Update: all of the suggestions have been implemented, thanks @Scortchi.
A couple of weeks ago a suggestion came up on meta to merge [natural-language] and [nlp] tags: Do we need the [natural-language] tag, separate from [nlp]? The Q has 6 upvotes, the A has 8 upvotes. Nothing happened.
In July @Dougal finished a massive work to eliminate the [kernel] tag and posted a specific suggestion of how to merge/synonimize/rename other kernel tags: The [kernel] tag is dead. It got 14 upvotes. Nothing happened.
In June @gung posted two suggestions in the "main" thread Current tag synonym candidates that seem very clear-cut and currently have 6 upvotes each. Nothing happened.
One and a half years ago I posted this suggestion: Let us merge [mixed], [mixed-effect] and [mixed-model] tags that got 25-1=24 upvotes. Nothing happened.
I am concerned by how these discussions go. In none of these cases (!) was there a negative opinion voiced. So nobody seems to be objecting, a lot of people seem to be agreeing in the comments and via the upvotes, but no action is undertaken.
Are moderators waiting for a certain number of upvotes to accumulate? I guess there are no strict rules here (?) and the required level of explicit consensus might depend on how controversial the suggestion might appear to moderators. This makes it ultimately up to a subjective judgement of moderators -- which is completely fine with me, but then it would be helpful if moderators explicitly stated their opinions in such discussions. Currently it is difficult to understand why some of the suggested tag changes are swiftly acted upon and some others get completely stuck.
More importantly, what can we do about it? How should we deal with tag moderation more effectively?