If you're unable to wash beer out of a shirt, you need a new detergent.
Causation can't be proved even in the case of smokers, any more than it can be
proved that drinking impairs your driving. However what you can do is show a correlation which is highly statisticially significant, such that the 'null hypothesis' (i.e. smoking has no link with cancer) can be rejected at that level of statistical significance.
But if you want to go to that level of semantics you don't have to accept that the earth revolves round the sun, that gravity exists, that there really is a country called Australia etc.
However, if you believe that 'active' smoking causes cancer, and you don't believe that passive smoking does, then the logical implications are:
a) There is some threshold value below which it is perfectly harmless and above which you get cancer, and that that concentration is only obtained in 'active smoking', or
b) ALL carcinogens are inhaled by every smoker AND all other by-products are harmless
Even if you're stupid enough to think that, there is still the point about smell etc - and your counterargument in that regard is ludicrous. Smokers who smoke in company KNOW that as a direct consequence of THEIR CHOICE other people suffer (i.e. end up smelling of smoke). The correct analogy would be a 'drinker' who decided to throw his drink in the air knowing that it would land on everybody (drinker and non-drinker) in his/her company. I'd suggest to you that anyone who did that would be heading for a slap in the chops, and rightly so. Would that smokers were the same
