The Stilnox defence: automatism or amnesia?

Stilnox, a branded version of the generic drug Zolpidem, is a medication sometimes prescribed for the treatment of insomnia. It acts quickly – subjects typically become sleepy just 15 minutes after taking it – which is great if you only have a few hours to catch some Zzz’s and must fall asleep pronto. Also, its effects wear off much faster than other sleeping tablets, and this is great not only if you have precious few hours to have that cat nap, but also because you’re less likely to feel drowsy the whole next day as you wait for the effects of the sleeping tablet you took last night to wear off.

However, in rare cases Zolpidem can apparently also cause sleep-walking, and, while sleep-walking, people sometimes commit crimes — for instance homicidal somnambulism (killing someone while sleep-walking). Go on, Google “homicidal somnambulism” right now. And now Google “stilnox defense”. Wicked, right? Google probably just spewed up a bunch of bizzare legal cases in which people took Zolpidem for insomnia; they then sleep-walked; while sleep-walking they evidently committed some crime; but then they got off scott free because they committed that crime while sleep-walking.

Admittedly, we might sometimes have good reason to suspect that a criminal defendant is lying when they claim that they committed their crime while sleep-walking. Heck, how convenient,… it if reduces their sentence, why not lie, right? (I’m reporting, not endorsing.) Or we might be suspicious about whether Zolpidem is what caused them to sleepwalk. Perhaps they’re just trying to find someone else to blame whom they can later sue for damages,… who knows! However, if the criminal defendant’s veracity is not doubted, then at least from one perspective the Stilnox defence seems reasonable. In essence, it’s a form of an automatism defence, since what the accused alleges is that their body’s movements should not even be viewed as actions (let alone as their actions) because they were not controlling their body’s movements (heck, they were not even aware of those movements since they were after all asleep), and hence that they should not be blamed for whatever their body did in their absence. So far, so good…

But what has started to worry me of late is how we can be certain that what we have in such Stilnox defence cases is indeed automatism? How do we know that we’re not in fact dealing with amnesia? After all, one of the side effects of Zolpidem is amnesia, so it is not inconceivable that that’s what Zolpidem does — that it makes people forget the things that they did, rather than making them run on autopilot like zombies. Now, I do not know the answer to this question — i.e. whether Stilnox defence cases involve automatism, amnesia, or sometimes one and sometimes the other. It’s an empirical question, and if you know of empirical studies that have investigated this question then please email me, k? But the reason why this worries me is because it is far from clear to me why amnesia should even be viewed as a defence?! For instance, why should we think that someone who has forgotten the murder that they committed – perhaps they murder someone but then they trip over, hit their head hard on the pavement, and forget having committed this crime – is less blameworthy and less deserving of punishment than someone who remembers it clearly? Why should memory matter?!

In a recent paper Annette Dufner argues that in cases of dementia-induced amnesia we might have good grounds to stop punishing (or to not start punishing) a person convicted of a crime. To me Dufner’s argument seems fair. My main objection is that even if a person can’t remember having committed a crime, it might still be reasonable to punish them for it if they can nevertheless be reasonably expected to accept our incontrovertible evidence (e.g. CCTV footage, perhaps) that they indeed committed that crime. I would view such assisted remembering as being on a par with normal remembering, though admittedly this may seldom apply to cases that Dufner discusses. But is Zolpidem-induced amnesia analogous to dementia-induced amnesia, at least for the purpose of criminal blame and punishment? Somehow to me the cases seem very different.

Furthermore, to my mind automatism seems like a water-tight defence, but amnesia… not so much. It would thus be great to see some empirical studies about what goes on in the brains of people who sleep-walk while on Zolpidem. For instance, if we found that Zolpidem interferes with memory consolidation, that might support the amnesia hypothesis. On the other hand, we might find that Zolpidem interferes with the normal mechanisms of conscious behavioural control, and that might support the automatism hypothesis.

My point is simply that if all that Stilnox does is to wipe away your memory of doing something, rather than to wipe you away temporarily and replace you with an autopilot, should people still be entitled to cite the Stilnox Defence? Or, should we develop better tests to help us distinguish these cases from each another? After all, everyone might sincerely believe that they sleep-walked while committing their crime, but only some might be right about this.