Simultaneity in natural languages?

Question

Is there any known natural language in which it is possible to express grammatically—i.e., not through emotional tone or other secondary traits—multiple parallel channels of meaning? Parallel structures could be made explicit in such a language by simultaneous articulation.

Consider a hypothetical language in which words can overlap:

• [ʔņ] = “have, there is”
• [po] = “apple tree”
• [ka] = “rain”
• [k͡pʼœ̃] = “it is raining and there is an apple tree”.

(In which [p] and [k] become coarticulated, [ʔ] becomes ejective colouring, [o] and [a] combine to form [œ], and the syllabic [ņ] nasalises the vowel.)

The process is essentially synthesis with a pathologically high degree of sound change.

Naturally this is a contrived example, and a phonologically unusual one; I’d expect a language that has anything remotely resembling this to have some provisions for keeping things manageable.

This doesn’t seem outside the realm of possibility, and in my experience with conlangs, I’ve learned not to be surprised when natural languages turn out to be far weirder and more wonderful than anything an individual can come up with.

Answer 1

I don't know of an example quite like that from spoken language, but Singleton & Newport (2004) discuss verbal morphology in American Sign Language, and according to their discussion verbs of motion carry seven morphemes. Here's their Figure 1.

From the description, it seems like one or more of these morphemes are articulated simultaneously.

---

Then, there's cases of infixation and reduplication, where the pronunciation of one morpheme is interleaved with the pronunciation of another. A fun example from English is fan-fucking-tastic. I'm not sure of that meets your criteria of "parallel channels," since it is more a mixture than a blend.

---

Morphologically triggered stem changes could also be a good example. For instance in sang, the phonological process ɪ -> æ indicates the past tense, and is pronounced simultaneously as the lexical material indicating the verb "to sing."

Answer 2

i.e., not through emotional tone or other secondary traits

Intonational contours have aspects that are highly grammaticalized and not just "emotional" in many languages (e.g. in English: contrastive focus/rise-fall-rise, list closure intonation, rising intonation on questions, etc). The semantic content isn't of the same sort as occurs with content words, which may be what you're looking for?

Possibly root and pattern morphology, where a (multi-)consonantal root that is e.g a N or V is interleaved with vowels providing inflectional information.