The Dangerous Dogs Act 1991 is both Delphic and obscure, and in urgent need of reform, says Robert Wade


'The Dangerous Dogs Act 1991 was, as is well-known, introduced in great haste by Parliament to deal with a number of unpleasant incidents in which a number of fierce dogs had seriously injured small children. It is a piece of Delphic legislation which is even worse than some of the directives coming out of Europe.'



Thus spoke Judge Popplewell in the case of Rafiq v DPP [1997] JP 412. What he said about the genesis of the act is undoubtedly true. It was aimed principally at dogs specially bred for their savagery and fighting instincts, and sections 1 and 2 of the act are directed to that end. His justifiable remarks about the obscurity of the act were made in the context of a prosecution under section 3(1), and might have led one to hope for clarification. Sadly, as a recent case of my own amply proved, his judgment, and that of Lord Justice Auld, who sat with him in the Divisional Court, only served to compound the confusion.



Section 3(1) of the act provides that 'if a dog is dangerously out of control in a public place' the owner and (if different) the person in charge of the dog is guilty of an offence. The offence does not require proof of actual injury, although, if any person is injured, the offence becomes triable 'either way', carrying a maximum sentence of two years imprisonment in the Crown Court.



When the bill was first mooted, Minister of State Angela Rumbold said that it had to be 'clearly understood' that liability to the criminal law would come to those who 'allow their dogs to cause injury to others'. That suggests it was intended that the prosecution had to prove mens rea. However, in R v Bezzina [1994] 3All ER 964, the Court of Appeal decided that the offence under section 3(1) was one of strict liability. According to Lord Justice Kennedy, the normal presumption that the prosecution had to prove mens rea was displaced because the statute was concerned with an issue of social concern in which public safety was involved, and where the creation of strict liability would promote the object of the statute by encouraging greater vigilance to prevent the commission of the prohibited act.



However, what is the 'prohibited act'? Before deciding whether mens rea is required, a court first needs to identify the actus reus. Obviously, the dog has to be in a 'public place'. What else is required? Logically, one would think, the court has to ask itself two questions. First, was the dog dangerous? Second, was it out of control? Yet Lord Justice Kennedy clearly assumed that that second question was irrelevant. Section 10(3) states that 'for the purposes of this act a dog shall be regarded as dangerously out of control on any occasion on which there are grounds for reasonable apprehension that it will injure any person.' Therefore, he said: '...when one encounters the words "dangerously out of control" one applies the meaning which is set out in section 10(3) and that means, in effect, that if a dog is in a public place... [and] is acting in a way that gives grounds for reasonable apprehension that it would injure anyone, liability follows.'



The flaw in this reasoning is that, whereas section 10(2) clearly defines two other expressions in the act, section 10(3) does not claim to provide a definition or meaning for the phrase 'dangerously out of control'. Those words involve two entirely different concepts - dangerousness and lack of control. Arguably, section 10(3) only relates to the question of when the dog is 'dangerous'.



In Ross v Evans [1959] 2 All ER 222, a case under section 54(2) of the Metropolitan Police Act 1839, the court held that a dog on a lead was not 'at large' even if the restraint was ineffectual. Under the Dogs Act 1871, a civil complaint can be made about a dog which is 'not kept under proper control', and it is a question of fact, to be determined by the magistrates, whether or not a dog is 'under control' (Philip v Wright [1940] JC 9). Section 3(5) of the 1991 act, which clarifies certain powers of the court under the 1871 act, and to which Lord Justice Kennedy made no reference at all, refers specifically to 'measures to be taken for keeping the dog under proper control, whether by muzzling, keeping on a lead, excluding it from specified places or otherwise'.



In my case, the dog was tied to a lamp-post. I argued that it was not 'out of control', and therefore not within the terms of section 3(1). Yet according to Lord Justice Kennedy, it makes no difference whether the dog is under restraint or allowed to roam free. That may not be wholly unreasonable if the dog bites somebody, but it is worth remembering that an offence can be committed under section 3(1) where no injury is caused at all, or where the injury is caused other than by a bite, in circumstances in which a dog would not be regarded as 'dangerous' in any accepted sense - for example, by tripping up a passer-by by entangling itself in his legs, leaping up at someone with over-enthusiastic affection or, as in my own case, jumping up and accidentally butting a child under the chin, forcing a tooth through her lip.



Lord Justice Kennedy further confused the situation when, in deciding that the offence under section 3(1) did not require proof of mens rea, he claimed to find support in section 3(2), which provides a defence to an owner who has placed the dog in the charge of a person whom he reasonably believes to be a fit and proper person. Yet section 3(2) appears to relate only to the issue of control. It certainly supports an argument that the act imposes strict liability once it is proved that the dog is 'out of control', but that does not in any way preclude the requirement of mens rea as to the issue of dangerousness.



