In Graves -v- Brouwer [2015] EWCA Civ 595 the Court of Appeal carried out an extensive review of the principles and authorities relating to evidence and causation.  There is a useful discussion on the role and questioning of experts at trial.

“The question or questions which elicited it ought never to have been put to [the expert] since it, or they, invited him to express a view on a conclusion of mixed fact and law which it was the task of the judge to reach


The claimant suffered extensive damage to her house following a fire. She brought an action against the defendant claiming that he caused the fire in her home by burning two A4 pieces of paper and two A4 pieces of card in a narrow alleyway next to her home some 30 minutes before the fire was discovered.


  • The fact that one possible cause of an incident was discounted did not make it inevitable that the other possible cause could be established.
  • An expert should not be asked to reached conclusions of mixed fact and law which were the sole province of the judge.
  • There are cases in which a claimant, simply, fails to establish causation on the balance of probabilities.


The trial judge found:

1. The small fire by the defendant was responsible for the fire at the claimant’s home.

2. The defendant was not,however, negligent.


The claimant was even less fortunate on appeal. The court upheld the original finding that the defendant had not been negligent (it is not negligent to light a small fire in an alleyway – although it is “bizarre”).


The Court of Appeal, however, dealt with the issue of causation first and disagreed with the trial judge’s conclusions. The reasoning is important.

The judge’s conclusion – causation

  1. The judge’s conclusion on causation is contained in the following passage in her judgment:-

“21. …

Mr Demachkie submits that as the claimant has failed to prove her case on the balance of probability that it matters not how likely it is relevant to other causes. Put simply, he says the claimant has not proved her case. He asserts this is based on the expert evidence stating from his closing submissions.

Mr Demachkie further contends:

“The experts have corroborated this during the trial. The various responses in which this was confirmed will have been noted. The only point to draw to the court’s attention was Mr Houghton confirming under cross-examination that, on the balance of probabilities, the fire as described by the defendant would not have caused the house fire at number 9.”

However, this is to ignore the final exchange between Mr Weston, the claimant’s counsel, and Mr Townsend, the defendant’s expert, which was as follows:

“If the court takes the view that arson did not cause the fire then, on balance of probability, it must be the cause even if improbable.”

To which the answer was:

“If the court deems arson unlikely or less plausible than the fire of an ember from the items lit, then that would be the conclusion.”

It was put:

“That is as you test it and Mr Brouwer describes?”

To which the answer was: “Yes.” I consider that exchange material. It was a significant shift of view. Had Mr Townsend not conceded as he did, given Mr Houghton had conceded in cross-examination that, on the balance of probabilities, the fire, as described by the defendant, could not have caused the house fire at number 9 then I would have acceded to Mr Demachkie’s submission.

22. I do not consider that Mr Weston’s submission re: the balance of probability is correct. His client had to prove, on the balance of probabilities, that the cause of the fire was the fire in the alleyway. It was not a matter of the court simply assessing the time frame of events, the location of the fire and its seat. That might have been so in the absence of expert evidence. However, there was expert evidence before the court and his own expert’s conclusion was that the fire was not caused, on the balance of probability, by an alleyway fire. That would have been the end of the matter had he not succeeded in extracting the concession from Mr Townsend. However, as the defendant’s own expert takes the view the cause of the house fire, on the balance of probability, was the alley fire (and the maintenance throughout by Mr Houghton, the fire in the alleyway was the cause of the fire, albeit not on the balance of probability, this court cannot be doubtful as to the real cause of the loss. The only realistic possibility proffered by Mr Townsend was arson. I reject that possibility. It would have required an unknown person to have appreciated when Mr Brouwer was about to start his fire and to start it in an area which was directly above the location. Realistically, it would have had to have been a member of the Graves/Moore family. As Mr Brouwer conceded in his evidence that he had not seen any strangers about, that is fanciful in the extreme. It was the alley fire. It follows, therefore, that the claimant succeeds on causation.”

