——
Imagine there was a small-ish device that everyone could carry that would render them impervious to any attack. But it was a bit too big to be convenient, due to a single component, which if we cared to investigate we would see could be easily amended. So a person did not carry it, and was attacked. Where should we place the blame?
It seems eminently sensible to blame that single component, since that is the clear ‘salient improvable’ focus. Is there still a sense that somehow the attacker must really be the proper target of blame? But what is blame for? Any conception of blame must be very misguided that detaches it from what we can most do to have a material effect on the world.
That example worked because of its isolated and malleable component. But the complexities come with the causal network not being easily modularisable and calculable. And so we can discern that what imparts moral responsibility is this: being deemed (by an accepted general non-moral explanation/model) to be the ‘salient improvable’ in a scenario. It is not “when there is some moral agent whose action brought about some event”, but how we choose a focus of intervention to achieve certain aims in the future. The important clarification here is that it has nothing to do with free-will, agency, sentience, or rationality ‒ it can apply to anything.
The minimal exemplar is to be able to trivially point to the single object that is costlessly improvable, as in that earlier example. But real cases spread this over a causal network, so that evaluating the cost and benefit is a problem of gathering and weighing, and choosing simplifications, so that evaluation is tractable. As an exercise, one might return to the example and imagine the attacker replaced by a tiger, or the victim by a crow, and ponder how replacing those sections of the network change our evaluation of where the salient improvable blame-focus might be …
Why favour this somewhat technically inflected attitude/approach to ethics? Because it is capturable as software, and that is how we will manage our oncoming large-scale systems of interaction of human and non-human agents.
——
Reference: