Listen to my discussion of sexsomnia and the criminal law on ABC Radio National’s The Law Report with Damien Carrick here.
Over the last decade, a number of Australian defendants accused of sexual offences have pleaded the ‘defence’ of ‘sexsomnia’ in response to criminal proceedings instituted against them.
Just last week, a Sydney man charged with rape was acquitted on these grounds. The criminal complaint alleged the man had engaged in non-consensual sex with a woman who had visited his apartment. Within hours of asking the trial judge for certain clarifications, presumably about the rules governing the defence, the jury returned a verdict of not guilty.
Other cases have occurred in recent years, including a significant appellate case in 2022 (canvassed below). Another, decided in late 2024, involved a youth minister being acquitted of multiple charges of sex offences against three minors.
Though still rare, the apparent rise of the sexsomnia defence has raised concerns both in Australia and abroad.
In the United Kingdom, legal commentators have expressed fears that certain defendants may be ‘evading justice’ and lamented that prosecutors may have discontinued prosecutions where defendants have submitted the defence.
Is the sexsomnia defence a denial of the sexual acts?
Typically, defendants who raise the sexsomnia defence do not dispute the allegation they engaged in the charged conduct. Nor do they claim, as is more commonly argued, that the complainant consented to the sexual acts. Rather, the sexsomnia defendant generally pleads that their conduct was involuntary because it occurred while they were asleep.
The legal contention is that they are not guilty because the sleep disorder deprived them of their will — that is, of their ability to exercise a choice to perform or not to perform the acts in question.
Such defendants usually also report having no memory of committing the acts.
But what is the sexsomnia defence?
The sexsomnia defence is a subcategory of the well-established ‘common law’ defence of automatism. (The ‘common law’ consists of rules or ‘precedents’ created by courts — not parliaments.)
The automatism defence can be traced to the 1840s and, since its inception, it has always been entwined with sleep disorders such as ‘somnambulism’ (sleep walking). Other ‘automatistic’ acts for which the defence has been raised, however, include reflexes, spasms, convulsions, hypoglycaemia and epilepsy.
The so-called ‘defence’ of automatism emanates from the ‘ancient rule’ that people must not be held criminally responsible for involuntary acts.
However, certain debates have arisen in the law of automatism, which mirror those in psychiatry and sleep medicine, about how best to classify the conditions under which people commit involuntary acts.
In both domains, the key issue is whether sexsomnia is a transient condition that occurs in an otherwise ‘healthy mind’ (as a defendant must claim if they seek an outright acquittal) or is instead a mental health impairment that may pose a continuing danger to the person affected and those around them.
When was sexsomnia first recognised as a clinical disorder?
Although associations between sleep and sexual activity have been studied for centuries, the first case of sexsomnia (involving unconcious masturbation rather than intercourse) was not reported until 1986.
More cases were identified in the 1990s and, in 2005, the International Classification of Sleep Disorders recognised several ‘confusional arousals‘ as ‘parasomnias’ or sleep disorders.
2007 saw the first attempt to classify sleep-related disorders and abnormal sexual activities.
But it was not until 2013 that ‘sexsomnia’ appeared in the Diagnostic Statistical Manual of Mental Disorders (DSM-5), alongside sleepwalking and night terrors.
Just how common is sexsomnia?
While the exact prevalence of sexsomnia is unknown, it is considered uncommon. A 2020 study identified only 116 clinical cases, and no study of the disorder’s prevalence in Australia has been published.
Some researchers have suggested that sexsomnia is underreported due to a lack of knowledge about the condition and embarrassment about its occurrence. However, as the courts have recognised, there are several triggers, as well as several effective treatments, for the condition.
But how can sexsomnia be proved?
To raise the sexsomnia defence, a defendant generally needs to present detailed medical evidence. But defendants bear only an ‘evidential burden’ in this regard, which requires them simply to demonstrate the ‘reasonable possibility’ that their acts were produced by sexsomnia.
Given that sexsomnia is a species of ‘sane automatism,’ it is worth noting that such a category of impairment has been defined as a ‘transient non-recurrent malfunction of an otherwise sound mind.’
In the reported case law, experts called by defendants to give evidence tend to be somewhat familiar with the legal terminology of sexsomnia. However, the psychiatrists are not legal experts and can also confuse matters for the legal adversaries and the courts. For instance, while a psychiatrist may agree that sexsomnia involves involuntary acts (so as to satisfy the defence of ‘sane automatism’), they may also agree that it is a brain dysfunction and a mental disorder — and they may even indicate, under examination, that the disorder satisfies the language of a statutory provision.1
In such circumstances, however, the judge or jury must decide, as a question of law, whether the psychiatric evidence meets this or that legal classification. Indeed, as long ago as 1978, Glanville Williams said as follows:
Because automatism is a legal concept, a psychiatrist should be asked to testify to the mental condition as psychiatrically recognised, not to ‘automatism.’ It is for the judge to make the translation.2
To override the automatism defence, the prosecution must ‘disprove’ the sexsomnia evidence. It must prove, to the high standard of ‘beyond reasonable doubt‘, that the accused’s acts were voluntary and deliberate rather than the converse. But this can be very challenging.
