This appellant was convicted of counts of rape and common assault upon his partner and of a minor offence of criminal damage to her house. There was clear evidence that at the time of the offences he had been mentally ill, affected by paranoid schizophrenia and harbouring a number of delusional beliefs. His appeal certainly raises the question what if any impact his mental illness had on the issues before the jury. It is said more generally to raise the question whether, when considering the issue of a defendant’s reasonable belief in the complainant’s consent to sexual intercourse, account can or cannot be taken of the mental condition of the defendant.

The defendant and his partner had been in a relationship since about 2004. He is Algerian and she is British. They have a daughter, aged about two at the time of the events which led to the trial. This daughter had been born in, and remained in, Algeria, apparently having no passport. The couple was living in his partner’s house in Chelmsford.

In 2009 the defendant’s behaviour deteriorated. He became argumentative and aggressive towards his partner, of whom he was possessive and jealous, accusing her of paying attention to other men. In November 2009, on three occasions he assaulted her. It appears that on one occasion he became aggressive, thinking that there was something in the iron which was making him ill. He spat in her face, slapped her and pulled her hair. The available psychiatric evidence at that time was that he was not suffering from any mental illness, although he showed some signs of inappropriate behaviour in interview. Nor were the assaults thought to be the result of any mental illness. In due course in March 2010 he pleaded guilty to assaults and was made the subject of a suspended sentence, with a requirement that he live apart from the complainant. The trial with which we are now concerned was the result of a second prosecution a few months later. With hindsight, his illness was clearly developing.

The couple evidently remained on terms of some affection, as may often happen, not least when they have a child together. As well as remaining fond of him, his partner wanted his help in enabling her daughter to leave Algeria. On 19 July 2010 she supported his application to the court to lift the requirement that he live apart from her, and that condition was removed.

Most of the events which led to the second prosecution of him occurred within about a fortnight of that lifting of the condition, although it later turned out that one had occurred just a few days before the application to the court had been made. After a number of incidents, the complainant went to the police station to complain on Wednesday 5 August 2010. The result was his prosecution for counts of rape, common assault and a single count of criminal damage. We should record that he was acquitted of two counts of rape and one of common assault.

The defendant did not give evidence at his trial, although he was fit to do so. He had been interviewed by the police and the complaints put to him, some rather generally and some with particularity. He had provided some answers. Those, in effect, constituted the basis of his case at trial, together with some suggestions put to the complainant in cross examination.

The allegations in the counts of which he was convicted were, taking them in chronological rather than indictment order, the following.

On Friday [a date in] 2010 (which was the Friday before the Monday of the application in court to remove the non-residence condition) the couple had been going out for a meal on the occasion of the complainant’s birthday. The defendant objected, for no rational reason, to the complainant having passed the time of day with their (male) next door neighbour, who had been by his bins as they left the house. The defendant remonstrated with her in the car, and when it stopped at some traffic lights he turned towards her and spat in her face.

As to this allegation, D told the police that no such incident had ever occurred and that he had no problem with his partner speaking to the neighbour.

On Wednesday 4 August 2010 the defendant made the complainant eat a bowlful of cold tinned peas in their canning fluid, mixed with crumbled dried leaf from the apple tree in the garden. She did not want to eat it. He was agitated and insisted. She complied but tried to pick out the leaf pieces. He came across the room, seized her finger, and made her put the piece of leaf back and eat the mixture. He made sure that she swallowed it.

In interview the defendant accepted that he had given the complainant this mixture. He agreed that she did not want it. He said that he had said to her “Have it”. He said that he had stood there, and told her that the leaf was oregano and that the mixture would make her feel well. He denied, however, that he forced her to eat it.

Very soon afterwards that same evening he told her that he wanted sex. Her evidence was that she did not want to do it and did not want to go to the bedroom. He insisted. In the bedroom he sprayed her private parts with a mixture which he had in a spray bottle, which smelled of bleach and caused a burning sensation, and then had intercourse with her in a rough manner. She agreed that she had undressed herself before the act of intercourse. In effect her evidence was that she objected but submitted in the face of his insistence.

In interview the defendant was asked generally about spraying the complainant. He was alleged to have done it also on another occasion when it fell on her nightdress. He said that he had a spray which he used regularly to scent the bed but said that it did not contain bleach. He said generally that he had never had intercourse with her unless she consented.

At the trial the case put to the complainant was that there had indeed been this act of intercourse but she had consented. The case put as to the spray was that the defendant had told her that the spray would make her (not, it seems, the sheets) clean.

