Conjunctions and breath groups

[Yiuel Raumbesrairc]

In my Back-Translation of Xurnese Dzusuisi thread, I asked a question about comma placement. At first, zompist thought it was simply an orthography issue, which would therefore be irrelevent to Xurnese, not using modern Earth punctuation. But I was asking for something deeper.

I speak French natively and English half-natively. I also learned Japanese in way which makes me sound native to native ears. Yet, when I speak all these three languages, I noticed that conjuctions were either part of the preceding phrase or the following phrase.

In English, I invariably link the conjuction to the following phrase. In Japanese, the opposite. In French, not defined, mostly following, and it depend on the type of linked phrases (NPs seem invariably following, IP seem to be hesitating, though I tend on preceding). I'll give here the three sentences I gave zompist when further explaining my question :

I am who I am but I don't seem who I seem.
Je suis qui je suis mais je ne semble pas ce que je semble être.
Watashi ga watashi de aru ga watashi ga watasirashiku ha nai.

* For the purpose of these examples, I took out commas, and will use roman orthography for Japanese.

I cut here :

I am who I am / but I don't seem who I seem.
Je suis qui je suis mais / je ne semble pas ce que je semble être.
Watashi ga watashi de aru ga / watashi ga watasirashiku ha nai.

As for simple NPs

You / and I
Toi / et moi
Kimi to / boku

zompist, looking in some of his books, found Haj Ross positing that no language cuts on both sides and that invariably a conjuction links more closer to one or the other. But he didn't have much more. Do any of you know if there is any cross-linguistic research beyond hypothesis?

[zompist]

The short discussion of the issue in McCawley's The Syntactic Phenomena of English is section 16a. The Ross book he's referring to is Infinite Syntax, if you have a good library available. I don't have it, so I don't know what languages he considered.