IIUC, this is precisely why the n^2 curve was adopted... the idea was to create a two tier protection against abusive self-voting.

For whales (who have enough influence to use their vote to assign non-trivial rewards on their own) the idea was that they would police each other with downvotes.

However, the assumption was made that non-whales would not be able to be organized enough to police themselves, and that the whales would be unwilling to devote a level of time and energy sufficient to police all of the non whales. So an n^2 curve was adopted to make sure non-whales did not have the power to assign a significant amount of rewards without consensus.

These assumptions, however, have not been borne out by time. There are institutions on steemit (like steemcleaners) that monitor such things. For example, the recent ubg spam thread and the "whitespam" thread by steemvoter. The latter was getting downvoted even in the absence of upvotes as a prophylactic measure.

Even with mere linearity, the typical user would still have to spam comment, then upvote the spam to get a non-trivial portion of the reward pool. this plan would die in the womb when such spam was detected and downvoted by observant community members.