Sometime during the same night, or perhaps next morning, when he woke up, the defendant had wanted sexual intercourse again. The complainant’s evidence was that she said no but he insisted. He was, she said, far stronger than she. Generally, she said that he would not take no for an answer. She agreed that, as on the previous evening, she had removed her nightdress herself before submitting to sex.

The case put for the defendant as to this was the same as for count 3. It was accepted that intercourse had taken place but it was contended that she had consented.

The complainant said that she had once come home from work to find that the defendant had dismantled the front doorstep. That, she said, was all of a piece with other occasions on which he had insisted on cutting down trees in the garden. She said that she had not agreed to this being done.

When interviewed the defendant said that he had spoken to her in advance about the doorstep and she had consented to what he had done. He said that he had said that it would be good if the step were lowered. That was the case pursued at trial.

The trial took place in November 2011, a little over sixteen months after these events. The complainant’s evidence as to the counts of which the defendant was convicted was as summarised above. Although the defendant did not himself give evidence, one of the several consultant psychiatrists who had assessed him was called, Dr Volkanskaia. She had not treated him but had seen him on two occasions, on 25 July 2011 and again on 15 November 2011, just a few days before the trial. She also had access to the records of his assessments in 2009 and of his treatment in the period since his arrest in August 2010 in connection with these events.

Dr Volkanskaia’s evidence amounted to this:

The relevant part of the law of rape is now set out in s1(1) & (2) Sexual Offences Act 2003:

The principal issue on the rape counts here was plainly whether the complainant had in fact consented or not. The line between reluctant consent and submission despite lack of consent is often a fine one, especially in cases of an existing sexual relationship, and it was so here. The complainant told the court that there were other occasions of consensual sexual intercourse during the period the jury was considering. There were two other counts of rape relating to other alleged acts of intercourse, in respect of which the jury acquitted. It must, however, by its verdicts, have found that in the cases of counts 3 & 4 she was not consenting. Given the immediately prior incident of forced feeding of a strange and unpalatable mixture, as well as her evidence of these acts of intercourse, there was clearly a proper basis for such a finding.

Since the defendant did not give any evidence, there was no direct evidence that he had believed that his partner was consenting. But he had said generally to the police that he never had intercourse with her without her consent, and it is of course the law that it is for the Crown to disprove reasonable belief in consent when once the issue is raised. Given his illness, the question of his belief was sufficiently raised to require the Crown to do so.

The judge was invited to direct the jury that if and when it came to considering the reasonableness of any belief by the defendant that the complainant was consenting, it was entitled to take into account his mental condition at the time. After argument and careful consideration, he declined to do so. He gave the jury an admirably clear and untechnical direction that the mental condition was not to be taken into account. He said:

Mr Spence QC submits that this was wrong. The judge ought, he says, to have directed the jury in very general terms that the question of whether the belief was reasonable was whether it was reasonable in the particular circumstances of the defendant, which included his mental illness. The judge should, says Mr Spence, have given no other or more specific guidance, leaving the matter to the jury at large.

Mr Cray for the Crown submits that the judge was right. A delusional belief may, he says, be a genuine one, but it is by definition an irrational and thus an unreasonable one.

There was and is no authority on this topic to guide the judge. The question was adverted to by this court in R v MM [2011] EWCA Crim 1291, but it did not there call for answer because it had not, and could not have, arisen at the trial. The psychiatric evidence in that case, which the defendant chose not to rely upon, would have been that the defendant thought that the complainant was unjustified in saying no, not that he believed her to be consenting, but that he might satisfy the criteria for insanity – see paragraphs [10], [15] and [49] to [54].

We were taken de bene esse to the report of the Home Affairs Committee of the House of Commons at the time that the Bill which became the Sexual Offences Act 2003 was passing through Parliament. At the time there was debate (inter alia) as to the form the consent provisions in the Bill should take. One proposal was that it should require a two stage determination. The first would be whether a reasonable person would have doubted consent. If yes, then the second stage would be whether the defendant acted in a way that a reasonable person would consider sufficient to resolve the doubt. That was not adopted (and would undoubtedly have been complex to apply). Nor was a further alternative formulation adopted, namely that the test should be what a reasonable person would have thought, if he shared the characteristics of the defendant . The Government’s view, expressed by Lord Falconer of Thoroton was reported as being “strongly” in favour of an objective element in the test for belief in consent, on the grounds that it was not unreasonable to require a person to take care to establish that a sexual event was consensual; the cost to him was very slight whilst the cost to a victim of forced sexual activity was very high indeed.

The debate is interesting, and entirely understandable, especially given the difficulties which had arisen in the slightly different context of provocation. There, the attempt to import into the standard of the reasonable man the characteristics of the particular defendant had been demonstrated to be impossible of achievement. It had led in the end to the concept, entailed in the decision in Morhall [1996] AC 90, of a reasonable glue-sniffer. It had undoubtedly proved unworkable, as well as wrong in principle, and the result was reversal by the decision of the judicial committee of the Privy Council in Attorney General for Jersey v Holley [2005] UKPC 23 [2005] AC 580 and the subsequent case of R v James [2006] EWCA Crim 14 in this court. (We will return to provocation and its modern successor later in this judgment.)

The Committee approved the formulation eventually adopted in the legislation. The briefly stated conclusion to the report makes it clear that it, or at least the majority of its 11 members, assumed that this formulation would enable the jury to assess the reasonableness of belief in the light of the individual characteristics of the defendant “such as learning disability or mental disorder”.

It appears that the Government, in its response to the Committee’s report, stated simply that it agreed that “it is for the jury to decide whether any of the attributes of the defendant are relevant to their deliberations, subject to directions from the judge where necessary.”

It is impossible to accept the invitation to us to treat these materials as an authoritative indication of the meaning of the Act. The Committee, distinguished as its membership was, represented but one small part of the Parliamentary process. The views of its few members may or may not have been shared by Parliament generally, which of course consists of two Houses. There had undoubtedly been extensive debate on the consent provisions of the Bill in the House of Lords. It is not legitimate to treat partial preparatory material of this kind as a substitute for specific words in the Act as passed. No one can know what the reasons of the many members of both Houses may have been for supporting the formulation adopted. The Act as passed simply has to speak for itself. That is the law which is accessible to everyone and by which we are all governed.

Quite apart from that difficulty of principle, neither the Report nor, still less, the Government response to it, provide a worked-through answer to the question raised by the appellant’s submissions to us. ‘Mental disorder’, the expression used by the Committee, covers an enormous range of conditions, whether it is being used generally or with the statutory definition contained in s 1 of the Mental Health Act 1983 in mind. The latter definition has been modified by the Mental Health Act 2007 since the Committee was deliberating, but both now and as originally enacted it includes not only mental illness but also a large variety of personality disorders of which psychopathic disorder is only one, and some of which are susceptible to limited treatment and many not. It also includes (for many purposes) learning disabilities. There is no sign that the Committee had to grapple with the question whether and when a mental disorder such as a psychotic delusional state or a psychopathic or antisocial personality disorder might be capable of inducing a “reasonable” belief in the consent of a complainant. The Government’s response, by indicating that the deliberations of the jury will necessarily be subject to the direction of the judge, even more clearly begs the question whether he will have to direct the jury that a characteristic, such as delusional belief that all women are irresistible to the defendant, or such as immutable inability to empathise with a woman’s wishes and feelings, is inherently unreasonable and cannot be relied upon.

On the facts of this case, we do not think that in the end the issue arises. The evidence of Dr Volkanskaia was that the defendant’s illness did not affect his ability to understand whether or not his partner was saying no. This was not a case where the issue was whether he might, through illness, have misunderstood subtle social signals which a man without his illness would not have misunderstood. The fact that Mr Spence skilfully elicited the evidence that he had an impaired ability to interpret events normally does not alter that position. He clearly did, at the very least at the time of these events, have such an impaired ability; his delusions that he was a healer with special powers and that he was being poisoned, and perhaps that his partner was taking an improper interest in other men, were all examples of his not interpreting events normally. Dr Volkanskaia’s evidence, assuming it were accepted, certainly left it as a possibility that he believed that he had good reason to oblige his partner to have intercourse with him, either because he believed it would heal her or because he believed she had been consorting with other men and he was irrationally jealous of her. But what Dr Volkanskaia was not saying was that his illness would or might have led him to think that his partner was consenting when she was not. She was saying the reverse. In the face of that evidence, and in the absence of any account from the defendant suggesting misunderstanding, Mr Spence’s submission to us that he may not have understood that the complainant was frightened and thus thought that her submission was consent can only be mere speculation. For these reasons alone, the convictions for rape in the present case cannot be considered to be unsafe.

