It’s OK if you can’t pronounce foreign words

Early in September, Vanessa Ruiz, a news anchor at 12News Arizona, caused a mini-controversy with the way she pronounces Spanish words on air. Ruiz is a native speaker of Spanish, and viewers were getting upset that she rolled her “r” when saying words of Spanish origin. She defended herself by claiming that she was only pronouncing the words “the way they were meant to be said”, although not everyone appreciated this response.

Ruiz isn’t the first person to get caught up in a pronunciation argument. Gustavo Arellano, who writes the syndicated column ¡Ask A Mexican!, expressed similar opinions in an interview with KQED News back in February of this year. He said that pronouncing words as they would be in Spanish is the “correct way” to do it, and further claimed that English-speaking Americans “don’t think twice about butchering Spanish”.

President Obama has long been criticized for his pronunciation of Pakistan where the first syllable rhymes with “sock” instead of “sack”. Republicans in particular have reacted negatively, and this features in many Obama-is-a-secret-Muslim conspiracies. People more sympathetic to the President claim it is because that’s the way that people in Pakistan say it.

The idea that the original pronunciation is always correct deserves some scrutiny, and the accusation that English speakers are doing harm to other languages is just plain unfair. Let’s be clear: When words are borrowed from other languages, it is entirely normal for the pronunciation to change. It’s not a matter of laziness or an uncaring attitude.

Part of the difficulty is that there’s a political dimension to pronunciation, especially in the United States. Choosing an English or Spanish pronunciation can signal a lot about someone’s attitude toward issues like immigration and multiculturalism. So let’s consider the reverse situation: What happens when other languages borrow English words?

Japanese is a good example. Speakers of Japanese love using English words, and a large number of these words have become a part of the everyday vocabulary. However, when words get borrowed, they almost inevitably change. The word disk was borrowed into Japanese as disuku, the word taxi became takushi, and the word thrill was borrowed as suriru. These pronunciations aren’t very faithful to English. If we’re going to apply the standard of “the original language is the correct pronunciation”, then it follows that Japanese people are saying all these words wrong.

There are many other examples just like this. Due to the reach of English-speaking cultures, English words have been borrowed into languages around the world. The word hamburger when borrowed into Russian has an initial /g/ sound, plus a rolled-r, plus some vowel changes, and it’s more like “gamberger”. Yoruba, a widely-spoken language of north-west Africa, borrowed the word revolver as rifofa. Where did the /v/ and the /l/ go? Should we say that the Russians and the Yoruba are wrong, and they have been mispronouncing English words all along?

This is not even something unique to English. Languages everywhere borrow from each other. Zulu, spoken in southern Africa, has borrowed many words from the neighboring Khoi-San languages. For example, the word for grave in Zulu is i-ngcwaba, but this is originally a Khoekhoe word |hoba-b. Should the Zulu modify their incorrect pronunciation?

Quechua, the language of the Incan Empire which is still widely spoken in South America, has borrowed numerous words from Spanish. The Spanish word meza, meaning “table”, became miza in Quechua. Spanish forketa, meaning “pitchfork”, was borrowed as xurkita. Would Vanessa Ruiz be willing to tell Quechua speakers they aren’t pronouncing Spanish words the way they are meant to be pronounced?

This is completely silly of course. It is normal for words to change pronunciation when they get borrowed into another language, and it isn’t something we should be criticizing anyone for. This is a world-wide phenomenon that has been around as long as language has existed. It is so common, linguists have a term for it: loanword adaptation.

Loanword adaptations are interesting because they follow patterns. It’s not as though people just randomly change sounds, or add new vowels for kicks. Certainly it’s not a matter of laziness, or not caring, and it isn’t people “butchering” the other language. Research has identified a number of ways that loanword adaptation can happen, and each type of change is based on a specific difference between languages.

One cause of change is that borrowed words might contain a sound that is not found in the borrowing language. When this happens, the foreign sound undergoes adaptation. For example, English has borrowed Spanish words containing a rolled-r, which is not a consonant of English. In this case, the words are adapted to the native sound system by changing the rolled-r into the glide-like “r” of (American) English. This makes intuitive sense. That’s probably the most similar sound in English.

