Redundant tags I has been made more specific to overtly specific tags. Here are some which don't even seem to be about physics.
x-ray-crystallography This isn't very specific. Common technique in experimental physics.
laws-of-physics Slightly metaphysics-y, but not that bad a tag. it's ambiguous though, I guess we could get rid of it. Not too certain about that
approximations Sort of a meta tag in the sense that it doesn't denote a topic (however that doesn't seem to be the tpe of met tag that's disallowed, since most other sites have similar tags). But otherwise approximations are there in many different physics topics.
complex-number similar as above. Used in many places (circuit analysis, optics are two examples I can come up with)
affine-lie-algebra Probably too specific as it has only 2 questions. I don't know how broad the topic/application is.
laboratory-safety is on topic (I think), and a pretty interesting tag to have.
Here I do not quite agree that all of the tags mentioned in this question should be burnicated (generally I think all the tags we have should denote meaningful physics terms):
x-ray-cristallography is a meaningful method to investage the structure of materials in the field of material sciences, so it seems a good one.
laws-of-physics seems a bit unspecific
approximations seems ok for questions that ask about about
particular approximations needed in different contexts
affine-lie-algebra is a good meaningful technical tag for
technical questions (it came from TP.SE), even though it is not
heavily populated I still dont give up that the kind of questions
that needs this kind of tags will increas again ;-)
labor-savety looks slightly off topic to me ..