Quick poll: supposing that morality and practical (all things considered) reasons can come apart, which of the following two claims sounds more plausible?
(1) Morality requires you to sacrifice your loved ones if this would promote the impartial good, but it would be unreasonable to do so.
(2) Rationality requires you to sacrifice your loved ones if this would promote the impartial good, but it would be immoral to do so.
In other words: Is impartial consequentialism more plausible as a theory of morality or of practical reason?
In Commonsense Consequentialism, Doug Portmore suggests that utilitarianism is widely recognized to be 'unreasonably demanding', in the sense that it asks us to do things that we lack sufficient reason to do. So, he suggests, if we come to accept 'moral rationalism' -- the view that we always have decisive reason to do what's morally required -- we will be led to reject utilitarianism.
But if I had to pick one of the above claims as a starting point, I'd sooner endorse (2) than (1). Impartial consequentialists may endorse Parfit's suggestion that the view constitutes an "external rival to morality". At bottom is the idea that there's no principled reason for favouring your own welfare over others', and so even if radical impartiality bears little resemblance to ordinary "moral" thought, such absence of personal bias is nonetheless what's rationally required, strictly speaking (just as we're rationally required not to be temporally biased, e.g. in favour of the near future).
Whilst taking the fundamental normative requirements to be impartial in this way, the utilitarian might follow Railton in constructing a more commonsensical and moderate "practical morality" that people would do well to follow. Given the familiar pragmatic reasons for introducing norms of partiality (an efficient division of moral labour insofar as we tend to be more motivated and able to help those who are closer to us), this constructed "morality" could plausibly allow for more partiality than the fundamental norms. It might even make it obligatory to look out for your family, even when this means passing up apparently greater benefits to others. This shows one route to claim (2) above.
On this view, we have every reason to prefer the impartially best outcome. It's just that we can't call it 'morally obligatory'. More than that: falling short of perfection is not sufficient grounds for social censure, so in this sense it would be unreasonable to demand that people meet the strict requirements of utilitarianism. But take care: it is the third party's demanding that is unreasonable, not the act thereby demanded. It'd be perfectly reasonable for the agent to act with perfect impartiality. It just isn't reasonable for others to ask him to do this -- a vital difference!