Sven Andréasson’s 1987 paper in the Lancet moved the relationship between cannabis and mental illness out of the realm of Reefer Madness. Psychiatrists and researchers had to take seriously the possibility that marijuana could cause schizophrenia.
But possibility isn’t fact. Andréasson had designed his study carefully. He had shown men who reported smoking cannabis developed the disease more often, and the more they smoked the higher the risk.
That association might mean the drug caused psychosis.
Then again, it might not. Other possibilities still existed. Skeptics offered four competing theories. In rough order of likelihood:
First, that the same genes that caused people to develop psychotic illnesses also encouraged them to use cannabis heavily. In that case, even though marijuana use often preceded or accompanied schizophrenia, it would not be causal. People would become schizophrenic for genetic reasons whether or not they used.
Imagine looking at a highway at night in winter. The pavement seems clear, but it’s actually covered with black ice. A dump truck scatters sand on the pavement. A few minutes later, a car drives past—and skids off.
The sand was put on the road before the accident, but did it cause the accident? Of course not. The ice did. But the ice is impossible to see. So, someone who relied only on the video to figure out what had happened might wrongly blame the sand. In reality, both the sand and the accident resulted from the ice.
In this analogy, the sand is cannabis use. The accident is schizophrenia. Our genes are the ice—the hidden fact that leads us to make an incorrect assumption about the events we can track.
The theory made intuitive sense. Schizophrenia was known to have a strong genetic component. Drug abuse also seemed to have some relationship to genes. But genetic science was in its infancy in 1987. Scientists had not yet decoded the human genome, much less discovered which mutations might be associated with schizophrenia or drug use.
As a result, no one at the time could prove or disprove the argument. To use Donald Rumsfeld’s famous line, it was a “known unknown,” a problem that could be outlined but not solved. (For a time, tobacco companies had tried the same argument about lung cancer, arguing that the same genes might both cause people to smoke cigarettes and raise their risk for lung cancer.)
Second, that although cannabis use seemed to cause schizophrenia, in fact the causation was reversed—the onset of the disease led to cannabis use. In its most extreme version, this theory held that cannabis might treat schizophrenia symptoms.
Doctors knew that many people who developed schizophrenia went through a slow decline before they suffered open breakdowns. As George Ewens had written in 1908: A gradual change of disposition; a loss of activity and energy, the patient becoming silly, shy, irritable, obstinate, and careless . . .
Psychiatrists now call that period the “prodrome.” Prodromal patients are becoming ill but haven’t yet snapped. They might hear voices from time to time or have intrusive thoughts and minor delusions—everyone at school is laughing at me, or I think my drink is spiked. Their grades often slip. They may become disorganized or stop taking care of themselves. To cope, they begin to withdraw from friends and family.
But they are still well enough to understand that their thoughts are abnormal. They’re anxious and depressed about their mental chaos. They smoke cannabis for the pleasure it produces or just to spend time with friends who are also smoking. Over time, as their psychosis progresses, they smoke more. Then they suffer an open break.
Because the heaviest use occurs just before the break, the drug seems to cause it. But according to this view, in fact smokers are using cannabis to try to control their symptoms. They’re “self-medicating.”
Andréasson had tried to account for this theory by excluding from his analysis recruits who were already suffering from schizophrenia when they were drafted. He only counted cases where recruits who were already smoking developed psychosis later. But Andréasson couldn’t exclude the possibility that some recruits might have been suffering from hidden psychotic symptoms before the time of conscription—and possibly before they began to use marijuana.
Still, the self-medication theory faced big stumbling blocks. Psychiatrists knew that cannabis could cause temporary psychosis even in healthy people. They knew that many patients whose schizophrenia appeared under control suffered relapses if they used cannabis again. Andréasson also noted that even draftees with no evidence of depression or personality problems showed “an increased risk of schizophrenia with increasing cannabis consumption.”
Researchers tried different arguments to square those facts with the self-medication theory. Maybe temporary cannabis psychosis didn’t have anything to do with schizophrenia. Maybe people who already had developed schizophrenia started smoking again because they were relapsing and trying to control their symptoms with marijuana.
Those explanations weren’t great. All in all, self-medication was less convincing than shared genetics as a hypothesis. But marijuana advocates would come to promote it. Unlike the genetic theory, it was easy to understand. Plus, it put a positive spin on the fact that so many people with mental illness smoked marijuana. See? Getting high is actually good for people with psychiatric disorders!