This is important, because it shows that English speakers are not insensitive to foreign language, despite Gustavo Arellano’s accusations that they “butcher” Spanish words. If English speakers were truly so lazy and tin-eared, then we’d expect rolled-r to get transformed into something utterly different, like /b/ or an /m/. I don’t think you have to be a trained linguist to see how that would be really weird. When foreign sounds are borrowed, the borrowers do some kind of unconscious computation to decide which native sound is most similar. It’s not a random choice, and it’s not committing abuse to the foreign language. It’s actually pretty amazing that large groups of people can, without explicit consultation, come to the same decision about how to translate a foreign sound into a native one. It just seems natural that we turn one r-sound into another, so we all do it.

The reverse happens too, of course. When English words with an “r” are borrowed into Spanish, they have to change because Spanish doesn’t have that sound. In fact, Spanish has two different r-sound (a tap and a trill) so there is more than one possible way to adapt English words to the Spanish sound system.

The Russian borrowing of hamburger as “gamberger”, which I mentioned earlier, is another example. Russian doesn’t have a /h/ sound, and /h/ is usually changed to /g/ in loanwords. Burmese, the national language of Myanmar, doesn’t have an /f/ sound, so when it borrows English words, all instances of /f/ are replaced by /p/. The word file is more like pai in Burmese, and phone is closer to po-u.

Another factor that can trigger a loanword adaptation is something known as “phonotactics”. This refers to language-specific restrictions on how sounds can be combined into syllables. In English, syllables can have up to 3 consonants in a row at the beginning (like string), but never more than that. Also, any word that starts with 3 consonants must start with /s/. We don’t have words like *mtrip or *gtrip but strip does exist. On the other hand, Sipakapense, a Mayan language spoken in Mexico, has words with up to 6 consonants in a row: xtqsb’jaj means ‘we are going to whack him’.

Languages that allow long strings of consonants are actually uncommon. Most languages have very strict rules. In Japanese, for example, a syllable can start with no more than one consonant. Think about words like karaoke, kimono, tofu, Pokemon, etc. They have no consecutive consonants. Japanese syllables can also end with up to one consonant, but it has to be nasal sound, as in Honda, manga, and bonsai.

This is why disk is borrowed into Japanese as disuku. It’s not possible to have the sequence /sk/ in a Japanese word, so speakers add in a /u/ sound to separate them into different syllables. There’s also /u/ added to the end of the word, because Japanese words cannot end in /k/ (due to the fact that it is not a nasal sound). This is not a case of mispronouncing English words. It’s something way more complicated and impressive than that. It’s not exactly trivial to mentally break down a foreign word, realize it contains sequences not in your native language, and then find a way of repairing that word so that it fits with the rest of the patterns in your language.

So in short, borrowed words change so that they fit into the patterns of the borrowing language. This gives a reasonable answer to the question of how borrowed words change. But it doesn’t quite explain why they change. Why do adaptations have to occur in the first place. Why don’t people borrow words exactly as they hear them?

One theory is that loanwords represent what people actually are hearing. Research has found that our native languages can shape the way that we perceive speech. Adult speakers tend to “filter” speech through the patterns of languages they already know, which can results in people literally not being able to notice sounds.

This is something that’s been studied in laboratory experiments. One interesting study looked at how French speakers perceive [h]. This is not a sound of French, although “h” is a silent letter in some words. French speakers famously have difficulty mastering it in English. I had a French-Canadian coach in high school who would say things like “I urt my harm hand went to the ospital”.

The study consisted of two experiments. In one, English and French speakers heard recordings of the words “hum” or “um”. In the other experiment, participants heard recordings of just an /f/ sound or a sequence of /hf/ (which had been spliced together by computer). The experiments were pretty simple, and basically aimed at finding out if English and French speakers could tell the difference between something that starts with /h/ and something that doesn’t.

