Saviour Siblings, Ends and Means: How far can you go?

It would appear that France is being forced to grapple with the idea of saviour siblings in a case that looks to be in essence a copy of the UK’s Hashmi case from 2002.  That case concerned a couple who wanted to screen embryos to ensure compatibility in order that cord blood from the baby could be used to treat its brother’s beta thalassemia.  The HFEA ruling at the time was that, so long as the parents were planning to have another child anyway, there was no reason why they could not do so by IVF in order to select the “right” embryo.

There’re two reasons why the case caused controversy – and will, I suspect, cause controversy in France.  One had to do with the standard pro-life concern about discarding unwanted embryos; if you think that embryos have a right to life, then a treatment that creates them in the expectation that at least most of them will never be brought to birth is worrisome.  The other has to do with a concern about ends and means; the worry here is that the younger sibling is being treated wholly as a means to another’s end.

Of course, there’s nothing wrong with treating another person as a means to an end.  The problem, on this line of argument, has to do with treating a person as a means to an end and nothing else besides.  I treat the person behind the deli counter as a means to an end; but that’s not immoral just as long as my attitude is tempered by keeping in mind that they’re more than just a salami-fetching drone.  The person behind the counter is a person, too.  So as long as the parents in any saviour sibling case don’t treat the saviour simply as a source of cord blood, there’s no reason to get too worried, morally speaking.

The criterion that another child should be desired anyway is ostensibly a way to meet this requirement.  If you want another child for its own sake, then the harvestable and useable cord blood is just a fortunate by-product.  You wouldn’t be treating the saviour wholly as a means to an end; you probably wouldn’t be treating him as a means to an end at all.

But this raises a question: would it really be so bad if the saviour was treated as a means to an end, either wholly or predominantly?  That is: suppose your only reason for having another child is to generate cord blood that can be used to save the life of an older sibling.  Well?

It’s still hard to see how the saviour could complain that he’d been wronged.  That is, the motive for his creation might have wronged him, inasmuch as it was wholly instrumental, but for one important point: he didn’t exist to be wronged, and wouldn’t exist but for the action in question.  We can’t even say that an action the analogue of which would have been wrong in respect of an existing child is not wrong here, just because there is no analogue.  There’s noone there to be wronged until the baby is created, and so being created can’t easily be said to wrong it.

As long as the child, once created, is treated in a manner that conforms with at least minimal decency (of course, we may hope for quite a lot more than that), then it’s not obvious that the reason for its creation makes much difference.  Kids are made for much less good reasons – or no reason at all – fairly often.

Things might be a bit more complicated if the cells required could only be derived from, say, bone-marrow, so that the retrieval procedure is that much more complicated and invasive – but, of course, that would involve an attitude to an identifiable, discrete, and previously existing being.

But how much further could we push things?

Imagine that A is a young child, and that, in order to save A‘s life, some component part of foetus B would be required, such that B would either have to be dead to retrieve it, or else would be killed by its retrieval.  For the sake of the thought-experiment, imagine that embryonic cells won’t do: the tissue has to be specialised in a certain way.  Would it, then, be permissible for a woman to conceive a child that she does not want in order to terminate the pregnancy a couple of months in for the sake of A‘s survival?

On a very simple account, we might simply wonder why we’re sacrificing one life to save another: the net result, one living child, is the same.  But this assumes parity between A and B, and we can’t help ourselves to that, because B is a foetus and A is an emergent person.  (In order to avoid problems about the difference between the rights of the late abortion and infanticide, let’s stipulate that A is no longer a non-person, and that B is too young to survive outside the womb.)  It also ignores the possibility that the woman in question might regard the procedure not as gestating a baby, but simply as providing some means to A‘s survival from her own body – slightly more radical than breastfeeding, perhaps, but not categorically different.

Would she have made a mistake?  Would there be anything obviously wrong with gestating a foetus wholly for the sake of a child?  As long as there’s no suffering involved, or less than there would be should A die, utilitarians might struggle to object.  Ditto some Kantians, on the basis that B is not yet an end in itself.  Other deontologists might say that the foetus is a discrete human, and that all humans have an equal moral status and certain rights; but buying this line requires accepting a number of fairly strong pro-life positions, and so might only be a minority view.  Finally, some virtue theorists may deny that the foetus is the proper focus, and say that the woman is blameable because of her callous attitude to B – but others may equally well say that her choice is between saving A and not doing so, and that it is comprehensible on that basis, and defensible just because A and B are not, and cannot be, comparable.

I’m not sure if this example is the best that could be produced – but still: the question I want to raise is simply this: how far can the creation of saviour siblings go, and are we obliged to keep saviour embryos and foetuses going until birth?  It strikes me that the normal concerns about abortion might be relevant here – but also that they’re insufficient to deal with the question wholly.