Does this make it different in any meaningful way from an agglutinative morpheme? No, obviously not. Whether to call a language "agglutinative" is already more a question of cosmetics than facts. It reminds me of the feature tagging guidance on Universal Dependencies, which notes that no language can ever simultaneously have "gender" and "noun class" features, because they are the same thing. If there are three or fewer, the feature is called "gender"; if more, "noun class".
Etymologically it's a past participle: three days agone (gone) => three days ago
There are some similar constructions:
- Three days hence. (archaic)
- Three days later.
- Three days beforehand.
- Three days afterwards.
I failed to think of one that didn't have to do with time.
You could think of "three days later" as being supposed to have a complement supplied to later, as in "three days later [than that]", but interestingly enough this isn't possible for "beforehand" or "afterwards".
---
postscript: I think this can be conflated with a general syntactic possibility in English. I can describe an establishment as being "one floor up" from some other contextually-determined establishment, or, as with later, I can make that relationship explicit by saying "one floor up from [wherever]". This is also similar to the bog-standard measurement construction that gets you phrases like "three feet tall".
Called a ‘postposition’. And that’s precisely what ‘ago’ is!
Agglutination is a different concept.
