In ad copy for the Michael Lucas raunchy gay porn film The Wetter the Better, this summary of some hot-hot man-on-man action (not perhaps to everyone’s taste, but this posting is about syntax and semantics, not watersports, as piss play is delicately referred to in some contexts):
Morgan Black spices up his sex life with Christopher Daniels by soaking him in piss before fucking each other.
Two sentence-final subjectless predicational adjuncts there, and they both need something to supply the referent of the missing subject (they are SPARs): by soaking him with piss, which picks up a referent for its missing subject from the subject of the main clause, Morgan Black; and before fucking each other, which the writer of the copy clearly intended to pick up a referent for its missing subject from the *combination of* the subject of the main clause (Morgan Black again) and the oblique object in that clause, Christopher Daniels. The first exercise in referent-finding is just the default Subject Rule for these things, so there’s no issue. The second exercise in referent-finding is non-default, requiring before fucking each other to be interpreted as ‘before they fuck each other’, where they refers to the set-theoretic union of Black and Daniels (and the semantics of each other then tells you that Black will fuck Daniels and Daniels will fuck Black, as indeed happens in the flick — this is called flip-fucking in the trade). I understand the writer’s intent, but the non-default SPAR is beyond my comfort zone in this case. A dangler too far.
The background here comes from the interpretation of overt anaphors (as I keep pointing out, SPARs involve zero anaphors), as in examples like:
Morgan Black spices up his sex life with Christopher Daniels by soaking him in piss before they fuck each other.
The anaphoric pronoun they here doesn’t have a simple antecedent NP to supply its referent; instead, the referent has to be composed from the referent of the subject Morgan Black and the referent of the oblique object Christopher Daniels. Such examples are known in the literature on anaphora as “split antecedent” cases, and they present a direct challenge (one of many) to the traditional, initially attractive, but deeply mistaken idea that anaphors are literally replacements for their antecedents (just devices for avoiding the burdensome repetition of full NPs). Split antecedents are, in general, not remarkable.
So now we see split antecedents turning up in cases of zero anaphora, where they can be problematic for understanding. Crudely, the issue is that an overt anaphor like they supplies some information about the intended referent (more than one, not including speaker or hearer), but a zero anaphor as in a SPAR lacks these clues and so makes interpretation just a bit harder.