The judge also found support in section 3(3), which imposes liability on an owner who 'allows' a dog to enter a place which is not a public place but where it is not permitted to be: 'In section 3(3) one can see that where Parliament wishes to introduce [mens rea] by having the word "allows", it was quite capable of doing so. It could just as easily have put in section 3(1) that if a dog is dangerously out of control in a public place, the owner, if he has allowed it to be so etc, but it did not do so. And the omission of the words that qualify the absolute duty seem to us to be persuasive.'



Surely, again this is a non sequitur? Bezzina was not concerned with whether the dog was in a public place or out of control - the case involved three separate appeals, in none of which the defendants attempted to argue those issues. The question was whether or not the prosecution had to prove that the owner realised that his dog was dangerous - a completely different issue altogether.



Driven from the argument that, in principle, criminal liability requires mens rea, the defendants had claimed that section 10(3) imported mens rea because there had to be a 'reasonable apprehension' that the dog would cause injury.



Lord Justice Kennedy said that 'section 10(3)... on the face of it... sets an objective standard of reasonable apprehension, not related to the state of mind of the dog owner'. That is not hard to understand. Other offences are treated similarly - an obvious example is driving without due care and attention. But they still require some objective evidence for the reasonable bystander to view. The judge had already said that the dog had to be behaving in such a way as to give grounds for reasonable apprehension that it would injure someone. What if - just as in my case - the dog was a perfectly amiable, biddable family pet which had never given the slightest prior warning that it might cause any injury?



The judge's reasoning is curious: 'Of course, if injury does result then, on the face of it, there must have been, immediately before the injury resulted, grounds for reasonable apprehension that injury would occur.'



This seems extraordinary. 'Apprehension' means 'uneasy or fearful anticipation'. If the dog gives no warning of a propensity to bite, how can there be - objectively or subjectively - 'grounds for reasonable apprehension' that injury would occur just because it actually does so? That is straining the English language to breaking point.



The issue was raised starkly in Rafiq, which concerned a German Shepherd dog with the perhaps unfortunate name of Venom. Mr Rafiq was the proprietor of a petrol filling station. A customer found Venom loose on the forecourt, and let him into the shop; he repaid her by biting her in the thigh, causing a small puncture wound.



On appeal from the magistrates, the Crown Court found as a fact that Venom showed no sign of aggression prior to inflicting the injury, but felt obliged to convict, following Lord Justice Kennedy's views in Bezzina. Before the Divisional Court, the defendant argued that the mere fact that injury had actually been caused did not justify a finding that the dog was 'dangerously out of control', and that, if that had been the intention of Parliament, it would have said so in section 10(3).



Judge Popplewell said that argument was illogical because it meant that the defendant would not be guilty of an offence if his dog bit somebody without any prior warning or reasonable apprehension, whereas he would be if his dog displayed signs of aggression but did not cause any injury at all, so that 'in the absence of forewarning, in one sense the more dangerous the dog, in the sense that he has bitten, the less responsibility... there is on the owner'.



However, he still had to explain the words 'reasonable apprehension'. His solution? 'In order to impose some logic... the proper way to approach these cases is to take the view that if there is a bite without reasonable apprehension immediately before that, the use of the words 'any occasion' is sufficient to impose a liability because there are grounds thereafter for reasonable apprehension that it will injure some other person.' (My italics)



With respect, this appears neither logical nor in accordance with the wording of the statute. Lord Justice Auld said he did not agree with that conclusion. He also had 'some difficulty' with the proposition of Lord Justice Kennedy in Bezzina: 'Depending on the circumstances, the time for apprehension, even by the notional reasonable bystander, may be so minimal as for practical purposes to be non-existent. The notion of reasonable apprehension of injury before it occurs in such circumstances, is artificial and the court should strain against adding that unhappy element to an already difficult statutory formulation. It seems to me that Lord Justice Kennedy in that passage was unnecessarily focusing on the injury as if it were a necessary culmination and demonstration of anterior reasonable apprehension of injury.'



Nevertheless, he upheld the conviction because 'the act of a dog causing injury, a bite or otherwise, is itself capable of being conduct giving grounds for reasonable apprehension of injury'. This reasoning, although certainly different from the views of Judge Popplewell, appears indistinguishable from that of Lord Justice Kennedy. It is unsatisfactory to have three judges openly disagreeing with each other, but engaging in such semantics to attain a desired result.



Liability under the criminal law must be based on certainty, not the opinion and good sense of an individual police officer or Crown Prosecution lawyer, especially where the defendant could face a prison sentence. A review of the Dangerous Dogs Act is long overdue.



Robert Wade is an associate solicitor with Keith Evans & Co in Cwmbran