Paul, I take it, L1 and L2 refers to "language 1" and "language 2." Not to Layer 1 and Layer 2, as in digital protocol layers. And I take it, you are referring to spoken or written languages? I don't know of studies, but here's my take.

In some digital protocols, including Asynchronous Transfer Mode (ATM) and anything to do with Ultra Wide Band (UWB), the receiver keeps looking at the arriving symbols until it sees a known pattern. When it finally sees the known pattern, it synchronizes all subsequent symbols to that known pattern.

I'm referring here to protocols that do not have one consistent "well known pattern" regardless of transmission. The latter is a much simpler case.

In written and spoken languages, I think it works much the same way. You won't see any well-known pattern to begin every sentence. Such as, you do not see anything like

Universal Sentence Start symbol

Universal Language Code symbol

Then the sentence

And perhaps even, universal End of Sentence.

Instead, the receiver scans through (or listens to) the symbols and quickly determines if there's a pattern it is expecting. For spoken languages, there are many clues, including intonation, which can provide very fast cues. In written language, it may possibly take a little longer, if you only see very few words, because some languages have identical words, spelled the same way. But you'll soon see other words, that give it away.

If I understand your question, and didn't go off on some tangent, it's all a matter of pattern recognition. Humans are great at pattern recognition. In fact, it's quite simple to mix languages, once the two parties have agreed which languages they will mix. Sentences consisting of words from multiple languages. No problem. It's just like expanding the dictionary of the one language. In my family, we do it all the time. Met a Russian family this past weekend. They also do this all the time.

Hi Albert,

Awesome to get a cross-disciplinary look at this phenomena. If I understand what you are saying, if five Japanese students of English are speaking together in English, then even fillers in the first language (yes, L1) like "ano" and "eto" and "de" instead of "Um," "erm," and "hmm" could be part of the pattern that starts triggering code switching back into Japanese. Japanese and English might not be dissimilar enough as an example, actually, since there are so many borrowed words from English in Japanese these days. I need to do some more research on this...

Exactly, Paul. It works. And the funny thing is, sometimes the solution is more universal than one might have expected. One example:

When school kids get home from school, they may have a small afternoon snack, before tackling their homework. In Italian, it's called "merenda."

So, naturally, when speaking English, it became "merend," in our family. Just throw the word in the sentence, no problem. The Miriam-Webster was effectively increased, by one word.

Come to find out, we were not the only ones who added that word to the dictionary. One of our teachers, who grew up in Australia, same two languages, had invented the same word. Weird.

Then again, we do say things like raison d'être, or fait accompli, in English sentences, and that has become common enough. As much as weekend or even sometimes fun became used in French (much to the chagrin of the Académie Française). Or for Japanese, I think the use of sayonara is fairly common, in English and in other languages.

For what it's worth, I believe that the English spoken in India is crammed full of this sort of local infusion, mixing in Hindi or other words, into English sentences. You see this in Bollywood movies.

Nice Contribution Albert Manfredi