If, however, we are wrong about that, and the defendant’s delusional beliefs could have led him to believe that his partner consented when she did not, we take the clear view that such delusional beliefs cannot in law render reasonable a belief that his partner was consenting when in fact she was not. The Act does not ask whether it was reasonable (in the sense of being understandable or not his fault) for the defendant to suffer from the mental condition which he did. Normally no doubt, absent at least fault such as self-induced intoxication by drink or drugs, the answer to that in the case of acute illness such as this defendant seems to have suffered will be that it is reasonable. What the answer would be if the condition were an anti-social, borderline or psychopathic personality disorder may be more problematic. But the Act asks a different question: whether the belief in consent was a reasonable one. A delusional belief in consent, if entertained, would be by definition irrational and thus un reasonable, not reasonable. If such delusional beliefs were capable of being described as reasonable, then the more irrational the belief of the defendant the better would be its prospects of being held reasonable. The possible condition of the defendant in MM (supra at [27]) may afford an example. It seems that he may have been suffering from a bi-polar affective disorder which disinhibited him and made him grossly impulsive. He apparently believed that he was entitled to have intercourse with his former girlfriend despite her objections because those objections were, to him, invalid. If, instead, he had believed, for the same reason, that she consented, the question would have arisen whether his belief was reasonable. Similarly, the beliefs of a sexual predator that all to whom he is sexually attracted welcome his attentions may be attributable to mental disorder, whether correctly labelled a personality disorder or otherwise, and whether susceptible to any form of treatment or not. So may the beliefs of those who excessively idealise or excessively devalue their partners –sometimes swinging abruptly from the one to the other.

Both the common law and statute law are well used to a rule which judges a defendant by his subjective state of mind. So, for example, in a case where self defence is at issue, the defendant is to be judged according to the facts as he genuinely believed them to be, whether his belief was reasonable or not, at least unless it was attributable to voluntary intoxication. Criminal damage, which arises also in the present case, is not committed if the defendant honestly believes he had (or would have had) the consent of the owner of the property damaged to do what he did, even if that belief was unreasonable. But the decisive indication as to the law of rape is, we think, that the Sexual Offences Act 2003 deliberately departs from this model. It deliberately does not make genuine belief in consent enough. The belief must not only be genuinely held; it must also be reasonable in all the circumstances. This was a conscious departure from the former law. Under the former law, a genuine belief in consent (reasonable or not) was a complete defence to rape; the reasonableness of the belief was material only as a factor to be considered en route to the decision whether it was genuinely held: see the Sexual Offences (Amendment) Act 1976.

We agree that Parliament has thus made the distinction analysed by Professor Ormerod in Smith & Hogan (13th edition at page 744):

We do not say that the law of what used to be called provocation raises identical issues to the law of belief in consent in the law of rape. For one thing, the former provides for a partial defence, and to murder alone. But it is apparent that the concept of a reasonable belief in consent entertained as a result of wholly irrational delusions would be as difficult to apply as the concept of the reasonable glue sniffer proved to be. It is noticeable that the law of this partial defence to murder, formerly called provocation, has now been reformulated in a different manner by s 54 Coroners and Justice Act 2009 and in a way which – tellingly - excludes from consideration those characteristics of the defendant which bear on his capacity for tolerance and self restraint. Thus mental illness reducing that capacity would not be relevant to the partial defence of loss of control, but would be material if it amounted to insanity or (in the particular case of a murder charge alone) fulfilled the different criteria for the separate (also partial) defence of diminished responsibility. This re-statement of the partial defence of loss of control is further support for the conclusion at which we have arrived as to the meaning of section 1(2) of the Sexual Offences Act 2003.

The submission which Mr Spence has to make as to the manner in which this issue should be left to a jury points to the same conclusion. His submission is that the judge ought simply to have told the jury that the mental illness of the defendant was a factor which they should (or could) consider when deciding whether any belief which he might have held in the consent of his partner was shown to be an unreasonable one. No further guidance should, he submits, have been offered. But that is to leave a jury entirely rudderless. There is no difficulty whatever with the general proposition that questions of what is or is not reasonable in all the circumstances often cannot be defined in advance and must be left to the individual judgment of the jury. But if the question is one such as this – whether beliefs induced by a delusional psychotic illness are capable of being reasonable or not – it is not simply individual to any single defendant but will recur from time to time in cases of similar illness. The answer needs to be the same if justice is to be done between defendants. Similarly, if the question arises in the context of a personality disorder, the jury is entitled to know how it should approach the issue of reasonableness, and the general approach ought to be the same in similar cases.

We conclude that unless and until the state of mind amounts to insanity in law, then under the rule enacted in the Sexual Offences Act beliefs in consent arising from conditions such as delusional psychotic illness or personality disorders must be judged by objective standards of reasonableness and not by taking into account a mental disorder which induced a belief which could not reasonably arise without it. The defendant’s mental condition, and its impact on his behaviour, is of course extremely relevant to sentence. If punishment is inappropriate, a non-custodial sentence may result when otherwise there would have been a substantial sentence of imprisonment, and whether a hospital order is needed by the time of trial or not. In other cases it may significantly mitigate the punishment required. In yet others, it may result in a substantial custodial sentence recognising the danger which the defendant presents.

