Cannabis tends to form an incomplete embryo, a miniscule seed in every calyx.
I've seen a lot of people deny that, and after saying they might need a magnifying glass to spot them, I've never heard from them.
Pollen can fly along with the wind for quite far. I'm dealing with hemp contaminants from over 15 miles away. France has been dealing with Moroccan cannabis pollen in their hemp, which can blow across a sea if the wind is right.
Some plants make their own pollen, usually at the end of flowering, so that doesn't lead to matured seeds.
Other types of hermies - as far as I've seen in the past 10 years - tend to lead to hermie offspring. I keep hearing and reading it should result in female seeds, but I've seen that happen in less than 5% of all cases I monitored. The numbers just don't add up in practice. One strain of hermie wildtype Afghan even completely hominized (100% male) while another never even flowered..
However, if chemicals are involved to induce the hermie trait, the offspring does in fact tend to be pure female in something close to 95% of all cases.
If you'd ask me, there's a huge difference in auto/self-hermie-ing compared with chemical-hermie-ing. The general concensus on most forums though, is that there is little to no difference. Personally I like to go with the data I collected, which shows that there's a relation between automatic/self-hermie offspring being hermie, but no relation between automatic/self-hermie offspring being female.