It wasn't but a year ago that I'd heard the word heterosemy for the first time. I was beginning to write an article with a good friend and he started throwing around this new $10 word. I quickly and unashamedly asked what it meant. He explained it something like this (except, take out the US English slang, the Texan draw you don't hear, and add on a funky Spaniard-ish/Icelandic accent and grammar):
Just like polysemy describes the various semantic meanings that might be associated with a particular word, heterosemy describes the different grammatical functions that a single form might be used to convey.
Simple, right? It is. There's no catch and that's not a trick question. It's not a possibility we really think about that much though. We simply use words cause they make sense, not really noticing how their grammatical function might be being manipulated in different contexts.
I think even for us Biblical Language nerds – even though we look closely at different "meanings" of words – we still have this tendency to overlook these types of syntactic maneuvers. We tend to focus on the semantic component of a lexeme and ignore the syntactic function that might actually be the basis for these shifts in meaning.
The example that my friend and I looked at was בלי. What we found was pretty amazing. We looked at every single occurrence and, using typological and diachronic analysis, determined that it carried many different grammatical hats. (Or rather, these different hats wore בלי. Remember, words don't really mean anything by themselves, context and sundry just imbue forms with certain features). To give you a taste of the different usages, look at these.
וְאַתָּ֞ה חָשַׁ֤קְתָּ נַפְשִׁי֙ מִשַּׁ֣חַת בְּלִ֔י כִּ֥י הִשְׁלַ֛כְתָּ אַחֲרֵ֥י גֵוְךָ֖ כָּל־חֲטָאָֽי
It was you that delivered me from [[the pit of] decay], for you put all my sin behind your back (Is 38.17)
Even though this is the only case we have of בלי acting like a noun when it occurs independently (i.e. not connected to another particle), it's still a solid occurrence. And quantity does not qualify quality. We have to remember that we're only looking at בלי as it occurs in a closed corpus. It could very well have been used in more ways than we could identify, or like a noun more than what we get the impression it did, statistically speaking.
זַ֥ךְ אֲנִ֗י בְּֽלִ֫י פָ֥שַׁע חַ֥ף אָנֹכִ֑י וְלֹ֖א עָוֹ֣ן לִֽי
I am pure, [without [transgression]]. I am innocent, with no iniquity in me (Job 33.9)
Here it's clearly functioning as a preposition. It stands before a noun, which acts as its complement.
יָמִים֙ עַל־שָׁנָ֔ה תִּרְגַּ֖זְנָה בֹּֽטְחֹ֑ות כִּ֚י כָּלָ֣ה בָצִ֔יר אֹ֖סֶף בְּלִ֥י יָבֹֽוא
Days upon a year you will tremble, O complacent ones, for the vintage will fail – the fruit harvest [will [not] come]. Is 32.10
Unlike the previous function as a preposition, in cases like the one above בלי stands in front of a verb, not a noun. Likewise, the negative-instrumental value present in the prepositional use has faded, though remnants remain which are seen in the negative (privative) value that serves to simply negate the verb.
While בלי does several other things (i.e. negative affix, semi/genuine conjunction), it should be clear that the heterosemy is indeed diverse and established. If you're interested in learning more about this you'll be happy to note that our 2 part article will be appearing in a journal soon.
Whatever the case, just ask yourself the next time you're looking at a Hebrew or Greek form: are these different meanings signs of different syntactic functions? Is it possible that this particle is behaving like two different grammatical roles at the same time? Maybe so. Categories are made up and their boundaries are fuzzy. Functions develop over time, and in stages – not uncharted jumps. Sometimes we get glimpses of these hybrid stages, even in the Hebrew Bible if you're paying attention.