The third argument was that Andréasson’s study had shown a falsely positive result by pure chance. Scientists say a clinical trial has proven its hypothesis if it has a “p-value” of 0.05 or less. In plain English, that number means that the trial has found a difference between two groups that has less than a 5 percent result of occurring by chance. But even at that standard, 1 trial in 20 may have a false finding. Andréasson’s “crude” findings—before he accounted for confounders—were very strong, with p-values far below 0.05. But after he adjusted, he was on the edge of that 5 percent line. Maybe his trial was the 1 in 20 that had accidentally produced an invalid finding.
Like the genetic theory, this objection couldn’t be answered immediately. But in the long run it would be the simplest to prove or disprove. If later studies on different populations failed to show a link, Andréasson’s study would become less convincing. If they backed it up, his thesis would get stronger.
The fourth theory was that cannabis use was just a marker for other social or environmental problems that might cause schizophrenia. People whose parents were in prison, used drugs, or did poorly in school were more likely to use marijuana. Maybe those stressors or the other drugs were the real risks.
This theory was basically just confounders by another name, and Andréasson had gone out of his way to account for other variables. Plus, those other factors didn’t seem to cause schizophrenia on their own, though they might worsen its course. It seemed more likely that cannabis stood apart from them. Still, the environmental theory couldn’t yet be entirely discounted.
Psychiatrists and researchers batted around versions of those theories for fifteen years after Andréasson’s paper without reaching definite conclusions. The need for more evidence—not just on the psychosis link, but on all of marijuana’s potential risks and benefits—became clear in 1999. That year, the Institute of Medicine, the federally backed group that was later renamed the National Academy of Medicine, put out a report titled “Marijuana and Medicine: Assessing the Science Base.”
The study was the last major effort to look at cannabis by the group before its 2017 paper. Reflecting the state of the science at the time, the 1999 paper is notably more positive than the later report, both in terms of marijuana’s benefits and its risks.
“Except for the harms associated with smoking, the adverse effects of marijuana use are in the range of effects tolerated for other medicines,” the report found. It recommended continued research into marijuana for many diseases.
In terms of negatives, the report focused on potential physical hazards from smoke and largely discounted other concerns. Only a few smokers became dependent on marijuana, withdrawal symptoms were “mild and short-lived,” and no evidence proved that marijuana use led to harder drugs.
As for mental illness, the 1999 report hardly mentioned it. It said cannabis psychosis was almost always temporary and dealt with schizophrenia in a single paragraph, saying “The relationship between marijuana and schizophrenia is not well understood.” It even suggested “the possibility that schizophrenics might obtain some symptomatic relief from moderate marijuana use.”
Time and further research have not been kind to that theory. They have been even harsher on the report’s view on prescription opiates, which is worth noting because it informed its findings on the potential positives of medical marijuana. The report lauded the late-1990s trend toward increased opiate prescribing:
[Fears] that liberal use of opiates would result in many addicts . . . have been proven unfounded . . . Few people begin their drug addiction problems with misuse of drugs that have been prescribed for medical use . . . the diversion of medically prescribed opiates to the black market is not generally considered to be a major problem.
Six hundred thousand overdose deaths later in the United States, it’s fair to say the report—and pro-opiate physicians more generally—underestimated the risks of encouraging opiate prescriptions and use.
The fact that scientists twenty years ago understood so little about cannabis and the brain didn’t help. Research into what was becoming known as the endocannabinoid system was in its infancy. Only in 1988 had researchers discovered the CB1 receptor, the structure on brain cells that THC activated. The hippocampus, which helped form memories, and the amygdala, which helped govern emotion, memory, and fear, both had heavy concentrations of the receptors.
In 1992, scientists discovered anandamide, an “endocannabinoid,” a chemical our own bodies produce to stimulate CB1 receptors. But researchers weren’t sure exactly how anandamide interacted with dopamine and other neurotransmitters. Anandamide was hard to study, too. The stuff barely even existed permanently in brain cells or channels. It was produced as needed and then quickly destroyed.
Scientists also discovered that triggering CB1 receptors indirectly caused some dopamine release, although it was concentrated in a different part of the brain than the dopamine release that opiates caused. Still, the fact that THC could increase dopamine levels helped explain the high associated with marijuana. It also hinted at an answer for its link to psychosis.
