What is the Bastard Verdict?
After closing arguments have been given and the jury has received its instructions from the court, all you can do is wait. This is a very anxious time, even if you've done a great job at trial and even if you aren't terribly nervous person by nature. Some folks get work done on other matters and God bless them for doing it. But for my part, I'm often found pacing about the courthouse or making small talk with opposing counsel while we wait for the jury’s determination. It's an odd thing to have a whole case, hundreds or thousands of hours of work over months or years come down to “liable” or “not liable” or “guilty” or “not guilty” scrawled on a verdict sheet.
At least in the United States it works that way. Until recently, Scotland had a much criticized third option, “not proven,” available to the jury. This third choice was famously cursed by Sir Walter Scott as the “bastard verdict,” when a jury returned that result in a poisoning case that he was involved with.
But where did this “bastard verdict” come from? And why did Scotland recently decide, after hundreds of years of history, to do away with it?
Early 17th century Scottish criminal jury trials differed substantially from the modern formulation. They had radically different (or in some cases non-existent) evidentiary and procedural safeguards such as access to counsel. But for all their differences, their ultimate results essentially tracked the modern practice. Either an accused was convicted and found liable for the crime or acquitted and found to be innocent.
But in the late 17th century this changed. Instead of being charged with determining guilt or innocence, Scottish juries were instructed to determine whether the facts alleged by the prosecution were “proven” or “not proven.” After this factual determination by the jury, the Scottish judge would then decide whether the “proven” facts justified a finding of criminal liability. While you can imagine a benign justification for this procedural change, in practice it appears to have been done to weaken the power of juries and permit the authorities to score convictions even when the laws were unpopular and the defendants broadly sympathetic. There is some academic dispute as to the precise cause, though it has been suggested that juries in the 1660s were acquitting Presbyterian Covenanters who were hostile to the Stuart monarchy, and that the proven/not proven replacement was instituted to allow for more brutal legal repression.
This proven/not proven era was fairly short lived. By 1728 Scottish juries were reasserting their right to determine guilt and innocence. But even once this was clearly reestablished, the “not proven” verdict remained available as a kind of vestigial procedural option for the jury. And so it remained in regular used for centuries, drastically longer than the odd political conditions that had produced it.
So, what does a “not proven” verdict even mean, when both guilty and not guilty remain as options? Well, it isn't altogether clear. Like jury nullification, the “not proven” verdict was a right that Scottish juries clearly had under the law, but that courts bent over backwards not to explain to them. As such, there wasn't a situation for which it was clearly appropriate under Scottish law. One could imagine, and Scottish juries often did, that the “not proven” verdict reflected a judgment that the accused was not innocent—that is to say, was probably guilty—but that some evidentiary threshold necessary to convict had not been met. But Scottish defendants are not required to prove their innocence. And a verdict that brands them as “probably guilty” but does not subject them to punishment seems quite unfair both to defendants and to society at large. Hence Sir Walter Scott's description of it as a “bastard verdict” that ultimately satisfies no one.
The only class of folks who really liked the “not proven” verdict were social science researchers, who treated it as a weird natural experiment that provided insight into human decision making. In particular, one issue potentially implicated by the “not proven” verdict is the notion of asymmetric dominance. Asymmetric dominance is a phenomenon where the introduction of a third—inferior—choice changes which of two other options is selected. For example, assume that a movie theater sells a small popcorn for $10 or a large one for $20. Assume that people break 60/40 for the small popcorn. Now assume that the theater introduces a medium option for $18.50 that is halfway in between the two sizes. This option is objectively terrible; it's more expensive than the small popcorn and a much worse price per unit of volume than the large popcorn. But even if almost nobody buys it, the existence of the medium popcorn may cause people to shift from 60/40 for the small popcorn to 70/30 for the large popcorn because the large popcorn now looks like “a good deal” in comparison. This is totally irrational, but it works frequently enough to merit study in the social science literature and to use as a sales tactic.
But it's one thing to use cognitive defects to sell more popcorn and quite another to use them to send folks to jail (or to put criminals out on the street). Over the past several decades, quite a lot of research has focused on whether the existence of a not proven verdict has an impact on the total balance of guilty and not guilty verdicts. For most crimes, it appears that the not proven option does not have a meaningful impact on the absolute conviction rate, and some percentage of not guilty verdicts (roughly one-fifth) appear to get converted into “not proven” verdicts. This likely occurs where the prosecution introduces some compelling evidence of wrongdoing by the plaintiff, but nothing that would prove guilt for a particular crime.
But there is one big exception and it's the reason that Scotland finally eliminated the “not proven” verdict after four centuries of use. Among western nations, Scotland is an extreme outlier in terms of the conviction rate for sexual assault. In 2022–2023 only 48% of rape trials in Scotland resulted in a conviction. In contrast England and Wales have conviction rates of almost 70%—still nothing to be proud of but vastly higher. There is also reason to believe that the “not proven” verdict plays a factor in this discrepancy. While only 22% of acquittals in Scotland involve “not proven” verdicts, the “not proven” verdict was given in 66% of rape and attempted rape acquittals.
Now it's hard to prove causation based on statistical evidence, particularly on a scale as large as a country. And it's possible that Scottish juries are just unwilling to convict defendants of rape under any circumstances. But the correlation was disturbing enough (and the conviction rate embarrassing enough) that the Scottish Parliament finally voted to abolish the “not proven” verdict in September of 2025. It's too early to tell whether this will make a difference in the conviction rate. And as it was included in a bill that contained other criminal justice reforms, it may also ultimately prove difficult to separate out the impact of any one change even if the conviction rates rise.
But apart from the impact on Scottish criminal justice, the fact that such an old rule can be done away with is a powerful reminder that the anachronisms that shape trials—both in the United Kingdom and America—are not beyond correction. The Scottish “not proven” option was venerable as trial rules get. It was substantially older than the right of criminal defendants to have counsel at trial at all. It was also dramatically older than every rule of evidence, including the very notion of hearsay. If the Scots can clear out the cobwebs and eliminate parts of their trial procedure that no longer spark joy, just maybe there is still hope for folks on our side of the pond too.
