Fix broken binary compatibility

[travisbrown]

Reverts #3100, #3105, and undoes part of #2834. This unfixes #2891 and #2701, which I think will have to wait for 3.0.

In the case of #2834 I don't think the move was necessary, since the CommutativeMonoid one is more specific. In any case moving the other back doesn't break any tests and seems fine:

scala> import cats.kernel.instances.all._
import cats.kernel.instances.all._

scala> cats.kernel.CommutativeMonoid[Int]
res0: cats.kernel.CommutativeMonoid[Int] = cats.kernel.instances.IntGroup@5c7654d1

scala> cats.kernel.Monoid[Int]
res1: cats.kernel.Monoid[Int] = cats.kernel.instances.IntGroup@5c7654d1

scala> cats.kernel.Monoid[String]
res2: cats.kernel.Monoid[String] = cats.kernel.instances.StringMonoid@77c600b4

I've confirmed that my quick-fix MiMa branch here doesn't find any problems after this change and #3162.

[codecov-io]

Codecov Report

Merging #3163 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master    #3163   +/-   ##
=======================================
  Coverage   93.04%   93.04%           
=======================================
  Files         376      376           
  Lines        7374     7374           
  Branches      209      209           
=======================================
  Hits         6861     6861           
  Misses        513      513

Flag	Coverage Δ
#scala_version_212	93.37% <100%> (+0.02%)	⬆️
#scala_version_213	90.62% <100%> (-0.02%)	⬇️

Impacted Files	Coverage Δ
.../scala/cats/kernel/instances/OptionInstances.scala	100% <100%> (ø)	⬆️
...n/scala/cats/kernel/instances/QueueInstances.scala	100% <100%> (ø)	⬆️
...in/scala/cats/kernel/instances/ListInstances.scala	100% <100%> (ø)	⬆️
.../scala/cats/kernel/instances/VectorInstances.scala	100% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 55c48a8...d6e940b. Read the comment docs.

[rossabaker]

We should have a discussion of how to make sure we don't forget these otherwise good changes for an eventual 3.0, but that conversation shouldn't block 2.x. 👍

[travisbrown]

@rossabaker I was thinking we'd reopen the issues that were closed once this is merged, which I think should cover it.