But the difficulties of translating basic neuroscience research into the thought distortions of psychosis—the difficulties of turning brain into mind—meant that stronger proof would have to come from somewhere else.
A collaboration between researchers on opposite ends of the globe provided the next step.
Dunedin is a picturesque city of 120,000 on the southeast coast of New Zealand’s South Island—a rock in the Pacific Ocean, a long way from anywhere. In the late 1960s, a Dunedin pediatrician named Patricia Buckfield created a database of all the babies born in the Queen Mary Hospital, the city’s obstetric center. Buckfield planned to track the relationship between prenatal problems and later health complications.
In 1971, Phil Silva, a school-teacher-turned-psychologist, joined Buckfield’s research, which initially focused on just 250 kids. Silva noticed that a lot of those children had developmental problems their family pediatricians had not diagnosed. He pushed for a larger study that would include all the babies born between April 1972 and March 1973 whose families still lived in Dunedin in 1975. The potential pool included 1,139 children. Most of their parents—covering 1,037 children in all—agreed to participate in what became known as the Dunedin Multidisciplinary Health and Development Study.
At first, Silva expected the study would include only a single assessment when the children were three. “The resources for the Study were scarce,” he wrote in a 1996 book about the project. Aside from a part-time secretary, Silva mostly relied on volunteers at first. They conducted assessments of the children in the Sunday school classroom of a local church. But Silva kept finding new funds to support and extend the project.
The kids were measured again at ages 5, 7, 9, 11, 13, 15, 18, and 21. Each time nearly all took part. By then, the study’s value was obvious. Dunedin was not the largest study of its kind, but it was probably the most comprehensive in terms of tracking participants. The high participation rate was crucial. Long-term social science studies had the same dynamics as high school reunions—the most successful people were the most likely to come back.
But the Dunedin researchers went to great lengths to make sure the children and their families understood the value of the study. Only once, at age 13, did fewer than 90 percent take part. In later years the rate increased. The researchers got to know the participants well and promised confidentiality even for potentially criminal behavior such as drug use or violence. The New Zealand police didn’t try to interfere. As a result, the Dunedin research truly described the entire group, for better or worse.
As the survey progressed, Silva added permanent staff and more researchers. On a trip to Los Angeles in the early 1980s, he met Terrie Moffitt, a psychology graduate student at the University of Southern California whose research focused on the reasons people became violent. Moffitt had broken her leg in a skydiving accident and was temporarily stuck in a wheelchair, so she and Silva had plenty of time to talk. In 1985, Moffitt joined the study—while simultaneously becoming an assistant professor at the University of Wisconsin. For the next several years, she split her time between Madison and Dunedin, a mere 8,700 miles apart.
By the mid-1990s, the Dunedin research had attracted international attention. A psychiatrist named Robin Murray was among those who noticed. A Scottish native, Murray has spent most of the last five decades researching and treating schizophrenia at the Institute of Psychiatry, Psychology & Neuroscience of King’s College London.
Murray was born in 1944 in Glasgow, the son of teachers. He grew up in rural Scotland but came back to Glasgow for medical school. His interest in psychiatry took off after he spent a year living in a psychiatric hospital as a medical student. In exchange for free room and board, he had to examine the patients periodically. He found their damaged minds fascinating.
In 1972, he moved to London to begin his residency in psychiatry at “the Maudsley”—Maudsley Hospital, a campus of brick buildings in Denmark Hill, a low-income section of South London. Three years later he joined the Institute of Psychiatry, which shares the Maudsley’s campus and is the largest research center for psychiatry and neuroscience in Europe.
Aside from a single year working at the NIMH (National Institute of Mental Health) in Maryland, Murray has remained in London ever since. In 2010, another professor joked that Murray’s career was “a rare example of institutionalization being wholly positive for both the institution and the individual.”
In the 1970s and 1980s, Murray focused on the genetic underpinnings of psychosis and the structural changes to the brain that schizophrenia seemed to cause—the core of American research. But over time, Murray grew to believe that psychosis resulted not just from genetics but from disturbances in infancy, childhood, and young adulthood. The fact that Denmark Hill was located near Brixton, a poor London neighborhood with high rates of cannabis use and schizophrenia, helped convince him. He became interested in the possibility of a link between cannabis and psychosis.
