Well, there have been zero deaths related to marijuana toxicity, but it would be naive to say that people don't die due to, in part, adverse effects of marijuana smoke... The difficulty is in determining whether someone's death can be linked to usage of cannabis. That said, the only statistics I could find that linked any deaths at all were from the Australian Bureau of Statistics. I think looking at the Australian statistics give a good idea of the relative risks of alcohol and cannabis. So StevenC, this one is for you.
First, let's compare how many people in Australia you might expect to be drinking or smoking on any particular day if you took a naive average of statistics. According to the Australian Institute of Health and Welfare statistics for 2004 of the total Australian population, 8.1% drink daily, 41.2% drink weekly, 33.5% drink less than weekly, and the rest don't drink. According to the National Cannabis Prevention and Information Centre, 34% use cannabis, and using their statistics of consumption trends, of the total Australian population, 12.24% smoke daily, 11.22% smoke weekly, 10.54% smoke less than weekly. If we take a naive average for the daily use, taking 1/7th of the weekly use, and let's say, 1/60th of the "less than weekly" use, we get the following percentages of use on any particular day: 14.54% of the Australian population drinks on a naively average day, and 14.02% of the Australian population smokes on a naively average day. Now this is pretty comparable, that's about 3 million for each given the total Australian population of roughly 21 million. But now let's compare some other statistics...
The Australian Bureau of Statistics lists 250 deaths between 1997-2005 that are linked to cannabis use, though it's unlikely that any of these were a result of direct toxicity, and more likely to impaired judgement and co-ordination. Now, the deaths due to simple alcohol toxicity (not even taking into account deaths related to impaired judgement and co-ordination and alcohol attributed diseases such as cirrhosis of the liver) between 1997-2005 was 1867 deaths, according to the same source. According to the Community Alcohol Action Network, there are roughly 3400 deaths every year due to alcohol caused disease and injury, which extrapolating over the span of eight years (1997-2005) is roughly 27200 deaths.
So if you please, compare 27200 deaths to 250 deaths for a very comparable rate of usage amongst the population and tell me that cannabis isn't safer than alcohol. This isn't even taking into account the injuries and hospitalizations due to alcohol and the immense cost on the health system. Yes, my calculations were naive and inexact, but even if I was off by a factor of 100, the relative risk would only be comparable.
To add a little anecdotal evidence to this... think of all the fights you've witnessed in your life. How many were alcohol induced... and how many were cannabis induced? When you think of abusive husbands and boyfriends, do you think of alcoholics or potheads?
Now ask yourself if alcohol being more dangerous than cannabis should make a difference? Can we baby our citizenry by not allowing them to engage in activities that are harmful to themselves? The number one cause of death is highly correlated to poor diet, but should we legislate diet to save lives?