1. Not allowing a down-vote is the worst kind of "group-think." It is impersonating the various social media sites that most of us fled to come here! (Twitter, Facebook, and other big, dumb sites)

2. Additionally, it is not the way any intelligent, feedback-oriented system operates in nature. If animals on Earth lacked pain receptors, they couldn't even have risen past the intelligence of worms.

3. The entire discipline of Cybernetics tells us that systems that don't allow negative feedback are "less intelligent" than systems that allow both positive and negative feedback.

4. From a political and cybernetics perspective, (A First-Amendment Cultural perspective that favors classically liberal or "libertarian" western civilization) no information should ever actually be censored. And even stupid, weak, poorly-written posts should not be "minimized" or "removed."

5. From an "optimized internet" perspective, only those posts labeled (done with conscious effort, not as a value assigned to a certain number of downvotes) as "SPAM" can legitimately be minimized, because that's solving a "perverse-incentive-filtering" problem, not a "same-incentivization-level" filtering problem.

6. Do both! In any case, it's possible to allow down-voting, and retain upvoting. Simply don't remove the separate record of the upvotes, and allow filtering by both! (Heck, for that matter, one could allow users a slider-bar that divides the weight between positive to negative from 1 to zero).

For more information, I recommend reading "Out of Control" by Kevin Kelly to get a sense of the value of feedback and correction(including negative feedback).

Additionally, Marvin Minsky's paper on "Negative Expertise" is probably vital to this discussion: http://web.media.mit.edu/~minsky/papers/NegExp.mss.txt