FREE CHAPTER from ‘A Practical Guide to Insane and Non-Insane Automatism in Criminal Law – Sleepwalking, Blackouts, Hypoglycaemia, and Other Issues’ by Ramya Nagesh


Within the criminal law, we are often concerned with whether the accused actually committed the act with which they are charged. The accused may claim that a witness is mistaken, is lying or that someone prompted the accused to commit the act out of self-defence. The law of insane and non-insane automatism is concerned, however, with the accused’s state of mind and ability to control themselves at the time. The accused in these cases will claim that they were not mentally responsible for what their body did.

The law draws a fundamental distinction between automatism caused by insanity, and that which is not.

It is important to remember, first, that the word ‘insanity’ in this context is very much a legal concept. It does not always correlate with what the medical community would term ‘insanity’. The law has imposed a far broader definition: a defect of reason caused by a disease of the mind which renders the accused unable to know what they are doing or to appreciate that it is wrong.

One of the ways in which the law seeks to distinguish whether a condition is insane or non-insane automatism is to identify whether the cause of that automatism is an internal or external factor. Internal causes are said to be indicative of insane automatism. External causes are said to give rise to a defence of sane automatism.

A long-standing justification for such a distinction has been that internal causes are more likely to recur, and so the greater powers of disposal (as to which, see below) available after an insanity verdict would help prevent that. However, this argument is not always sustainable. For example, a sleepwalker is highly unlikely to commit violence whilst asleep. Having done so once, there is no medical evidence to indicate that the sleepwalker will be likely to commit violence again. Whilst sleepwalkers can be treated medically, they are often given lifestyle advice to control their somnambulism. It is difficult, then, to see why there would be greater justification to find that a sleepwalker was legally insane than to find that they were suffering from a form of non-insane automatism.

It is perhaps for this reason that the Court of Appeal in Burgess (1991) WLR 1206 moved away from the possibility of recurrence as a determinative factor at 1212:

If there is a danger of recurrence that may be an added reason for categorising the condition as a disease of the mind. On the other hand the absence of the danger of recurrence is not a reason for saying that it cannot be a disease of the mind.

That is not to say that the ‘internal’/’external’ distinction is worthless. In cases at either end of the automatism spectrum we can see how it can be useful. A person suffering from a long-term mental illness that causes him to be unaware of his actions is undoubtedly driven by an internal factor. A person who has been so distracted by a swarm of bees in his car whilst driving, such that he crashes the car, is undoubtedly triggered by an external factor. However, the difficulty arises within the shades of grey between the extremes. For example, can it be said that a sleepwalker who commits a violent act whilst asleep is guided by an internal factor (his proclivity to sleepwalk) or the numerous external factors that may have triggered the somnambulism (such as caffeine or external disturbances to sleep)?

These questions are important ones. The decision as to whether a particular person’s condition falls within insane or non-insane automatism has impact on the burden of proof at trial, and the outcome for the accused if the defence is successful.

We will delve into both in more detail. This chapter is merely a brief introduction to both concepts. With that in mind, I set out a quick reference guide to the key features of insane and non-insane automatism below.


The key features of a defence of insane automatism are:

  1. The burden of proof is on the defence;

  2. The standard of proof is the balance of probabilities;

  3. The claim must be supported by the written or oral evidence of two or more registered medical practitioners, at least one of whom is ‘duly approved’ (section 1 of the Criminal Procedure (Insanity and Unfitness to Plead) Act 1991);

  4. If the defence succeeds, there will be a special verdict of ‘not guilty by reason of insanity’. The defendant will not necessarily walk free, however. He may be made the subject of a hospital order under section 5(2) of the Criminal Procedure (Insanity) Act 1964.


The key features of a defence of non-insane automatism are:

  1. The defendant bears only an evidential burden. Whilst merely raising the defence is not enough, he needs only to introduce such evidence as might leave a reasonable jury in reasonable doubt as to whether he was in the state alleged. It is then for the prosecution to disprove the defence to the usual criminal standard;

  2. If the automatism is self-induced, the defendant’s liability will turn on the principles relating to intoxication;

  3. If the defence is successful – i.e. the prosecution cannot disprove automatism – the verdict will be one of outright acquittal. There will be no scope for hospital order or other form of detention.