Causation – discussion

  1. In an earlier passage the judge made express reference to Fosse Motor Engineers v Conde Nast Distributors Limited [2008] EWHC 2037 (TCC), identifying as the key passage paragraphs 61-68 where Akenhead J cited extensively from the well-known speech of Lord Brandon in Rhesa Shipping Co v Edmunds, The Popi M, [1985] 1 WLR 948, including the following passages:-

“My Lords, the appeal does not raise any question of law, except possibly the question what is meant by proof of a case “on a balance of probabilities.” Nor do underwriters challenge before your Lordships any of the primary findings of fact made by Bingham J. The question, and the sole question, which your Lordships have to decide is whether, on the basis of those primary findings of fact, Bingham J. and the Court of Appeal were justified in drawing the inference that the ship was, on a balance of probabilities, lost by perils of the sea.

In approaching this question it is important that two matters should be borne constantly in mind. The first matter is that the burden of proving, on a balance of probabilities, that the ship was lost by perils of the sea, is and remains throughout on the shipowners. Although it is open to underwriters to suggest and seek to prove some other cause of loss, against which the ship was not insured, there is no obligation on them to do so. Moreover, if they chose to do so, there is no obligation on them to prove, even on a balance of probabilities, the truth of their alternative case.

The second matter is that it is always open to a court, even after the kind of prolonged inquiry with a mass of expert evidence which took place in this case, to conclude, at the end of the day, that the proximate cause of the ship’s loss, even on a balance of probabilities, remains in doubt, with the consequence that the shipowners have failed to discharge the burden of proof which lay upon them.

This second matter appears clearly from certain observations of Scrutton L.J. in La Compania Martiartu v. The Corporation of The Royal Exchange Assurance [1923] K.B. 650. That was a case in which the Court of Appeal, reversing the trial judge, found that the ship in respect of which her owners had claimed for a total loss by perils of the sea, had in fact been scuttled with the connivance of those owners. Having made that finding, Scrutton L.J. went on to say, at p. 657:

“This view renders it unnecessary finally to discuss the burden of proof, but in my present view, if there are circumstances suggesting that another cause than a peril insured against was the dominant or effective cause of the entry of sea water into the ship . . . and an examination of all the evidence leaves the Court doubtful what is the real cause of the loss, the assured has failed to prove his case.”

While these observations of Scrutton L.J. were, having regard to his affirmative finding of scuttling, obiter dicta only, I am of opinion that they correctly state the principle of law applicable. Indeed counsel for the shipowners did not contend otherwise.

My Lords, the late Sir Arthur Conan Doyle in his book “The Sign of Four”, describes his hero, Mr. Sherlock Holmes, as saying to the latter’s friend, Dr. Watson: “how often have I said to you that, when you have eliminated the impossible, whatever remains, however improbable, must be the truth?” It is, no doubt, on the basis of this well-known but unjudicial dictum that Bingham J. decided to accept the shipowners’ submarine theory, even though he regarded it, for seven cogent reasons, as extremely improbable.

In my view there are three reasons why it is inappropriate to apply the dictum of Mr. Sherlock Holmes, to which I have just referred, to the process of fact-finding which a judge of first instance has to perform at the conclusion of a case of the kind here concerned.

The first reason is one which I have already sought to emphasise as being of great importance, namely, that the judge is not bound always to make a finding one way or the other with regard to the facts averred by the parties. He has open to him the third alternative of saying that the party on whom the burden of proof lies in relation to any averment made by him has failed to discharge that burden. No judge likes to decide cases on burden of proof if he can legitimately avoid having to do so.

There are cases, however, in which, owing to the unsatisfactory state of the evidence or otherwise, deciding on the burden of proof is the only just course for him to take.

The second reason is that the dictum can only apply when all relevant facts are known, so that all possible explanations, except a single extremely improbable one, can properly be eliminated. That state of affairs does not exist in the present case: to take but one example, the ship sank in such deep water that a diver’s examination of the nature of the aperture, which might well have thrown light on its cause, could not be carried out.

The third reason is that the legal concept of proof of a case on a balance of probabilities must be applied with common sense. It requires a judge of first instance, before he finds that a particular event occurred, to be satisfied on the evidence that it is more likely to have occurred than not. If such a judge concludes, on a whole series of cogent grounds, that the occurrence of an event is extremely improbable, a finding by him that it is nevertheless more likely to have occurred than not, does not accord with common sense. This is especially so when it is open to the judge to say simply that the evidence leaves him in doubt whether the event occurred or not, and that the party on whom the burden of proving that the event occurred lies has therefore failed to discharge such burden.

