[0] https://www.webmd.com/pets/dogs/ss/slideshow-foods-your-dog-...
The point is to have an early warning in a situation where gas concentration potentially increases from harmless to dangerous levels, as in a mine shaft, where ventilation is limited. Something that is harmful in low doses to a bird isn't necessarily so to us, until it is too late, which makes the bird useful as an early warning signaller so people can get out early when the dose is still harmless and thus not easily detectable.
Thus the saying "the dose makes the poison".
Caffeine overdoses are too common among people and I've known friends that have died or almost died from them. (I'm strongly on the never side for caffeine pill usage as the risk of an OD from coffee is much lower and involves a lot more liquid sitting on your bladder.)
Theobromine in chocolate (and tea and other places) is a Caffeine relative that in us humans has a 1/10th to 1/100th the efficacy on our caffeine receptors. In dogs that relationship is reversed and theobromine is their caffeine and death by chocolate is their caffeine overdose.
People can overdose on theobromine too, it just takes much higher quantities still than caffeine.
Which yes, aligns with your point that doses matter, but the other thing useful thing to note about the Theobromine/Caffeine relationship is that subtly different but related chemicals have wildly different effects on different metabolisms. It's not just dose that matters, but also exploring the repercussions of entire families of chemicals (and the products they break down into) because you also need to know if a "cousin" chemical has larger efficacy at lower doses in people. (Results on small animals are still an early warning sign, the directions they may warn about aren't just exposure time or dosage amount, but also related chemicals and receptor paths and many other things.)
In any case, would it not be safer to assume the possibility of harmful impact of Chemical X until more concrete data is gathered? Granted, the fear mongering in the media needs to pump the brakes, but err on the side of caution, and all that.
Then again, where does the burden of proof lie; on PFAS not being harmful? That doesn't sound right either.
Bit of a headache, I guess.
You could argue that with food, there is a necessary trade-off, but with PFAS there isn't, because we don't need it to survive. But that is true for most of modern civilisation, and what you then end up with is the "this substance is known to the state of California to cause cancer" on basically everything.
You are right that this becomes a difficult problem of epistemology. Basically, it is impossible to prove "this is safe", only the opposite, and often it is not easy to prove conclusively. I'm not sure where the right balance lies, either.
This is true, but by itself this model presents an incomplete picture for assessing the risk of PFAS. Two additional models are important: bioaccumulation and half-life.
PFAS stick around (pardon the pun) in the environment. They are biomagnified upwards through the food chain. If we are releasing them into the environment at a greater rate than they are degrading, then the biosphere is building up a reservoir of PFAS whose acute effects in humans will lag their release into the environment.
Additionally, it is hard to assess the lethal dose of something over a lifespan vs. an acute exposure. This is because slow poisons tend to pick up correlations and interactions with other illnesses as they progress over time (e.g. someone who struggles to get out of bed because of a persistent infection will also get less exercise and sunlight). So the long-term effects of things on the body tend to be less tractable research questions and thus have a negative bias in the scientific literature, which can easily be mistaken for "evidence of absence". This is why the precautionary principle is important for anything that is as pervasive as PFAS.
Agree with long-term effects - this is really hard. The precautionary principle is dangerous in that you could justify the most draconian COVID measures like lockdowns with them, which in hindsight were unwarranted and unprecedented over-reach, having no positive effect at all.
I don’t think there is a recipe that can be easily followed, you have to have very good people, both intellectually and ethically, and even then its still hard.