Cà Phê, Xà Phòng, Ô Tô: The French Hiding Inside Vietnamese

Order a coffee in Hà Nội and you are speaking French without noticing. Cà phê is café. The bơ in your bánh mì is beurre, the phó mát or phô mai beside it is fromage, and the ga where you catch the train is gare. When your motorbike gets a flat, the mechanic checks the săm (chambre à air, the inner tube) and the lốp (enveloppe, the tyre), then sends you off to fill up with xăng (essence).

I wrote recently about Hán-Việt, the deep Chinese layer under formal Vietnamese. The French layer is much thinner and much younger, a few generations of colonial contact from the 1860s to 1954. But it sits in the most ordinary parts of life: food, clothes, machines and paperwork. It is also a small, clean laboratory for how a language swallows foreign words.

How big is it, really?

Smaller than people tend to think. The most careful recent count I found is a corpus built by Vera Scholvin and Judith Meinschaefer at the Freie Universität Berlin. They began with 1,038 Vietnamese words of French origin, then kept only the nouns still in current use, checked against native-speaker judgments, word frequency, web usage and a Vietnamese dictionary. 533 nouns survived.

For a sense of scale, the World Loanword Database (WOLD) lists a Vietnamese vocabulary compiled by the linguist Mark Alves: 1,477 words for basic meanings, about 27% of them borrowed. Most of those loans are Chinese. The French ones show up exactly where you would expect: soup (xúp), cheese, butter, wine (vang, from vin), shirt (sơ mi, from chemise), soap, car, bus (búyt), battery (pin, from pile), stamp (tem, from timbre), screwdriver (tua vít, from tournevis), film (phim) and coffee.

In other words, French gave Vietnamese things rather than ideas. Abstract and official vocabulary kept coming from Sino-Vietnamese roots, even for modern concepts: a bank is formally ngân hàng, although WOLD also records the French-derived băng (as in the older nhà băng).

Squeezing French into Vietnamese syllables

What makes these words interesting is the shape they had to take. Vietnamese is built from single syllables, each carrying one of six tones. Very few consonant clusters are allowed, and only a short list of sounds can end a syllable. French breaks every one of those rules, so each borrowed word had to be rebuilt. A few patterns stand out.

Clusters get cut, not padded. Faced with crème, Japanese inserts vowels to break up the cluster: kurīmu. Vietnamese mostly drops a sound instead: crème becomes kem, timbre becomes tem, fromage becomes phó mát, and frein (brake) becomes phanh. Scholvin and Meinschaefer found this is the general pattern in their corpus. Vietnamese prefers deletion over epenthesis (inserting a vowel), which they note is the opposite of what many other languages do. Where an extra syllable does appear, as in cravate → cà vạt, the result is still a string of legal Vietnamese syllables.

Long words get clipped, and the end survives. French stresses the last syllable of a word or phrase, and that is often the part Vietnamese keeps. Essence first became ét-xăng and was later clipped to xăng. Enveloppe became lốp, automobile became ô tô, and chambre became săm. If you know some French, you can often recover the original by asking what its stressed tail sounded like.

Every syllable gets a tone. French has no lexical tones, so speakers had to give each new syllable one. Linguists have studied how this happens (Sabine Huynh’s work and the Berlin corpus both focus on it), and the full picture is more complicated than any single rule. Still, one tendency is easy to hear in common words: an unstressed open first syllable often takes the low falling huyền tone. You can hear it in cà phê, cà rốt (carotte), cà vạt and xà phòng (savon). If I meet a new two-syllable word starting with cà- or xà-, I now half-expect a French original behind it.

Regions borrowed separately. The north and south did not always settle on the same form. Savon became xà phòng in the north and xà bông in the south, where Wiktionary notes it often means laundry detergent rather than bath soap. Cheese is phó mát or phô mai depending on who you ask.

A myth worth dropping

The most famous “French” Vietnamese word is not a French loan at all. Bánh mì is a native compound: bánh (a broad word for cakes, breads and dumplings) plus mì (wheat). The bread came from France; the name did not. French gave Vietnam the baguette and Vietnam named it in its own words, which is a neat summary of the whole layer.

What I take from this as a learner

If you speak French, use it as a decoder, not a dictionary. Reverse the rules above: put back the dropped r, imagine the full word behind a clipped syllable, ignore the tone. Tem, kem, ga, phim and pin fall straight out. Just don’t assume a French word exists in Vietnamese; most don’t.
Learn the tone as if the word were native. The most common mistake I hear from foreigners is pronouncing cà phê like café. To a Vietnamese ear, the tone is the word: ga (station) and gà (chicken) are different words, and a French cognate won’t protect you.
Know which register you’re in. French loans are everyday, spoken and practical. When a formal Sino-Vietnamese equivalent exists (ngân hàng rather than nhà băng), it is usually the safer choice in writing.
For Vietnamese learners of French, it works in reverse. Garage, antenne, cravate and carotte are free vocabulary. The work lies in rebuilding the clusters and final consonants that Vietnamese removed, so practise crème and timbre slowly, sound by sound.

Every language is a record of who it met. Hán-Việt records a thousand years of Chinese scholarship. The French layer records about ninety years of coffee, cheese, tyres and stamps, trimmed into syllables a Vietnamese speaker can pronounce without thinking.

How big is it, really?

Squeezing French into Vietnamese syllables

A myth worth dropping

What I take from this as a learner

Sources