Other psychiatrists at the institute also wanted to investigate how environment and genetics might combine to cause mental illness. In 1994 King’s College and a British government group created the Social, Genetic & Developmental Psychiatry Centre at the Institute of Psychiatry.
Meanwhile, in New Zealand, Phil Silva was nearing retirement. Terrie Moffitt had become the project’s associate director. In 1995, Michael Rutter, a child psychiatrist at King’s, convinced Moffitt to bring the Dunedin data to the new center, though London was 11,000 miles from the South Coast—an even longer haul than Madison.
By then, the Dunedin project had come a long way from its modest roots. Further assessments were conducted starting in 1998, when participants were 26, as well as when they turned 32, 38, and 45. The plan now is to continue the study through the lives of the participants. Secondary research will examine their children and even their grandchildren. The project will long outlast Silva, its founder, who retired in 2000. Silva is nearly 80 now and complained via email that he was “going deaf and losing vision fast. That is called ageing.” Still, he had enough energy to take credit for his work: “I founded the world-renowned Dunedin study in the early ’70s and directed it for 30 years. I taught the present director, Richie Poulton, most of what he knows, and Terrie Moffitt taught him the rest.”
With Moffitt on board, Robin Murray looked for fresh ways to use the Dunedin data. He liked to encourage collaboration among young researchers. He put together Mary Cannon, an Irish psychiatrist and epidemiologist, with a Canadian psychologist named Louise Arseneault. (In fact, Arseneault had just written a paper based on the Dunedin data that examined risk factors for violence; more on that finding later.)
“Robin Murray said I should go meet with Louise Arseneault,” Cannon told me. “I came from a psychosis background, and she was studying delinquent behavior in young people. We worked well together.”
Murray encouraged Arseneault and Cannon to see what the Dunedin data revealed about cannabis and mental illness. “I remember having discussions with Robin Murray,” Arseneault said. “We decided to explore the link between schizophrenia and cannabis.”
In one way, Dunedin was not ideal for this work. With fewer than 1,000 participants in its age-26 pool, it was unlikely to generate more than ten full-blown schizophrenia cases. With such a small sample, even a strong marijuana-schizophrenia link wouldn’t pass the 0.05 p-value test.
To increase the sample size, the researchers included people with a diagnosis of schizophreniform syndrome. Those were psychotic episodes that ended within six months and thus didn’t quite reach full-blown schizophrenia. Though less severe than schizophrenia, schizophreniform syndrome was still a destructive illness.
But the Dunedin data offered advantages, too. New Zealand had high rates of cannabis use, so the dataset offered a large number of people who’d used the drug. More important, unlike almost any other studies, it had data on preteen, premarijuana-use symptoms of psychosis. A psychiatrist who interviewed the children when they were 11 had asked them about psychotic symptoms.
The age-11 data meant that Arseneault and Cannon could account for preexisting psychosis when they tried to find if cannabis use caused later psychotic disease. Plus, the data on marijuana use at age 15 had been collected when the participants were age 15, not years later. It didn’t depend on the memories of participants, and researchers didn’t have to reconstruct it afterward, when they already knew who had developed full-blown psychosis.
“You have such good data collected prospectively,” Cannon says. “That’s the most important thing—this data was collected without any bias.”
Arseneault and Cannon ran the numbers. They knew of Andréasson’s Lancet paper, but they weren’t sure if they would have similar results, Arseneault says. “You never know what you’re going to find.”
What they found jumped at them.
People who had used cannabis at age 15 were more than 4 times as likely to develop schizophrenia or schizophreniform syndrome as those who never used. Even after accounting for those kids who had shown psychotic symptoms at age 11, the risk remained threefold higher.
“A fourfold increased risk was very big,” Cannon says. When it came to psychosis and schizophrenia, she was used to seeing slight increases in risk—10 percent or 20 percent.
With Arseneault as the lead author, they wrote a paper on their findings. “Using cannabis in adolescence increases the likelihood of experiencing symptoms of schizophrenia in adulthood,” they reported. “Our findings agree with those of the Swedish study.”
In what was becoming a trend, they submitted it to the British Medical Journal and the Lancet, but not the top American medical journals. The BMJ ran the paper in its November 23, 2002, edition, along with an update of the original Swedish paper and a third study from Australia that examined links between cannabis and depression.