While the prosecution might cross examine the defendant’s expert witnesses, or introduce their own expert witnesses and reports, once that defendant’s medical evidence has been heard by a court, the burden lies heavy on the prosecution to remove any reasonable doubt that the charged acts were involuntary.
Is sexsomnia the product of a sound mind or a mental illness?
The central controversy in a recent sexsomnia criminal appeal case was not whether the defendant’s acts were voluntary but instead whether the psychiatric evidence would allow the court or jury to conclude that the defendant has an ongoing mental health impairment.
Indeed, in this important New South Wales case of 2022, the prosecution accepted that the defendant — a man accused of sexual offences against his daughter — had a form of sexsomnia when he committed this abuse.
The prosecution case focused on the man’s condition, contending that it was not a transitory malfunction in ‘perfectly healthy mind‘ but a mental health impairment.
At the time, this argument responded to new mental health impairment laws in New South Wales, which defined ‘disturbances of volition‘ as mental health impairments.
The legal implications of this appeal case were widely understood at the time to be significant — not just for New South Wales but all state jurisdictions in Australia.
Why would it matter if sexsomnia were viewed as a mental illness?
If the prosecution was able to persuade the court that the psychiatric evidence established that the defendant had a ‘disturbance of volition’ — or, in the language of the common law, ‘insane automatism’ — then the court would likely be unable to order an acquittal.
In that case, rather, the judge or jury might be required to find that the defendant, not knowing the nature and quality of their actions, should be subject to a so-called ‘special verdict’ and be referred to a mental health tribunal for review, which in turn may result in their potential detention in a secure psychiatric facility, such as the Long Bay Hospital.
In the end, however, the prosecution failed to prove that the defendant’s sexsomnia was a ‘disturbance of volition.’ The majority of the NSW Court of Criminal Appeal found that, because the defendant was asleep, there was no ‘volition’ to be disturbed.
And although the prosecution later appealed to the High Court of Australia, that application was ultimately refused.
Has the law of automatism been criticised?
As long ago as 1966, legal scholars identified the problem with treating ‘sane automatons’ (such as sleepwalkers or sexsomniacs) as ‘harmless’ while the law also authorised, in some circumstances, the psychiatric detention of those with mental health impairments.
As one scholar wrote, ‘The propriety of treating an insane automaton as criminally demented, and a sane automaton as perfectly harmless can be strongly questioned.’
In a different vein, jurists have also criticised the complexity of the legal boundary that separates transient physical conditions from ongoing mental impairments, with one judge in 1990 describing it as ‘an intellectually unhappy and artificial distinction which at times produces both unsatisfactory and inconvenient results…’
However, this legal boundary appears to have only been reinforced by developments in psychiatry and medicine, where the line between sleep disorders and mental impairments seems only to have brightened in recent years.
Where to from here?
This is a complex area of law with a complicated history of medico-legal jurisprudence. But in view of the warranted concerns about the increasing use of the sexsomnia defence in serious, violent cases of sexual transgressions, it may be appropriate to consider the reform proposal advanced by esteemed mental health law scholar Bernadette McSherry no less than twenty years ago.
That proposal considered various legislative reforms, including one that would limit the scope and legal definition of automatism. Although it may be difficult to draft, such a reform might, as McSherry suggested, prompt courts to draw sensible distinctions between different types of conduct.
The effect of this intervention could be that courts would be prompted to demand more or different evidence from those defendants whose charges involve complex ‘goal-directed’ or ‘purposive’ acts but who yet contend those acts were involuntary.
Given that the defence appears to be difficult to disprove, a reform of this description might ensure that those cases whose facts were deserving of more scrutiny were able to be scrutinised appropriately.
- See, eg, R v DB (2022) 108 NSWLR 51, 71 [56] (Brereton J), 80 [111] (Wilson J). ↩︎
- Glanville Williams, Textbook of Criminal Law (Stevens, 1978), cited in R v Rabey (1981) 54 CCC (2d) 1, 30 and Stanley Yeo, ‘Putting Voluntariness back into Automatism’ (2001) 32(2) Victoria University of Wellington Law Review 387, 397 n 36. ↩︎
