Matthieu Bourel

Remember these numbers. You’ll be asked about them at the end of the test: 70, 73, 76, and 78.

These are the ages of the leading candidates in the 2020 presidential election: Elizabeth Warren, Donald Trump, Joe Biden, and Bernie Sanders, respectively. In most any other line of work, people in their eighth decade are usually retired. For most of human history—and still in most of the world today—people of this age were usually dead.

Last month, Jimmy Carter, the 95-year-old former U.S. president, said that the office requires a person “to be very flexible with [one’s] mind,” and that by age 80 he wouldn’t have felt able to do the job. He joined the growing ranks of those suggesting they would support an upper age limit for the office, either for purposes of breaking up the gerontocracy or to ensure a person has the physical and cognitive capacity. “You have to be able to go from one subject to another and concentrate on each one adequately and then put them together in a comprehensive way,” Carter said.

The relevance to the current political moment was significant. Many neurologists and psychiatrists have raised urgent concerns about Donald Trump’s inability to hold a basic train of thought. Compared with his public appearances in decades prior, his meandering speech suggests declines in ability to reason and form basic arguments. A decade ago, he was capable of uttering sentences as cogent as Carter’s.

Joe Biden stays focused on single topics and ideas much more consistently than Trump, but he is often slower to find words than he was in the past. He veers sometimes into nonsensicality. In September, in response to a question about racism, he wandered into, “Play the radio, make sure the television—the, excuse me—make sure you have the record player on at night … make sure the kids hear words.” After Bernie Sanders’s heart attack last month, news reports also brought concerns about age to the fore. While Warren seems to have been spared from such speculation, she would still be the oldest person ever elected to a first term. Ronald Reagan was reelected at 73, and the end of his second term was, by some accounts, marked by the early linguistic signs of the Alzheimer’s disease that he would announce five years later.

Concerns about politicians’ ages are not limited to presidential candidates. The average senator is 62. Mitch McConnell, the 77-year-old Senate majority leader, is already past the U.S.-male life expectancy. If the country were to deem, say, the Social Security retirement age of 66 as the mandatory cutoff for a career in politics, it would amount to a total overhaul of government. It would eliminate not just the top four presidential candidates, but much of Congress.

At a conceptual level, any hard cutoff introduces a host of problems. Beyond meriting charges of age discrimination, age cutoffs are difficult to defend on practical grounds. An age minimum might rule out, say, an illiterate 4-year-old, but it’s at least hypothetically possible that a 110-year-old might be perfectly able.

At the same time, for all the promise of modern science, it seems unlikely that the best approach to ensuring candidates’ abilities to execute the duties of the office is to rely on them to tell us they are fine. Bernie Sanders has insisted he is running a “vigorous campaign.” (He has invoked vigor often.) Joe Biden has reminded crowds that with his age comes wisdom. Elizabeth Warren has vowed that she would “outwork and outlast” her opponents. Donald Trump said in April, “I’m so young. I can’t believe it, I’m the youngest person.”

Beyond such first-person reassurances of health, the process for outside assessment is opaque, optional, and arbitrary. It does not need to be.


The concept of “fitness for office” dates to the assassination of John F. Kennedy. The transfer of power to Lyndon Johnson raised a hypothetical question: What if Kennedy had survived, but in an impaired state? What if a bullet had grazed his frontal lobes and he was only briefly hospitalized, but thereafter he seemed unable to hold a train of thought? What if he started making bizarre comments and took to denying the existence of the moon? Such gray areas were the basis for the ratification of the Twenty-Fifth Amendment, in 1967. It created a way of removing a person who became unfit to serve.

The amendment has never been acted upon. Instead, the idea of fitness for office has manifested itself in the customary presidential disclosure of medical records and an annual physical exam. That process, though, is perfunctory at best. After Trump was elected, his physician during the time he ran for office, Harold Bornstein, claimed that the president had dictated his own glowing health assessment. “It was very obvious from the start that that examination had no credibility,” says Mark Fisher, a neurology professor at UC Irvine. “It was unprofessional and unacceptable as an assessment of his health.”

What do citizens really need to know about their president’s health? A physical exam by a doctor is designed to test for diseases, but the real question of fitness does not seem to be whether a person is disease-free. Most Americans have been diagnosed with some chronic disease. Compared with the days of Kennedy, modern medicine, communication, and transportation make it possible to climb the political ladder while living with many different conditions. Vice President Dick Cheney suffered from terminal heart disease, then received an entirely new heart and continued on in the political milieu.

The fundamental issue when it comes to fitness, according to Kathryn Bayles, a professor emeritus at the University of Arizona, is not age or physical health, but cognitive ability. This correlates only very roughly with age: Everyone’s abilities decline later in life, but the rate of decline varies widely. The degree of impairment also depends on how much you had to lose. And mild cognitive impairment often starts in a person’s 60s, but many people remain sharp to the very end. “People over 65 are the most cognitively diverse segment of the population,” says Bayles, who studies speech, language, and hearing. “Look at Nancy Pelosi, who is 80. She is articulate, measured, sharp.”

Fisher estimates that simply based on population averages, we can assume that a significant proportion of political leaders have some degree of impairment of executive functioning, which includes things like decision making. But standard medical approaches such as blood tests or MRI or CT scans have not proved to be effective ways of identifying cognitive issues. In a 2014 journal article, Fisher and colleagues lament that cognitive decline in political leaders is widely left to conjecture. They suggested that newer, “highly sensitive brain-imaging” might improve detection. Such advances have yet to come to pass.

Earlier this year, ophthalmologists at Northwestern University Feinberg School of Medicine similarly suggested that better detection might be accomplished by taking pictures of our eyes. At the back of the eye sits a delicate nest of capillaries whose status can be a window into what’s happening in the blood supply to the brain. Still, the appearance of our retinas or brain is a secondary indicator—a proxy for what really matters, our current ability to think.

