A failed grounding for the wrongness of killing

Some people may think that what makes killing the innocent is in part that the prospective victim has interests, and what in part grounds the prospective victim's having interests is her having desires. This makes for a neat argument against the claim that the prohibition extends to embryonic and early fetal humans, since these appear not to have desires.

But I think the position leads to absurdity. Suppose Martha is in a severe but temporary depressive episode such that she has only one desire: to die. In such a case, killing Martha is murder. (Some defenders of euthanasia may disagree in the case where the depression is permanent, but what I said is uncontroversial. If need be, add that Martha does not consent.) According to the above story:

1. Martha's one and only desire is to die.
2. That killing Martha is wrong[note 1] is partly grounded in the fact that she has interests.
3. That Martha has interests is partly grounded in the fact that she has desires.

Add the following plausible grounding principles:

4. If p is partly grounded in q and q is partly grounded in r, then p is partly grounded in r. (Transitivity)
5. That ∃x(Fx) is partly grounded in the fact that Fa, where a is some particular such that Fa. (Weak form of Existential Grounding)

Then:

6. That Martha has desires is grounded in the fact that she has a desire to die. (1 and 5)
7. That killing Matha is wrong is partly grounded in the fact that she has a desire to die. (2, 3, 6 and 4)

But this is absurd. It's wrong to kill Martha, but surely not even partly because she desires to die. Of course, one can imagine odd cases where the wrongness of killing someone is partly grounded in their desire to die. For instance, suppose Patricia is someone whom it is permissible to execute for a crime, but that the executioner has sworn an oath never to execute someone who desires to die. In such a case, it could be impermissible for the executioner to kill Patricia in part because she desires to die. But the Martha case isn't like that. In the Martha case her desire to die does not contribute to the wrongness of killing her—if anything, it slightly lessens the wrongness (though I am not sure I want to say that). Moreover, while it would be wrong for the executioner to kill Patricia, it wouldn't be murder (assuming this was a case where the death penalty is morally justified), and what I am interested in is the grounding of that wrongfulness of killing that makes the killing be a murder.

Now, the case of someone whose only desire is to die may seem odd, and maybe someone could say that in odd cases one gets odd conclusions like (7). I don't know that that would be a satisfactory answer. But let's give a less odd case. Marcus desires to die, but he also has a full suite of other desires. I shall assume the case is to be elaborated in such a way that to kill him would be murder. To run the argument in the case of Marcus we need this principle like (5):

8. That ∃x(Fx) is partly grounded in the fact that Fa, for any particular a such that Fa and such that the fact that Fa is not itself partly grounded in ∃x(Fx).

The latter proviso is needed to rule out some paradoxes. Add the fact that:

9. That Marcus desires to die is not itself partly grounded in the fact that he has desires.

We can now modify the argument about Martha to conclude:

10. That it is wrong to kill Marcus is partly grounded in the fact that Marcus desires to die.

And that's absurd in this case. And while the case of Martha was very odd, alas the case of someone who among with other desires has a desire to die is not so very odd, though deeply unfortunate.

[Edited. The first version attributed the view that right to life is grounded in having desires to Boonin. But I think that attribution may be incorrect--he may only be claiming that it is grounded in (ideally, dispositionally) having a desire for life. See here.]