A simple way to do this is to just ask people “did you hear an /h/?” after playing the words. However, that won’t get very interesting answers, because we already know that the French speakers have difficulty detecting /h/. Instead, researchers in this study used EEG caps, which are little caps fitted with electrodes that can measure some brain activity. We are not necessarily aware of everything that our brains perceive, so using EEG allows researchers to look a little more carefully at what’s actually going on with loanword perception.

The experiments were designed using something known as the oddball paradigm. The participants hear a single word repeated over and over again for a long time, like “hum hum hum hum…”. At unpredictable moments, there’s an “um” thrown in. This is the oddball. (Alternatively, they heard “f f f” repeated with a random “hf” tossed in). While this is going on, participants are usually instructed to do something else. Often, people watch a silent movie during the experiment. They don’t have to actively pay attention to the sounds they hear. (In fact, participating in these studies is extremely boring. During my master’s degree, I was frequently recruited for EEG studies by my colleagues because I keep very short hair, and the electrodes are easy to attach to my head. I fell asleep during one experiment and the guy was really annoyed with me.)

Using the EEG caps, researchers can see if the participants’ brains noticed when the oddball slipped in, based on changes in brain activity. When repetitions go on for a long time, our brains get bored, and the EEG measures constant activity. When new stimulus comes along, our brains get excited, and the EEG will record a sudden change in brain activity.

Unsurprisingly, the brains of English-speakers show a change every time the oddball appears, in both experiments. This is unsurprising because the presence or absence of /h/ in English is critical for distinguishing words like hum/um, heat/eat, hit/it, heave/eve, inerrant/inherent etc. If you speak English, then you have lots of practice and you are really sensitive to the difference between /h/ and no /h/.

The French brains reacted differently. When listening to the words “hum” and “um”, nothing happened. When the oddball word was played, there was no noticeable change in brain activity. It’s like they didn’t even realize that a different word was played. This part is not very surprising. It confirms what ESL teachers have long known.

Now here’s the weird bit: when listening to the non-words /f/ and /hf/, the French speakers performed exactly like English speakers. Their brains actually did notice when the /h/ was there and when it wasn’t. In other words, their perception changed when they thought they were listening specifically to language (the words “hum” and “um”) compared to just listening to noise (/f/ and /hf/).

This means that our language filter plays a powerful role in how we perceive speech, and therefore in how we deal with borrowed words. It is unreasonable to demand that French speakers borrow an /h/ when they are not even consciously able to hear that sound. Similarly, we shouldn’t expect English speakers to roll the “r” in words of Spanish origin.

That being said, we aren’t helpless creatures. It’s not impossible to remove this filter. People can become fluent in a second language, even as adults. People can also become comfortable with new pronunciations. If you were to move to Australia or South Africa, you could continue to use English, but you would need to get comfortable with some radically different pronunciation norms. Nonetheless, North Americans manage to do this all the time, generally without complaint. Personally, I think this because is psychologically easier when we consider the way of speaking to still be “English”.

What if the Spanish-American accent were to be considered just another dialect (ethnolect?) of English, namely one that contains a rolled-r? Then perhaps this could become more of a conversation about differences in dialect, rather a polarized “right” vs. “wrong” argument. Given how many Spanish speakers there are in the US, I would bet a Spanish-American dialect group outnumbers some well-recognized dialects like Boston English.

There’s nothing wrong with using foreign pronunciations in English, especially when bilingual people do it. There’s no need for anyone to get antsy that Spanish-speaking journalists are using Spanish pronunciations for words of Spanish origin. But neither should we agree that those are the only “correct” pronunciations in the context of an English broadcast. Loanword adaptation is a real thing, and it’s OK if you do it. The original pronunciation doesn’t always take precedence.

Most of all, we need to stop treating loanword adaptation as though it is doing violence to other languages. Stop accusing people of “mangling” or “butchering” a language. Adaptation of sounds is rooted in facts about human speech perception, not ignorance of other languages or an uncaring attitude. Instead, when the issue of foreign pronunciations comes up, we should use the opportunity to highlight the diversity and complexity of human languages. It’s amazing how many different ways there are to talk, and how many different sounds humans can produce and perceive. It’s even more amazing how freely we can trade words, despite our linguistic differences. Maybe we could celebrate this diversity instead of complaining about it.