Unicode is a lot like a dictionary, in that both things exist to allow people to work with documents that already exist, whether the tool likes it or not. Dictionaries let you understand existing texts, by defining the words in them. Unicode preserves and faithfully transmits existing texts, by defining and fixing the meanings of the codels of those texts.

Even if a dictionary didn't define a particular word, people would still use that word. You just wouldn't be able to find out what that word means from your dictionary. The word wouldn't be harmed; only the dictionary would be made less useful.

Likewise, even without an assigned codepoint for an emoji, people would still create encodings of it—in chat programs and the like. They'd just be proprietary encodings that wouldn't be able to be copied-and-pasted to other software, and would likely suffer bit-rot. (Can any program that exists today—and that runs on a modern computer—correctly parse out the emoji-like symbols from the binary transcript files of a 90s IM program like AIM or ICQ?) The emoji symbols—at least at the time—wouldn't be harmed by this (people would still use them just as often); only Unicode's goal of "one universal text encoding" would be harmed.