Cannabis Kills No One

A reader challenges my assertion that overdosing on marijuana is a "physical impossibility":

Do you have any actual evidence that dying from smoking marijuana violates one or more laws of physics?

Exhibit A: Judge Young's findings from the 1988 DEA hearing [pdf] that considered reclassifying marijuana as a Schedule 2 drug:

Nearly all medicines have toxic, potentially lethal effects. But marijuana is not such a substance. There is no record in the extensive medical literature describing a proven, documented cannabis-induced fatality. This is a remarkable statement. First, the record on marijuana encompasses 5,000 years of human experience. Second, marijuana is now used daily by enormous numbers of people throughout the world… Yet, despite this long history of use and the extraordinarily high numbers of social smokers, there are simply no credible medical reports to suggest that consuming marijuana has caused a single death…

In a 2009 paper [pdf] on non-medical use of cannabis, Wayne Hall and Louisa Degenhardt round up the literature on its short- and long-term health effects. On the acute effects:

A dose of 2–3 mg of THC will produce a high in occasional users who typically share a single joint with others. Regular users might smoke up to 3–5 joints of potent cannabis a day for several reasons, including development of tolerance and to experience stronger e?ects… The dose of THC that kills rodents is very high and the estimated fatal human dose is between 15 g and 70 g, which is much higher than that smoked by a heavy user.

Note that the fatal dose is at least 5,000 times the dose that will produce a high in occasional users. On driving while high:

Studies of the e?ects of cannabis upon on-road driving found more modest impairments than those caused by intoxicating doses of alcohol because cannabis-a?ected people drive more slowly and take fewer risks. Nonetheless, some experimental studies have shown diminished driving performance in response to emergency situations… Driving after having taken cannabis might increase the risk of motor vehicle crashes 2–3 times compared with 6–15 times with alcohol.

On the effects of chronic use, "defined as almost daily use over a period of years":

In Australia, Canada, and the USA, cannabis dependence is the most common type of drug dependence after that on alcohol and tobacco… The lifetime risk of dependence in cannabis users has been estimated at about 9%, rising to one in six in those who initiate use in adolescence. The equivalent lifetime risks are 32% for nicotine, 23% for heroin, 17% for cocaine, 15% for alcohol, and 11% for stimulant users.