In my opinion Bingham J. adopted an erroneous approach to this case by regarding himself as compelled to choose between two theories, both of which he regarded as extremely improbable, or one of which he regarded as extremely improbable and the other of which he regarded as virtually impossible. He should have borne in mind, and considered carefully in his judgment, the third alternative which was open to him, namely, that the evidence left him in doubt as to the cause of the aperture in the ship’s hull, and that, in these circumstances, the shipowners had failed to discharge the burden of proof which was on them.”

Akenhead J also there cited the following passage from the judgment of Thomas LJ, as he then was, in Ide v ATB Sales [2008] EWCA Civ 424 at paragraph 6:-

As a matter of common sense it will usually be safe for a judge to conclude, where there are two competing theories before him neither of which is improbable, that having rejected one it is logical to accept the other as being the cause on the balance of probabilities. It was accepted in the course of argument on behalf of the appellant that, as a matter of principle, if there were only three possible causes of an event, then it was permissible for a judge to approach the matter by analysing each of those causes. If he ranked those causes in terms of probability and concluded that one was more probable than the others, then provided those were the only three possible causes, he was entitled to conclude that the one he considered most probable, was the probable cause of the event provided it was not improbable.”

I regard that passage as containing a sound guide to safe decision making. The judge also referred to paragraphs 100-102 in the judgment of His Honour Judge Toulmin in Amsprop Ltd v I.T.W Ltd [2009] EWHC 2689 (TCC) as being key. At paragraph 101 Judge Toulmin cited the following passage from the judgment of Waller LJ in Kiani v Land Rover Ltd [2006] EWCA Civ 880 at paragraph 30:-

“I do not myself think it is false logic to reason where only two possibilities are under consideration, both of which seem unlikely, if one seems much less likely than the other the less likely can be discounted thus making the first likely to happen on the balance of probabilities.”

  1. In Kiani the claimant, an industrial cleaner, had been found relatively uninjured but asphyxiated in a sub-floor tank containing coolant at the Land Rover plant in Solihull which he was cleaning as part of his work. It was common ground that there were only two possible causes of his being in the tank – either he had accidentally fallen in or he had committed suicide. The trial judge found it hard to visualise how Mr Kiani could have fallen in accidentally, but bore in mind that accidents happen in the most unlikely of ways. He regarded suicide as less than probable. It was in this context that Waller LJ made the observation which he did. He also said, at paragraph 32:-

“In my view therefore the recorder was entitled to take the view that suicide was “less than probable”. If the only evidence before the recorder had established that an accident was “impossible”, I would accept that the recorder would either have had to decide that the “less than probable” had happened or hold that the claimant had failed to prove his case. But the evidence did not go so far as to say that an accident was impossible “whatever the facts”. The recorder took the view on the evidence before him unconsidered as it would seem by Mr Rawden, that there was a way in which Mr Kiani could have accidentally fallen into the tank when going over to have a look, and overbalancing whilst trying to see into the tank.”

  1. We were referred, although the judge below was not, to the decision of this court in Milton Keynes Borough Council v Nulty & others [2013] 1WLR 1183 in which Toulson LJ, as he then was, essayed a helpful distillation of the authorities. This too was a fire case, in which the trial judge had concluded that there were only two possible causes, either that it had been caused by arcing of an electric cable or by a cigarette carelessly discarded by an experienced electrical engineer. Having cited the first passage above from the judgment of Waller LJ in Kiani, the judge concluded that “it might be regarded as unlikely that an experienced electrical engineer, who had in the past been a part-time fireman, would choose to smoke in a building to which he knew a no-smoking policy applied and then discard the cigarette end – albeit one that he thought he had stubbed out – into flammable waste lying on the floor. But if the only other possible causes of this fire are very much less likely, as I find they are, then in law the discarded cigarette becomes the probable cause of the first fire. As Waller LJ observed in the passage that I have emphasised above, this process of reasoning is not false logic.”

