I know for example, Mandarin (Chinese) has it like from this sentence: 我睡了八個小時 (I slept for 8 hours), the hanzi: 覺 is omitted in this sentence because 睡 is already a word with its own meaning “sleep” which in itself conveys the meaning of the verb. There are “Euro” languages where separable verbs exist such as Dutch or German for example. Is there an equivalent of that in the English language though?
There are also working memory constraints, or similar real-time processing constraints, at play which can affect how acceptable (in terms of native speaker intuition) some examples are.
I won’t get into the possible differences in the intervening structures (e.g. the relative clause), but as a mostly fair generalization: in these cases, the preference to split the verb into parts is inversely correlated with the number of intervening elements.
Something similar occurs without phrasal verbs but with indirect objects optionally expressed as prepositional phrases:
The similarities suggest that the pattern has more to do with heavy intervening phrases, or even more generally syntactic headedness, and not wanting to rebuild or finish building earlier structures later in the parse.
Ooh, this is a very interesting line of inquiry! Is there any evidence of differences in the working memory of language speakers where there is more tolerance on the amount of words between separable verbs, like the 50 word intervals in German mentioned elsewhere? Your longer examples are starting to get very Silas Mariner energy.
Yeah, I’m not sure it’s entirely relevant to the existing conversation, but maybe at least in part.
I don’t have anything to cite directly in front of me for other examples from non-English languages, but I know some work exists out there, particularly in cognitive linguistics. If I remember right, some papers under the “construction grammar” umbrella touch on this in an effort to explain relative preferences between what would otherwise be equally weighted constructions. Also look into works by Dryer and Haspelmath (together and separately) for similar discussion with a more typological bent.
German does have one typological distinction which might also play a role in how permissive it is with seperable prefixes: the verb-second constraint, which basically exists in the Germanic branch of Indo-European and only a handful of languages elsewhere. This is why German allows basic clauses which are SVO (e.g. Der Hund frisst den Fleisch), but as soon as there is an auxiliary verb, it looks more like an SOV hybrid with the main verb at the end (der Hund wird den Fleisch fressen), or fully SOV if a subordinate of a matrix clause (Wir glauben, dass der Hund den Fleisch fressen wird).
That doesn’t really explain why you have to do ich sage ab instead of the ungrammatical *ich absage, but having to move that ab to the end is at least reminiscent of having to move verbs in the same way. I don’t think there’s even an alternative, so extremely long examples persist by necessity.