approach to proof of causation was justified to assist a highly deserving claimant who from the nature of the facts had evidentiary difficulty in proving that the negligence of one of two or more possible tortfeasors, whose negligence may have operated successively, satisfied on the balance of probabilities the “but for” test” (own emphasis). It is therefore dangerous to justify such a break from causation if it cannot be bounded within a test, because of the inevitable worry that judges will simply give redress to those claimants who they have some sympathy with. If concern of whether the claimant is deserving lies at the discretion of the court, then we are giving the power to judges to merely give compensation to those claimants who they think should get compensation – it is obvious that this is fatal to principles of certainty and impartiality of the law. If, as other cases may state it is, for example, “the layman[‘s]” question, and that it is they who must chose who is a “deserving plaintiff”[45] then, at first glance this seems more satisfactory because the answer is given by one’s peers. Trial by peers is an important part of the English legal system[46]. Yet this produces the twin problems that (a) the legal system is allowing a ‘layman’ to decide without direction who they think should receive a remedy[47] (b) there is no certainty that the layman of metropolitan London has the same sympathies as that of the layman of the more rural St. Ives.

Certainly, the unquestioned use of this term has, thus far not caused any issues. However, to hide behind the lack of an existing problem is to merely state that, ‘because the courts seem to be getting the choice between a deserving and a non-deserving claimant right so far, so we should just let them carry on’. It can be seen that this is not persuasive. It is however, the best that can be offered. More comfort can be drawn from the fact that if the courts were to start making controversial decisions, academics (and perhaps more importantly journalists) would certainly start turning their attention to this issue.

The second reservation is about the nature of causation. If the English system is happy with ‘but for’ test, why are there any situations in which there can be a deserving claimant who lies outside of them? Surely this means that the principle is insufficient and that it must itself be revised? It is proposed that the reason why the ‘but for’ test would not find the cases which have departed from its doctrine liable is because it seeks to determine both the factual and the legal question of causation at once, even though it was only ever designed to answer the factual question. The conflation between these two aspects of causation is wrong; just as the stretching of the ‘but for’ test to cover them is wrong. As shown by Steel and Ibbetson’s article[48], the ‘but for’ test should cover merely the conceptual question, whilst the balance of probability should cover the evidential question.  The evidential question must be passed for the ‘but for’ test to be effective. Reference to the words of Lord Nicholls’ in Kuwait Airways v Iraqi Airways[49] should be made: his Lordship holds that there is a distinction between the factual and the legal. It does not follow that even if the factual is satisfied, the legal is satisfied. The factual is only ever a threshold test, at [73]. There is then the secondary question, at [72], of whether “the defendant should be legally responsible for the loss”. What cases (i) through to (v) present, are situations in which the factual question cannot be answered. It is unrealistic for the ‘but for’ test to then be expected to pick up the slack of the threshold question.

To present the various theories of factual evidence, analyses them, and determine which is best is to far extend the scope of this essay. Consequently, what is advanced from here is a reflection of personal opinion. This opinion is bounded by agreement with the statement by Soar that the whilst the factual test needs must be probabilistic[50], whilst the legal step is one not of fact but of imputation, of ‘ought’ the defendant be liable, Kelsen[51]. Soar argues that conception of causation is either deterministic or probabilistic. Either one believes that a cause is a sufficient, or part of a sufficient set, leading to an event (the deterministic view) or that a cause must be found to be a cause based on the probability of leading to the event (the probabilistic view). Support for the latter position is that, by proving a lower test, it is more likely to not be under-inclusive. This permits a wider field of possible causes be presented for the second stage of analysis: Kelsen’s question – ‘ought’ the cause create liability? Kelsen’s ambiguous position between positivism and normativism supports the idea of a test which deals with probabilities. It is irrelevant whether the law is imposing or manifesting values, ideals, and behaviour on individuals. The question which is one step less removed is whether something (X) might cause a harm (Y), it ought to be held liable as if causing that stop. Balance of probabilities and percentages are mere lines drawn along the spectrum of how likely X=Y is to be true. Since the basic precise that where X tends to Y, and Y is a harm, is agreeable however, the conclusion must be the courts are justified in relaxing the standard doctrine of causation for one which is based on various principles – (i) sufficient closeness (ii) contribution to the likelihood of the harm. (iii) where the effect of the negligence is unknown, an objectively greater than 50% chance of recovery before the negligence, must be lost or where the loss of chance is known, liability is proportional to the loss of chance (iv) contribution to the risk of harm – because they each are underscored by the ideal of contribution to a harm. It is submitted that contribution is enough to justify the extension of causation, and that this is not a weakening of causation, but a mere acceptance of the fact that in certain things, no higher certainty than contribution can be identified.

