The problem seems to be that people will occasionally reflexively use the upmod arrow on an article just to say "hooray for the sentiment expressed in the headline" rather than after carefully thinking about whether it's worthwhile.
Examples include this article (I'm guessing that most of the upmodders were Obama supporters), the "congrats to pg on getting married" thread the other day, and the "happy birthday somebodyorother" thread a few weeks ago.
Pardon the brevity of my previous post. I'm sorry if it appears that I'm arguing for it's own sake. I only meant to indicate that we currently have an implemented definition of "better" between posts--the HN ranking algorithm. I believe that the burden of proof belongs to the party challenging the status quo to provide an alternative and then argue why the alternative is indeed the preferred system.
During the debate over which ranking algorithm/definition is better where I believe the Luttgenstein/Chmosky argument is most applicable because we don't have a clearly agreed upon definition of what a better ranking algorithm is. However, until an alternative is provided, we cannot possibly reach and/or explore these semantic differences.
Why I use the implemented definition of "better" in favour of my (admittedly completely subjective) definition?
If you surveyed people's definitions of "better [thing that they care about]" you'd get a lot of different answers, but I suspect few would say "more people like it" (even if an algorithm based on "more people like it" would be good at finding things they think are better).