Meanwhile, in a paper published a few months earlier, Dutch psychiatrists had reported that cannabis seemed to cause psychosis in some healthy adults. In 1996, the researchers randomly chose thousands of adults in the Netherlands. They asked the participants about cannabis use and interviewed them to see if they had any psychotic symptoms. The researchers followed up again after a year, and then after three years. They wound up interviewing more than four thousand people in all; more than 98 percent did not have a psychotic disorder.
Three years later, 7 out of 300 previously healthy adult marijuana smokers had developed psychosis or serious psychotic symptoms. Over the same period, only 3 of more than 3,600 people who did not smoke developed psychosis or serious psychotic symptoms.
Many people who used cannabis also used other drugs, the study found. But when the researchers adjusted for other drug use, the psychosis risk of cannabis users remained strong. The scientists estimated that cannabis use might be responsible for as much as half the serious psychosis in previously healthy adults.
• • •
Taken together, the Swedish, Dutch, and New Zealand studies provided powerful evidence linking cannabis to mental illness. In a BMJ editorial discussing them, two Australian psychiatrists wrote:
Although the number of studies is small, these findings strengthen the argument that use of cannabis increases the risk of schizophrenia and depression, and they provide little support for the belief that the association between marijuana use and mental health problems is largely due to self-medication.
All three papers received plenty of scientific attention, especially the one by Arseneault and Cannon. Their paper has been cited more than 1,300 times since they published it, even more than Andréasson’s 1987 paper.
Cannon, now a professor of psychiatric epidemiology in Ireland as well as a practicing psychiatrist, says she expected that the 2002 findings would have a big impact on public attitudes. After all, three sets of researchers in three different countries had come at the question of marijuana and psychosis from three different directions. They all had reached the same conclusion. Cannon expected that the consensus would move close to ending the debate.
But it didn’t. Skeptics argued the Dunedin study was too small to be meaningful and that the Swedish data depended on self-reported cannabis use. Plus, the genetic objection remained unanswered.
“People were very polarized by this,” Cannon said. “I thought the data was very clear, and people were saying this cannot be true, it cannot be causal . . . To me, it’s just so blindingly obvious—we’ve found this incredibly robust effect, and it’s been confirmed a number of times now.”
In the United Kingdom, perhaps the best-known doubter was David Nutt, a brain researcher and member of a government commission that advised British lawmakers on the relative dangers of drugs. Britain classifies drugs as Class A, B, or C, depending on their dangers. Class A drugs, like heroin, carry the harshest consequences. Class B drugs have lesser penalties, though users and dealers can still face jail time, and Class C lesser still. When the BMJ published its research in 2002, Britain categorized cannabis as a Class B drug. But the government advisory committee had just recommended that it be moved to Class C.
For Britain, a move from B to C was the rough equivalent of decriminalization in the United States. Marijuana would still technically be illegal, but both police and users would know that it was viewed far less seriously than drugs like heroin. The committee stuck to its advice despite the new research. In 2004, Britain moved cannabis to Class C.
Yet, surprisingly, cannabis use did not increase in Britain after the reclassification—in part because of Robin Murray and the Institute of Psychiatry. The institute and genetics center eventually grew to 150 psychiatrists, neuroscientists, epidemiologists, graduate students, and other researchers. And Murray is at the heart of the institute, as I found when I traveled to London to see him and the place firsthand.
Now in his mid-seventies, Murray remains energetic, writing papers, treating patients, and traveling worldwide to conferences on mental illness. Even his personal life is inextricably interlinked with schizophrenia research; Marta Di Forti, Murray’s second wife, also works at the institute, and they regularly collaborate. Over the years, Murray has helped train hundreds of young psychiatrists. Forty have become full professors. With his encouragement, many became interested in the cannabis-psychosis link:
• “Cannabis Use and Outcome of Recent Onset Psychosis,” European Psychiatry, June 2005.
• “The Environment and Schizophrenia: The Role of Cannabis Use,” Schizophrenia Bulletin, July 2005.
• “What Is the Mechanism Whereby Cannabis Use Increases Risk of Psychosis?” Neurotoxicity Research, June 2008.
• “The Acute Effects of Synthetic Intravenous Δ9-Tetrahydrocannabinol on Psychosis, Mood, and Cognitive Functioning,” Psychological Medicine, October 2009.
Murray coauthored those papers and more than eighty others on cannabis and psychosis. Not all were pathbreaking. Some were small environmental studies that followed patients. Others were reviews of earlier research.