[1] Miller, C., Causation in Personal Injury after (and before) Sienkiewicz Legal Studies, Vol. 32 No. 3, September 2012, p. 396.

[2] Dicey, A. V, An Introduction to the Study of the Constitution (1885).

[3] Bingham T., The Rule of Law (2010).

[4] Bailey, S.H., ‘Causation in negligence: what is a material contribution?’ (2010) 30 LS 167.

[5] Steel, S. & Ibbetson, D. More Grief on Uncertain Causation in Tort, The Cambridge Law Journal, Volume 70, Issue 02, July, 2011, pp. 451-468.

[6] Pollock, Torts, 15th edition.

[7] Southport Corporation v Esso Petroleum [1954] 2 QB 182.

[8] As long as it is not injury to an innocent of a third party, per Scott v Shepherd [1773] 2 WBI 892

[9] And the necessity has not resulted because of the defendant’s own negligence: Farnworth v Mancehteser Corporation [1929] 1 KB 533.

[10] Bonnington Castings Ltd. v Wardlaw [1956] AC 613.

[11] [1956] AC 613

[12] Kuwait Airways v Iraqi Airways (No. 6) [2002] 2 AC 883.

[13] Kuwait Airways v Iraqi Airways (No. 6) [2002] 2 AC 883.

[14] [1973] 1 WLR 1

[15] [1988] AC 1074

[16] [2003] 1 AC 32

[17] [2005] AC 175

[18] [2006] 2 AC 572

[19] [2003] 1 AC 32

[20] [1995] 1 WLR 1602

[21] [1987] AC 750

[22] Miller C., Loss of chance in personal injury: a review of recent developments, Law, Probability and Risk (2006) 5 (10): 63.

[23] Ibid.

[24] Ibid. P. 2, section ‘Hotson Revised’.

[25] [2005] AC 175

[26] (1983) 664 P.2d 474

[27] [1956] AC 613

[28] [2000] ICR 1086

[29] Miller, C., Causation in Personal Injury after (and before) Sienkiewicz Legal Studies, Vol. 32 No. 3, September 2012, p. 396. Abbreviated to ‘NESS’.

[30] [1962] 1 QB 33

[31] [1970] AC 467

[32] [1982] AC 467

[33] Soar, K., A Probabilistic Theory of Legal Causation, University College London Jurisprudence Review, 132, 1996.

[34] The first problems of causality manifested during the Industrial Revolution’s exponential building of factories by rivers, which were consequently polluted. However, multiple factories on rivers meant that each owner could state that his contribution was not necessary  to the pollution as the others on the river had caused it. (Steel, S. & Ibbetson, D. More Grief on Uncertain Causation in Tort, CLJ, Vol. 70, Iss. 02 (2011).)

[35] Fairchild v Glenhaven Funeral Services Ltd. [2002] UKHL 22

[36] [2013] Irish Approved Judgments.

[37] Expressed in Wainwright v Criminal Cases Review Commission [2012] EWH 2988 (Admin). Also consider the array of concern directed at legislation referring to defamation, which believes the high cost of litigating under it deters those who are ‘deserving’ whilst allowing the rich to pursue fanciful claims (Mullis A., and Scott A., Tilting at Windmills: The Defamation Act 2013, (2014) 77(1) MLR87–109.

[38] Site Accessed: 11:24, 26th July 2014.

[39] Site Accessed: 11:27, 26th July 2014.

[40] Result of search at 11:28, 26th July 2014.

[41] Site Accessed at 11:31, 26th July 2014.

[42] Ibid.

[43]||BOOLEAN|||| Site Accessed 12:07, 26th July 2014.

[44] [2008] EWCA Civ 1117.

[45] Hayes v Bowman [1989] 2 All ER 293.

[46] Bingham, T. The Rule of Law (2010).

[47] This is distinct from the idea of a trial by jury in that a jury is directed.

[48] Steel, S. & Ibbetson, D. More Grief on Uncertain Causation in Tort, CLJ, Vol. 70, Iss. 02 (2011).

[49] (No. 6) [2002] 2 AC 883.

[50] Which is the ‘evidential’ of Ibbetson and Steel.

[51] Freeman, A. Lloyd’s Introduction To Jurisprudence, 7th edition, Thomson, London, Sweet&Maxell Ltd. (2001). P.256. Pp. 278-9 (extract of Hans Kelsen’s Pure Theory of Law).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s