> I'm not sure how what I mean for 'us' is difficult to grasp. Human societies?
I think the idea they're getting at is that collective preference isn't always well defined / well-define-able.
You're aware of Condorcet's paradox, right?
(Generally, one expects a person's preferences to be transitive, I.e. , if you prefer a to b, and you prefer b to c, then it is generally expected that you will prefer a to c. However, if you have a collection of people, all of whom satisfy that property, and have the group vote on what the "group preference" is between each pair of options, the "preferences of the group" will not always satisfy the property.)
This, along with a number of similar things, seem to suggest that maybe, at least sometimes, there is no "group preference". That the group cannot be treated as if it is an individual with coherent preferences.
So, if "we" can't always be treated as an individual who has coherent preferences, then maybe there's no coherent way in which "we" prefer something else to what the lawnmower does, in some cases.
Now, maybe in many practical cases, this isn't an issue. Idk. I mean, it certainly seems like humanity would "collectively prefer" not to be destroyed by nuclear war, or tortured to death, or whatever. Maybe even for most important things. I think the question would then be "in which cases are societies 'collective preferences' well defined, such that there is an achievable outcome which is 'collectively preferred' to what the market outcome would be?". This seems like an important question to me.
Another thing I find relevant is the VCG mechanism for social choice.