An experiment to determine 'death awareness' in animals would require a lot of dead animals. Fortunately for the stoopid animals (ie most of them) they would not be viable candidates. Only those self-aware would qualify.
And so we get a group of elephants and test each individual's competence in recognising themselves with the mirror test. Once we have an adequate number of dumbos trying to rub spots off themselves, we introduce them to the concept of the bazooka.
As a group, the elephants are shown the bazooka. Then, one of the dumbos is blown to bits. Each of the other elephants needs to be present to witness the death of nelly number one. The bazooka is then applied to nelly number two. To the same effect. And so on.....
If there is death awareness, the reaction of nelly number one when having a bazooka pointed at it, should be quite different to the reaction of nelly number one hundred, when they are lined up in the crosshairs.
1. make sure all subjects are self aware;
2. make sure that all recognise a dead one of their own kind;
3. ensure that they know what a bazooka looks like;
4. ensure that they connect the appearance of a bazooka with a chance of dead elephants;
5. analyse the behaviour pattern of subjects, to determine if they project their own self-awareness forwards in time, in appreciation of death and the consequence of it.
6. either watch the Church of Pachydermia come to life, or find that humans are at least more intelligent than dumbo!
If I'm understanding you correctly, this seems like other tests where mice, for instance, would avoid a particular feeding zone for fear of shock. Its just recognizing patterns of pain/pleasure
I would like to test your elephant hypothesis though... you bring the bazooka, I'll spring a few elephants from the zoo