At the moment, neurologists’ most widely used test for cognitive impairment is a basic screening exam known as the Montreal Cognitive Assessment, or MoCA—a five-minute, 30-question test that involves knowing where you are, drawing a clock, and identifying a picture of a tiger. The most difficult part is recalling five words that you are told to remember at the beginning of the exam. The test is meant to identify people who will have difficulty completing the “activities of daily living.”

Last year Trump took this test. In a glowing press conference, his physician, Ronny Jackson, reported that the president aced it. But a test used for screening the general population for dementia does not necessarily suggest abilities to serve as commander in chief any more than, say, passing a test to get a driver’s license would. Anything less than a perfect score raises concerns, but a perfect score tells the public little about a candidate’s actual capacity.

Better tests do exist. Simply and transparently administered, they could help the public know what is really going on with the cognitive status of people who seek to hold the nuclear codes.


Cognitive testing for a government job has long been done in the military. In World War I, psychologists developed tests known as Alpha and Beta to assess a person’s ability to follow directions and process information. Even in a time of great need for service people, the U.S. military deemed it worth knowing whether a person could not follow basic commands. The testing process evolved into the current ASVAB, which is given to enlistees. It is not intended to disqualify anyone, but to help identify which areas of service a person might be able to execute.

Of course, the testing process becomes more difficult for roles that are expansive. It is simpler for a football team to test the qualities that make a good placekicker, for example, than a great coach. The president has to be good at many things, but not necessarily expert at any one.

The challenges in such a process were clarified recently by the Department of Defense. In 2016 it deployed a working group to develop a cognitive-assessment process for warfighters. The attempt failed, with the group reporting the following year: “The goal of determining warfighter cognitive state and predicting performance under training and operational conditions has not been achieved.” It blamed insufficient evidence linking performance on tests to performance in the field.

For a purely leadership role, though, some believe the most important skills can be tested for. Kim McCullough, a communication sciences and disorders professor at Appalachian State University, says certain tests could help assess a president’s abilities. While neurologists may soon have the tools to identify cognitive impairment in early stages, speech pathologists already do. And there is no clearer way to assess this than through tests of language. “A battery of tests could be given to candidates,” says McCullough, whose work focuses on identifying mild cognitive impairment. The most relevant abilities are demonstrated through those language tests: namely, the cognitive-processing abilities that show that a person can synthesize complex ideas and identify key points.

“Language is cognition in action,” Bayles says. Simple tests can be remarkably telling. If you had only one test to assess cognitive decline, she says, it would be this. First you tell somebody a simple story. (In her example, a lady went to the grocery store and lost her wallet, but then someone found it and she got it back.) Then you ask simple questions about what the person just heard. You ask the same questions just after the story, and then again 30 minutes later. Most young people with “normal” cognition get about 15 out of 17 questions right both times. Older people with normal cognition get one or two fewer in the second round.

The differences are striking in people with cognitive impairment. People with mild Alzheimer’s disease get about half the questions right in the immediate asking, and only one right the second time. People with moderate Alzheimer’s get two correct, and then none.

Losing one’s train of thought or forgetting words is not necessarily indicative of cognitive impairment. Some people think poorly onstage but brilliantly off. Others are great at processing and synthesizing information, and the only difficulty is in quickly putting things into words. A politician’s supporters might dismiss or defend patterns of forgetfulness under some belief that people need not be eloquent in their speech to demonstrate outstanding comprehension and processing—and, indeed, to be able to function optimally as president. One might even argue that a president should be trusted to make hasty or thinly informed decisions.

But such trust becomes more difficult when the president makes decisions that have no apparent basis in U.S. interests and are widely condemned by both parties, as in the case of Trump’s decision to withdraw troops from Syria. The appeal of functional cognitive testing is that it would help the public know whether a president can even clearly understand—by putting into words—the implications of his decisions and their alternatives.

“There’s no gold standard for one assessment where you can get everyone,” Bayles says. But she suggests that several, such as the Cambridge Neurological Inventory, Wechsler Memory Scale, and Edinburgh Social Cognition Test, could be of use. Based on two decades of longitudinal studies, she and colleagues have also developed a test that they’ve proved is sensitive enough to identify differences among 55-year-olds who are just starting to wonder whether something is going on. Known as the Arizona Battery for Communication Disorders of Dementia, or ABCD, it tests language comprehension, expression, and memory. It can be done in 90 minutes, and it dives more deeply into things such as reading comprehension, concept definition, and delayed retelling of stories a person is told.

In a recent study, the exam was able to identify cognitive impairment in 55-year-olds before they progressed to overt dementia. It doesn’t test executive function or problem solving, but it could be combined with others that do. The whole process could take a few hours and give a clear picture of cognitive fitness for office.


No process will be perfect. But if fitness is to remain a tenet of the electoral process, a validated, transparent assessment of cognition could leave little room for ambiguity or politicization. Results could be conveyed to the public in a way that would be more relatable than talking about people’s retinal capillary imaging, and more meaningful than knowing whether they take Lipitor or have ever had an appendectomy.

The standardized process could be a new custom among presidential hopefuls that actually has practical value. The process would have to be mindful of state impediments to free and fair elections with any kind of required qualification test. No result would be considered categorically exclusionary. If anything, a testing process like this would be less arbitrary than current limitations on who can serve. Being 35 or being born in the U.S. are historic relics that do not necessarily reflect a person’s actual abilities. A testing process like this could.

Without such a process, concerns of politicization and bias in medical assessments stand to get worse, not better, as age factors more prominently in politics year over year.

Okay, what were those numbers?

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.