Most societies recognize that it would be wrong to execute someone like Lennie Small, the mentally disabled character in John Steinbeck’s 1937 novella Of Mice and Men, even though he murdered a woman while stroking her soft hair.

Lennie never understood what he did wrong, and that ignorance usually brings protection from the full force of the law. Most countries with a death penalty have some sort of special treatment for the mentally disabled enshrined in the judicial system.

Few intellectually disabled people are as obviously impaired as Lennie, so expert assessments — and science — are usually used to decide their fate. Most US states, for instance, use an IQ test to assess cognitive skills such as problem-solving and anticipating the consequences of actions. Lennie would have scored low: when the woman screamed, his reasoning power was so limited that he could come up with no other option but to kill her. The tests can provide an accurate measure of some cognitive skills and, better yet, seem to offer an objective metric for prosecutors to work with. No test of cognitive ability, however, can determine a person’s understanding of guilt, and thus their culpability for a crime. That problem becomes especially difficult when defendants have a mild intellectual disability.

Faced with such cases, some court systems use IQ score as a proxy to assess the deeper issue of awareness. Florida is one, and early next month its controversial approach will be tested. As we report on page 284, on 3 March the US Supreme Court will begin hearing arguments on behalf of Freddie Lee Hall, a convicted murderer. Hall has a low IQ, but not consistently low enough (below 70) to escape the death penalty in Florida. He and his lawyers want the state to raise its IQ cut-off point.

The state has refused. A weakening of its criteria, officials say, could prompt hundreds of appeals. One estimate suggests that in the United States, up to 20% of the 3,100 or so people on death row may have some level of intellectual disability (R. Coyne and L. Entzeroth Geo. J. Fighting Pov. 3, 40; 1996). And if it relaxes its strict interpretation, the state worries, could not a clever lawyer or sympathetic psychiatrist claim that a client facing the death penalty has a mental disability due to post-traumatic stress disorder, temporary insanity or a bout of depression? Defendants, lawyers and officials in other states are watching with interest.

Too little is known about how intelligence plays into criminality.

If the United States is to have a death penalty — and 55% of Americans supported it in a 2013 survey — it should ensure that all defendants have an equal, objective chance to save their own lives. Many states try to ensure this by drawing a ‘bright line’ at an IQ score of 70. But this greatly overestimates the IQ test’s precision. The tests have a ten-point margin of error — they cannot necessarily distinguish a 71 from a 69. If IQ tests were to be scrapped as a way to judge criminal competence, what could replace them?

In its latest version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), the American Psychiatric Association changed both the definition and the name of intellectual disability, formerly known as mental retardation. It now avoids setting any IQ limit for the disorder, and emphasizes the impact of cognitive ability on behaviour.

Related to this approach is the adaptive behaviour test. Designed to measure how well a person can manage in the real world by quizzing his or her family and acquaintances, when administered by experts, this standardized test gives the kind of consistent, numerical results that prosecutors crave.

Psychologists in the United States are already designing a modified version called the Diagnostic Adaptive Behavior Scale, the first evidence-based, adaptive behaviour test designed specifically for young people with a low IQ. Relevant to the debate over mental dysfunction and the death penalty, it assesses traits such as gullibility and the ability to solve social problems. Properly administered, it could determine awareness for courts better than existing tests of IQ.

Too little is known about how intelligence plays into criminality and the various environmental factors that affect it (such as decades spent in prison). If science is to provide courts with more certainty about the state of mind of defendants, then more research is needed on the nature of intelligence itself.

For example, Kent Kiehl, a psychologist at the University of New Mexico in Albuquerque, is compiling 3,000 brain scans of prisoners in what is already the largest collection of images of criminal minds in the world. Kiehl’s main aim is to assess factors such as psychopathy, or whatever it is that makes people commit crimes, but he also wants to develop a test to predict intelligence. Adding this to the arsenal of such tests could help to assuage prosecutors’ concerns about a defendant faking disability, or an expert giving a biased diagnosis.

In Steinbeck’s book, Lennie pays the ultimate price for both his crime and his disability. Justice demands that we separate the two. Science will keep trying to do so.