It does not follow that there will not be cases in which the personality or abilities of the defendant may be relevant to whether his positive belief in consent was reasonable. It may be that cases could arise in which the reasonableness of such belief depends on the reading by the defendant of subtle social signals, and in which his impaired ability to do so is relevant to the reasonableness of his belief. We do not attempt exhaustively to foresee the circumstances which might arise in which a belief might be held which is not in any sense irrational, even though most people would not have held it. Whether (for example) a particular defendant of less than ordinary intelligence or with demonstrated inability to recognise behavioural cues might be such a case, or whether his belief ought properly to be characterised as unreasonable, must await a decision on specific facts. It is possible, we think, that beliefs generated by such factors may not properly be described as irrational and might be judged by a jury not to be unreasonable on their particular facts. But once a belief could be judged reasonable only by a process which labelled a plainly irrational belief as reasonable, it is clear that it cannot be open to the jury so to determine without stepping outside the Act.

For those reasons, even if we are wrong about the import of the psychiatric evidence in this case, and even if there was, contrary to our view, any room for the possibility that the defendant (a) genuinely believed that his partner was consenting and (b) did so because of a delusional state consequent upon his mental illness, the judge was correct, on the evidence in this case, in his direction to the jury.

The judge directed the jury as to common assault without reference to the mental condition. He said this:

The judge was right to decline to do so. It is not the law that one is entitled to force feed another adult because one thinks that the food would be good for her. At least absent such genuine beliefs as that the assault was necessary to save the life of the person assaulted, or to avoid attack by a third party, a wish to do good does not generally justify an assault. The element of assault frequently and usefully described as ‘hostility’ is a means of conveying to the jury that some non-hostile contact is an ordinary incident of life to which we all impliedly consent. Similarly we can be assumed, unless and until we say otherwise, to have consented to life-saving or rescue actions involving contact. The judge correctly explained this. To the extent that he told the jury that an unwanted touching done out of a motive of love or affection was not an assault, the direction was arguably somewhat favourable to the defendant, but for entirely understandable reasons on the facts of this case. In many cases, and this was no doubt one, we can properly be said to have assented to unexpected touchings by those we care for, so long as they are motivated by affection and not hostility. But in the end, if a person tells another that he or she refuses to be touched or struck in a particular way and the other carries on and does it, the fact that he is motivated by misdirected affection will not save it from being an assault. In this case, the forcible taking of the finger to make the complainant eat the mixture was, if it happened, the clearest possible assault. The issue was whether it had happened. The jury decided that it had. The direction was impeccable.

Consent is a defence to criminal damage. So is belief that the owner has consented. So, moreover, is belief that the owner would consent if asked, even though he has not been asked. The belief here in question is a genuine belief, whether reasonable or not. The law here is akin to the law of self defence. The defendant is to be judged on the facts as he honestly believed them to be, whether his belief is reasonable or not.

In the course of argument the judge at one stage entertained difficulty in speaking of a delusional belief as a genuine or honest one. But that was mere discussion and the clearing of the mind. It is plainly possible for a delusional belief to be genuinely held. Its irrationality is likely in many cases to be evidence suggesting that defendant did not hold it at all, but it will not be conclusive and may not carry much weight if the defendant is shown to have been suffering from a (genuine) mental illness which generated delusions. In the outcome, despite his original doubts, the judge acceded to Mr Spence’s invitation to direct the jury simply in terms of genuine belief. He told them: “If it might be the case that he thought she would have consented if she had known about it, then the Crown will have failed to prove the element of him acting without lawful excuse. They have to make you sure that she did not permit it, she did not give permission, and he knew jolly well that she would not give permission to dismantle the doorstep in that way, before you can convict on that count.”

That was correct. Mr Spence complains that he ought to have told them explicitly that that meant that the defendant’s mental illness could be taken into account in asking whether he might genuinely (but mistakenly) have believed that the complainant had consented to the dismantling of the doorstep. We think that that was apparent from the terms of the direction, although there was really scarcely any evidence laying the ground for the possibility that he did harbour such a belief if there had not been the explicit consent which he had asserted to the police. In the end, however, after hearing the summing up, what Mr Spence actually invited the judge to add was somewhat different. The judge accepted his request without material modification and told the jury additionally this:

It follows that this appeal against the convictions on the counts of rape, common assault and criminal damage must be dismissed.