When Bilinguals Borrow from One Language to Another

When bilinguals interact with other bilinguals who know the same languages, they may well intermingle their languages through code-switching or borrowing. Code-switching—the shift from one language to another for a word, phrase or sentence—has been the object of much research (see here). This has not been the case for borrowing, where a word or short expression from one language is integrated linguistically into the other (see here). Professor Shana Poplack of the University of Ottawa has just published a much-awaited book entitled, Borrowing: Loanwords in the Speech Community and in the Grammar (see here). She has kindly accepted to answer a few of our questions and we thank her wholeheartedly.

You are known throughout the world for your pioneering work on code-switching. And now you have published this book on borrowing. Can you tell us the reasons behind it?

I came to borrowing as a result of many years of studying code-switching. The latter involves juxtaposing sequences of one language with sequences of another, as in this sentence, “Quand il marchait là, il marchait over dead bodies,” spontaneously produced by one of the participants in our studies. Code-switching is conspicuous, salient and endlessly fascinating. Yet, despite intense scholarly scrutiny, it remains controversial. Analysis of bilingual behavior on the ground suggests that this is a consequence of failing to recognize borrowing in all its manifestations. Speakers, on the other hand, make a fundamental distinction between borrowing of all types and code-switching (CS) (as your own early experimental work suggested).

In contrast to CS, borrowing, as in “Je groovais comme si j’étais à un show de rap” (‘I was grooving like I was at a rap concert’), has been treated as a poor relation. By virtue of its major property, integration, it often passes unnoticed. This is a shame because systematic analysis of borrowing in the 13 language pairs my team and I studied showed that the linguistic and social processes involved are at least as complex and startling as those underlying CS. So I had to write a book about them!

How frequent is borrowing in bilingual discourse as opposed to code-switching?

Ironically enough, given the neglect it has suffered, borrowing turns out to be the major manifestation of language mixing by far. In the French-English materials we studied, it outweighs CS by a factor of 20. In many bilingual data sets, that’s all there is. I regularly send my graduate students out to collect samples of CS from communities they describe as rife with mixing, and nine times out of ten they return with nothing but examples of lexical borrowing.

Are some bilinguals code-switchers and others borrowers? If so, what accounts for being one or the other?

Definitely, and for good reason. The ability to access a sequence like “over dead bodies” necessarily requires knowledge of English. The ability to juxtapose it with a French sequence so as to result in a sentence that is well-formed in both languages simultaneously requires (even greater) knowledge of both languages. And indeed, our research has shown, not surprisingly, that CS is the province of the most proficient bilinguals.

A borrower, on the other hand, doesn’t need to know the other, donor, language at all. English speakers, for example, utter words borrowed from Italian, Spanish, French and Japanese every time they say “espresso,” “arroyo,” “diamond,” or “tsunami” respectively, whether they know it or not. We studied bilinguals who engage in these strategies, as well as spontaneous (“nonce”) borrowing of novel words, and found no correlation. Copious code-switchers are not necessarily copious borrowers. Rather these strategies, and the speakers who favor them, are independent.

When a bilingual borrows a new word from the other (donor) language, it is termed a nonce borrowing. Can you tell us about the linguistic operations that take place when she produces such a word in the other (recipient) language?

When English speakers use old established loanwords like “terrace,” “boil,” or “court,” they are often unaware that they were originally French. This is because these words have been refashioned according to English grammar, taking English plural markers (“terraces”), English verb endings (“boiling”), and entering into English word orders (“criminal court”). Such integration of donor-language material to the morphology, syntax and optionally, the sound system of the recipient language is the major mechanism underlying borrowing.

The actual linguistic operations involved are those of the recipient, and can, therefore, vary wildly from one language to the next, depending on the specific grammatical properties of each. These may include assigning a gender if the recipient language features that category (“la drop”) or applying complex vowel harmony rules (“nà-a-hallucinate”). These characteristics are amply evident in the loanword stock of all languages.

Our studies of nonce borrowing on the ground show that speakers also appeal to them when incorporating other-language words spontaneously: they treat novel borrowings exactly like their established loanwords (by imbuing them with the grammar of the recipient language) and distinguish both from their CS, which retain the grammar of the donor language.

When do nonce borrowings become established loans (which you term loanwords), what linguistic transformations, if any, have they gone through to get to that endpoint?

Thank you for asking that question! I’m actually in a position to answer it, based on analysis of bilingual speech data spanning an unprecedented century and a half. One of our most startling discoveries is that nonce borrowings are not transformed into loanwords gradually, as scholars have long believed. Instead, bilinguals decide whether to borrow (as opposed to CS) right off the bat. If they opt for the former (which by the way, they almost always do), they imbue the word with the full complement of recipient-language grammar. Remarkably, they engage in these complex operations at their very first mention of the other-language word!

It’s the social integration of borrowed items that is gradual. As the word diffuses across the community, it stands a greater chance of becoming a bonafide or established loanword. Still, in the grand scheme of things, this almost never happens; another surprising finding of our research.

As bilinguals, we produce many nonce borrowings but they are ephemeral and very few make it into the language as established loans. Why do some obtain this distinction?

That’s the $64,000 question! We discovered that the overwhelming majority of borrowed words disappears after the first mention. We don’t know which will persist or why. All we can say is that the received wisdom—that words designating cultural items like “hamburger” or “yoga” get borrowed, while core words like “mom” do not—definitively doesn’t hold. Of course, all languages feature plenty of such words (“pizza”, “jihad”, “origami”, etc.). But their numbers pale in comparison to the body of loanwords that don’t designate such things.

The words that persist over time and go on to achieve the status of established loanwords often turn out to be the least expected. Why do we need to borrow words like “friend”, “weird” or “game”, as the Quebec francophones we have studied have done, when there are perfectly good French words for them? We don’t! Need is not the motivating factor in borrowing. Rather, words are borrowed through implicit community compacts and become part of the community norms.

These norms may differ from one community to the next, even when the same languages are involved. This is why you hear “parking”, “shopping” and “weekend” in France, while in Canada, “stationnement," “magasinage,” and “fin de semaine” are de rigueur (a French loanword in English). Canadians prefer “chum” to “ami” and “cute” to “mignon.” When it comes to language mixing strategies, the community rules.

For a full list of "Life as a bilingual" blog posts by content area, see here.

Photo of roadsigns from Shutterstock