But even the secondary efforts kept British journalists and politicians aware of the issue. The institute’s researchers had a big structural advantage. They worked in a neighborhood with lots of cannabis use and psychosis, yet not even five miles from the heart of London—England’s business, government, and media center. They could reach news organizations easily, and they succeeded in getting their message out. In 2007, The Independent, a British newspaper that had endorsed cannabis decriminalization, reversed its view. In October 2008, the British government reclassified cannabis back to a Class B drug.
By then, David Nutt had become the British government’s chief drug advisor. But the reclassification frustrated him, and he took some outlandish positions. In 2009, he wrote a paper called “Equasy: A Harmful Addiction,” arguing that riding horses was much more dangerous than Ecstasy, an amphetamine derivative. If horse-riding was legal, Ecstasy should be, too, he wrote. The government fired him soon after.
In 2010, Nutt continued in the same vein with a paper claiming alcohol was more harmful than cocaine or even heroin. Alcohol does kill more people than heroin, because it is used so much more widely; about 70 percent of American adults drink at least once a year, and more than half at least once a month.
But even before the opiate epidemic, no one doubted that heroin would have catastrophic consequences if it were used as widely as alcohol. Large studies have shown that about one-third of people who use heroin even once become addicted and that opiate addicts die at 10 to 20 times the rate of nonusers. (For cocaine, the addiction rate seems to be in the 20 percent range, and addicts die at 4 or 5 times the rate of nonusers.)
Since then, Nutt, who remains a professor of neuropsychopharmacology at Imperial College London, has been involved in a quixotic effort to develop a product he calls “alcosynth,” a replacement for alcohol. “The drinks industry knows that by 2050 alcohol will be gone,” he told The Independent in 2016. “But they don’t want to rush into it, because they’re making so much money.”
While Nutt chased synthetic alcohol, Robin Murray’s reputation grew. In 2011, he received a knighthood—making him Sir Robin Murray, though after seeing his rumpled clothes, I had a difficult time imagining anyone calling him sir. He became the most highly cited schizophrenia researcher in Europe.
And the repeated warnings from King’s College didn’t just get media attention. They helped change the way Britain views cannabis.
• • •
In 2001, more than half of British adults favored legalization of cannabis, according to the British Social Attitudes Survey. By 2010, only about 1 in 3 did. Since then, the percentage has increased, but it remains below 50 percent. A 2016 survey of UK students showed that they associated cannabis strongly with mental health problems. A Twitter search for “cannabis psychosis” shows British teenagers and young adults discussing the issue as a serious risk.
Meanwhile, British cannabis use has plunged. From 2000 to 2005, more than 10 percent of adults and almost 30 percent of young adults used at least once a year, according to British government surveys. Now the figures are about 6 percent for adults and 17 percent for young adults.
Maybe the most telling proof of the effectiveness of Murray’s efforts came in the jeers that British pro-cannabis campaigners launched against the Institute of Psychiatry. In 2015, CLEAR (Cannabis Law Reform), an advocacy group, complained on its website about the “anti-cannabis cult at King’s College.”
CLEAR trotted out the old argument that psychosis research has confused correlation and causation—as if scientists worldwide have not spent the generation since Sven Andréasson’s paper teasing out cause and effect. “The team at King’s College displays all the classic markers of a cult. It pursues a belief in cannabis as the ‘devil’s lettuce’ as a quasi-religion . . . it endlessly repeats itself, its ‘studies’ and its presentation of them as proof,” CLEAR wrote.
Three days later, CLEAR went further, implying without evidence that Murray and other researchers benefitted financially from their findings. “We simply cannot rely on these so-called eminent scientists to be honest about their work. They are in the gutter and they aren’t looking at the stars, they are looking at their bank balances.”
In fact, Murray has no financial interest in cannabis research. He has always been careful not to position himself explicitly against legalization. In 2009, in the wake of the classification controversy, he wrote in The Guardian, a British newspaper: “I care little whether cannabis is classified as a class B or class C drug. Fourteen-year-olds starting daily cannabis use do not agonise over its exact classification . . .” He simply wanted to educate people about the drug’s mental health risks and build a consensus against teenage use, he said.
Still, Murray clearly doesn’t favor legalization. When I asked him if he was surprised that so many American states had legalized, he gave me his only off-the-record answer, then added, “I’ve been surprised by the power of the business lobby for cannabis.”