  2. The Court of Appeal held that the trial judge had been wrong to adopt this approach. After discussing the speech of Lord Brandon in The Popi M, Toulson LJ said:-

“34. A case based on circumstantial evidence depends for its cogency on the combination of relevant circumstances and the likelihood or unlikelihood of coincidence. A party advancing it argues that the circumstances can only or most probably be accounted for by the explanation which it suggests. Consideration of such a case necessarily involves looking at the whole picture, including what gaps there are in the evidence, whether the individual factors relied upon are in themselves properly established, what factors may point away from the suggested explanation and what other explanation might fit the circumstances. As Lord Mance observed in Datec Electronics Holdings Limited v UPS limited [2007] UKHL 23, [2007] 1 WLR 1325, at 48 and 50, there is an inherent risk that a systematic consideration of the possibilities could become a process of elimination “leading to no more than a conclusion regarding the least unlikely cause of loss”, which was the fault identified in The Popi M. So at the end of any such systematic analysis, the court has to stand back and ask itself the ultimate question whether it is satisfied that the suggested explanation is more likely than not to be true. The elimination of other possibilities as more implausible may well lead to that conclusion, but that will be a conclusion of fact: there is no rule of law that it must do so. I do not read any of the statements in any of the other authorities to which we were referred as intending to suggest otherwise.

39. In expressing himself in those terms, the judge went further than he needed in order to answer the question which he had raised at [215]. Having concluded that a cigarette end carelessly discarded by Mr Nulty was the most probable cause of the fire, he posed the question whether on the authorities that finding was “sufficient” for him to hold that the council had discharged the burden of proof.

40. The answer to that question was yes, provided (as I think is implicit in his reasoning) that he was satisfied on all the evidence that the case for believing that Mr Nulty caused the fire was stronger than the case for not coming to that belief. The proviso is important as illustrated and explained in the Popi M. In that case the combined effect of the gaps in the court’s knowledge and the cogency of the factors telling against the theory of a collision with a submarine was that the court could not properly be persuaded that the case for believing the submarine theory was stronger than the case for remaining agnostic.”

The court did however conclude that the judge had reached the right conclusion on the facts. In his final paragraph Toulson LJ concluded:-

“69. The judge’s finding that the cause of the fire was “very much less likely” to have been arcing of the Boa cable than the cigarette end discarded by Mr Nulty was reached after a painstaking examination of the evidence and was properly open to him. Rational analysis of the Boa cable theory showed that it was highly improbable. It would have required a remarkable combination of unlikely events. The objection to the discarded cigarette end theory was of a different nature. There was no comparable scientific or practical improbability about it. Mr Nulty was working alone at the relevant place and time. He had the opportunity and could well have had the temptation to do what the council alleged. The objection to it was that he would not have been expected to behave in such a way, but the circumstantial evidence that he did so on this occasion was compelling. In my view the judge was right to find that he did so and I would dismiss the appeal.”

  1. The present case is in some ways the obverse of Milton Keynes, although I would stress at the outset one very important distinguishing feature which is that in that case there had been a thorough forensic investigation into the fire. In this case the flying ember theory was scientifically improbable. The arson theory, which ought really to be described, as did the experts, as the arson or juvenile fire setting theory, the judge rejected as fanciful, but there was no comparable scientific improbability about it. It was a theory which had been neither investigated nor properly considered. By the same token there had been no consideration of smoking materials or cooking and heating appliances. One possibility discussed at trial was the phenomenon known as “anting” – birds picking up discarded cigarettes and returning them to their nest. This is apparently an unusual but occasional cause of fires. That too had not been investigated, save to the extent that Mr Cronk of Hawkins recorded Mr Moore as telling him that he was aware of some birds nests in the eaves on the opposite side of the house, next to number 7, but did not think that there had been any birds nesting under the eaves adjacent to number 11, where the fire occurred. Mr Townsend did not think it a likely cause.

