While I previously defended hypotheticals like this one for the fact that they are isolating a certain aspect, I have to add another thought:
The hypothetical may come with the problem that the isolated aspect is not a major element in my own ethical considerations. Here: the number of people alive may fade in signifance when compared to e.g. quality of the lives (just to pick the first example that comes to mind). The - to me - most relevant criteria are possibly excluded due to the reduction of aspects so that my solution to the dilemma might not at all represent my ethical approach.
Kind of like the hypothetical question "If you had the choice between a red and a white car (without knowing anything else about those cars) - which one would you buy?" does a good job in isolating two colours as a criterium. However, it may be possible that colour is but a minor (if any) criterium in my choice of a car, and thus the question completely excludes me from presenting my approach towards the choice of cars.
Absolutely, but here are my two issues with that.
1. I've generally held that it's good philosophical practice in these discussion to not muck around with the parameters of the thought experiment, and given that it is stated that the scenario
definitely is not your fault and
definitely will kill 5 if unswitched, 1 if switched, the resulting moral inferred will come across as somewhat stark. But that's the point, in reality we would be as we already discussed - afraid, and probably prone to fearful inaction, and would probably expect/aim for some other outcome to occur that was outwith the parameters of the thought experiements. Again, I have seen utilitarians demonised in print for responding honestly to these kinds of thought experiments without any regard given to whether they were applying the same kind of approach to a thought experiment - that is, assume the certainty of the parameters of the scenario as given by the thought experiment itself. This is profound unfair on utilitarians....
2. ....and secondly, if the thought experiment is actually extremely implausible in and of itself, then it is hardly an indictment of a moral system if it yields a "questionable" answer based only on the response to an implausible hypothetical.
The irony is, if people who criticised utilitarianism in this way actually knew what they were talking about, they'd know that within some forms of utilitarianism there is indeed a distinction made between absolute hypothetical knowledge in theory and limited infallible knowledge in practice, and the differences in choices that can yield (I read yesterday that this was characterised by one philosopher as "archangel" vs. "prole").
In a hypothetical scenario with absolute certainty and no other option - darn right I'd hit the switch. In reality, I would not, I would attempt to search for another way. Like honking the horn, which magical-hypothetical trolley cart in the realm of dreams does not have. But I'll be buggered if I'm going to sit here and have someone call my moral system that of a murderer unchallenged because of how I answered a mere question about something which is never going to actually occur in the real world.