This morning’s post at Language Log on code switching reminded me that I intended to write about an instance of code switching by a friend of mine that I was fortunate enough to witness.

This friend is South African and her first language is Afrikaans, although she has been speaking Australian English for long enough that she only occasionally appears to have a twang of an accent. She still speaks Afrikaans with members of her family, as all live in Australia and speak often.

I was doing a favour for said friend which basically entailed my sitting in the passenger seat as she drove from her house to the RTA in her “Smart” Car. I put ‘smart’ in double quotes for good reasons, which shall become clear below.

While we were en route, we had the misfortune of running over a nail, which caused one of the rear tyres to deflate, which we noticed only after it was too late; the tubeless tyre was shredded, and would need replacing. My friend’s driving test is minutes away.

“No worries,” I said, “I’ll put the spare on.”

“There’s no spare.” was my friend’s reply. “Smart cars don’t come with a spare,” ironically.

With no other option, we limped into a service station and asked whether they could fit a new tyre. The reply was that Smart Cars, being so terribly smart, use a slightly different sized tyre than any other car; their wheel size is absolutely unique and, owing to the minority of the Smart Car market in Australia, tyre suppliers don’t generally keep them in stock. I hope you can see now why I put ‘smart’ in quotes.

To cut a long and largely irrelevant story short, my friend had no way of taking the test that day, so we set off back to her house. On the way, she phoned her brother to tell him, in Afrikaans, what had happened. Now, my Afrikaans is about as good as my Walmajarri, so I won’t try and transcribe it here, but when she related to her brother the cost of a new tyre, she did so while code switching into English.

As her Australian English accent is so good, she’d normally have no issue saying “a hundred and eighty five dollars” as [əˈhʌndɹədnˌeɪɾifaɪv.ˈdɔləz], exept that it came out as a particularly stereotypical Afrikaaner [əˈhʌndɹətəˌnaɪtifɒf.ˈdɔləs]1.

This is interesting to me because I’ve barely done any psycholinguistics or bilingualism in my undergrad studies, so I enjoy it when I come across cool little bits of evidence that allow me to make broad generalisations about the mind and the language faculty, such as the following.

This implies to me that my friend, and bilingual speakers in general, have an L1 bit2 of their brain and an L2 bit. Each bit contains a lexicon; the vocabulary of each language, and each bit contains a phonology. So, what happens during code switching? From this I’d take a naive guess that code switching is the act of moving out of the Lx bit, and taking a word from the lexicon of Ly (in this case she moved out of her L1 to select a word from her L2’s lexicon).

A question emerges here; where does a word’s phonetic representation come from? I would have previously thought (again, naively) that the mental lexicon contains the phonetic representation, much as a dictionary entry contains an IPA transcription. But here, the borrowed words are fed into the phonology of the borrowing language, so the words don’t bring their phonology with them.

My broad and uneducated conclusion then, is that within one’s L1 or L2 is contained separate modules of language: a lexicon, a phonology, a syntax and all the rest of it, and when speaking using L1, you use all the modules in that language, diverging from them as little as possible. So code switching allows words to go between L1 and L2 or vice versa, but the phonology being used is still that of the language you’re speaking.

If you were fluent in two languages and code switched from one to the other, I believe it would take a conscious effort to use those borrowed words with their ‘normal’ pronunciation, by which I mean, the pronunciation they usually take in the language they belong to. Conversely, if you’re a learner of a language, you haven’t yet formed a distinct and independent L2, so the pronunciation of the new language is all a conscious act, in which case, when code switching back to their L1, they’d still use their L1 phonology.

The more I think about this, the more it appears to be commonsense, so I apologise if, for instance, your an expert in bilingualism and either a) you’re wondering why anyone would other writing a thousand words on something so natural, or b) I’m completely wrong.


  1. My apologies if you can’t read IPA; just trust me that the way she said it was almost what I’d expect of a satire.
  2. For want of a better term. I realise that there’s no single bit, but I’m talking abstractly.