  2. It is plain from the concluding part of paragraph 21 of her judgment, which I have reproduced at paragraph 23 above, that the judge regarded one single answer from Mr Townsend as critical to the outcome of the enquiry into causation. Without it she would have found that the Claimant had failed to prove that the cause of the house fire was an escape of fire from the alleyway. The judge characterised the single answer as both a significant shift of view and a concession. With respect to the judge the answer to which she attached such significance was neither a shift of view nor a concession. The question or questions which elicited it ought never to have been put to Mr Townsend since it, or they, invited him to express a view on a conclusion of mixed fact and law which it was the task of the judge to reach. Moreover the question was put upon a false premise, that identified by Toulson LJ in Milton Keynes, that if arson was excluded then by definition the alleyway fire had to be regarded as, on the balance of probabilities, the cause of the house fire. In my view Mr Townsend was doing no more than giving it as his understanding of the law that if the judge rejected arson as a more (although he said less) implausible cause of the house fire than the fire in the alleyway, then the decision of the court, or “the finding”, would be that the alleyway fire was the cause of the house fire. Moreover the judge seems to have misunderstood his answer. In paragraph 22 she says that “the defendant’s own expert takes the view the cause of the house fire, on the balance of probabilities, was the alley fire.” With respect, he did not. The judge went on to say that “the only realistic possibility proffered by Mr Townsend was arson.” In the context of the relevant questions and answer that was not accurate either. The question of law had been posed of Mr Townsend in a manner which suggested that there were only two possible outcomes, ignoring the possibility that the Claimant had simply failed to prove the cause of the house fire.

  3. It follows that the process of reasoning which led the judge to conclude that the Claimant succeeded on causation was fatally flawed. The judge did not stand back and ask herself the ultimate question whether she was satisfied that the suggested explanation was more likely than not to be true. She did not have regard to the significant gaps in the court’s knowledge brought about by the lack of any adequate forensic investigation in the immediate aftermath of the fire. She did not ask herself whether the case for believing that the fire was caused in this way was stronger than the case for not coming to that belief, always bearing in mind that she was not obliged to come to a conclusion at all, and that a permissible outcome was that the inadequacy of the investigation conducted on the Claimant’s behalf gave rise to a situation in which the Claimant was unable to prove on the balance of probabilities what had caused the house fire.

  4. I do not overlook that, having expressed her conclusion on the basis of Mr Townsend’s answer, the judge went on to reject the “arson” theory, and that there is no formal cross-appeal on that point. However the Respondents’ Notice does contend that had the judge properly considered the evidence, she would have held that the Claimant had failed to discharge the burden of proof on the issue of causation. The judge could not possibly have concluded, on the basis of the evidence before her, that the house fire had been deliberately set, and the Defendant had no case to that effect. Before expressing her conclusion on arson, the judge had already indicated that, but for the critical answer or answers from Mr Townsend, she would have acceded to Mr Demachkie’s submission that the Claimant had simply not proved her case. I infer therefore that she would have reached that conclusion, quite properly, without further consideration of the arson theory, which was simply unavailable to her as a proved cause. The judge had no need to reach a positive conclusion either way on the arson theory, and she had no evidential material on which to do so. The judge was I think drawn into expressing an unnecessary conclusion about arson by the formulation of the question to Mr Townsend, the answer to which she regarded as determinative of the case. On any view however the answer given by Mr Townsend was not one which could justify departing from the conclusion to which she had otherwise come, which was that the Claimant had failed to prove her case. I also bear in mind that the starting point for the discussion at trial was the fire experts’ agreement in their joint statement that “all potential causes of fire appear not to have been given full consideration and thus cannot be eliminated.” As Lord Brandon pointed out in The Popi M, at page 956, “there are cases… in which, owing to the unsatisfactory state of the evidence or otherwise, deciding on the burden of proof is the only just course for [the judge] to take.” That was in my judgment the position here.

  5. I have well in mind the caution which an appellate court should observe before reversing a trial judge’s evaluation of the facts. We were rightly reminded of the judgment of Clarke LJ in Assicurazioni Generali SpA v Arab Insurance Group [2002] EWCA Civ 1642, [2003] 1 WLR 577, which contains a masterly exposition of the approach to be adopted. But we are not here concerned with the judge’s findings of primary fact, or of the impression made by the witnesses. We are essentially concerned with a process of legal reasoning on all fours with that which the House of Lords considered in The Popi M.

  1. I would therefore allow the cross-appeal and dismiss the appeal on that ground alone.