Flagging was originally downvoting, which as the whitepaper describes, was intended as a tool to bring down over-inflated rewards. Our vote weight is supposed to be able to both say "I think this should be rewarded more" and "I think this should be rewarded less".
I think it would be ideal to have some smarter voting options available. For example, "I think this deserves $X. My vote will count towards it up until the point of $X, but not further".
I think after they changed downvoting to flagging, they should have revised the whitepaper to say flagging was for spam only. Reducing payouts seems to be censoring BS to me. If a user doesn't get paid anymore, chances are they may stop posting.
they did not change downvoting to flagging. They changed a graphic in the UI. AFAIK, nothing at all changed on the steem blockchain.
The front end change has resulted in a semantic change for the community. Especially with the guidance added on SteemIt.com, about flagging being for abuse. This change in philosophy was never implemented consistently even among the founders though.
Stakeholders are entitled to believe, express and vote that a post is being given too large a share of the rewards pie. To disallow or discourage votes which say "this is rewarded too highly" is censorship in itself.