SVSYSTEMS

Index / Language / Etymology

Etymology

How 1 spoken Latin became French, Catalan, Spanish, Galician, Portuguese, Italian, Sicilian and Romanian, and why English, which is not Romance at all, shares half their vocabulary.

1

The Latin nobody wrote down

The Romance languages do not descend from the Latin of Cicero and Virgil. They descend from what soldiers, traders, farmers and slaves said to each other: a Latin that was already simpler, slangier and changing while the classical language was being taught in schools. Linguists call it Vulgar Latin, from vulgus, the common people, which means popular rather than coarse.

We know it mostly by accident. Graffiti on the walls of Pompeii misspell words the way people said them. A schoolteacher's list from around the 3rd or 4th century, the Appendix Probi, corrects common errors: write auris, not oricla (ear), write oculus, not oclus (eye). The errors are exactly the forms that survived: oricla is French oreille, Spanish oreja, Romanian ureche. Check the source

You can see the gap in the words they chose. Classical Latin said equus for horse; the street said caballus, a nag, and every Romance language took the nag: cheval, caballo, cavallo, cal. Classical caput (head) lost out in French and Italian to testa, a clay pot, much as English says nut or noggin. Classical edere, to eat, lost to mandūcāre, to chew, a word from comedy.

equus → caballushorse → nag
caput → testahead → pot
auris → oriclaear → little ear
2

Latin spreads, at different times

Rome did not reach everywhere at once, and that matters. Latin arrived in Iberia in 218 BC, in southern Gaul in the 120s BC, in the rest of Gaul with Caesar in the 50s BC, and in Dacia, north of the Danube, only in 106 AD. Each province learned the Latin current when it was conquered, and each learned it from different people.

An old theory holds that this is why Iberian Spanish and Portuguese sometimes keep older Latin words that Gaul and Italy replaced: comer (from comedere) instead of the newer manger, hermoso (from formōsus) instead of beau (from bellus). Iberia was settled early and sat at the edge, so later fashions from Rome did not always reach it. Debated

Romanian sits at the opposite extreme: the last to arrive, the first to be cut off. Rome held Dacia for about 165 years and left in 271.

218 BCIberia
58 to 50 BCGaul, under Caesar
106 ADDacia

From Rome to now. Scroll sideways on a phone.

3

The empire breaks

From the 5th century the western empire comes apart. The last western emperor is deposed in 476. Schools close, roads decay, long-distance trade shrinks, and the Germanic kingdoms (Franks in Gaul, Visigoths in Iberia, Lombards in Italy) draw new borders. Spoken Latin keeps changing, but now each region changes on its own, and nothing pulls the regions back together.

For several centuries people still believed they were speaking Latin. Then, around 800, Charlemagne's scholars reformed the pronunciation of church Latin back towards its spelling, and suddenly ordinary people could not follow the sermon. In 813 the Council of Tours told priests to preach in the rustica romana lingua, the rustic Roman tongue. That is often taken as the moment the split became visible. Check the source

Eastern Latin had already gone its own way. It spent these centuries among Slavs, Greeks and Albanians instead of Germanic kings, which is why Romanian is the family member everyone finds a little strange.

476last western emperor deposed
813Council of Tours
5th to 8th c.the languages separate
4

1 word, 8 outcomes

Left alone, each region wore its words down in its own way. The changes are regular: once you know a sound law, it applies to hundreds of words at once, which is why linguists call them laws. They are not rules anyone followed; they are patterns that ran through a whole community's speech over generations.

The best one to learn first is what happened to -ct-. Latin noctem, night, became nuit, nit, noche, noite (in both Galician and Portuguese), notte, notti and noapte. Swap in factum, lactem or octō and the same pattern returns every time: Portuguese -it-, Spanish -ch-, Italian -tt-, Romanian -pt-. The map below shows where each outcome is spoken today.

Other laws cut the family along different lines. Whether a p between vowels became b (Spanish lobo) or stayed (Italian lupo) splits east from west along a line across northern Italy. Whether an n between vowels vanished (Portuguese lua) sets Portuguese apart from everyone. The sound-law machine runs each word through its laws step by step.

noctemnuit · nit · noche · noite · notte · noapte
lactemlait · llet · leche · leite · latte · lapte
5

The grammar rebuilds itself

Sound change did more than reshape words. It wore away the endings that Latin grammar depended on. Latin marked who did what with case endings (lupus the wolf as subject, lupum as object); once final -m and -s weakened, the endings merged, and the languages rebuilt the same jobs with separate words.

Word order fixed who did what. Prepositions took over the cases: de for of, ad for to. The pointing word ille, that, became the article: le, el, il, o, and in Romanian, uniquely, an ending (lupul, the wolf). The Latin future disappeared, and a phrase took its place: cantāre habeō, I have to sing, fused into chanterai, cantaré, canterò.

The result is a family that shares Latin's vocabulary but built its grammar independently, along broadly parallel lines. See the new grammar family on the map.

ille lupusle loup · el lobo · lupul
cantāre habeōchanterai · cantaré · canterò
6

Words from outside

Every language took in words from whoever ruled or traded nearby, and the sources differ sharply. French took hundreds of Frankish words, including bleu, jardin and guerre, and possibly some of its pronunciation. Spanish and Portuguese took several hundred from Arabic after 711, still marked by the Arabic article: algodón, azúcar, arroz, alcalde.

Romanian took a Slavic layer so deep it reaches da (yes), a iubi (to love) and prieten (friend). Italian, at the centre, took comparatively little from anyone and changed least.

Later, Spanish and Portuguese became the gateway for words from the Americas, Africa and Asia: tomate, chocolate, canoa, and the Portuguese chá that made tea a different word in Lisbon from everywhere else in western Europe.

Frankishguerre, blanc, garder
Arabicazúcar, aceite, alcalde
Slavicda, prieten, a iubi
7

Latin comes back through books

Latin never stopped being written. It stayed the language of the church, the law, the universities and science for 1,000 years after it stopped being spoken. From the 12th century, scholars, lawyers and clergy began borrowing words straight from the page into the vernaculars, whenever they needed a word for something abstract, technical or formal.

These words arrived too late for the sound laws, so they look almost exactly like Latin and almost exactly like each other: nation, nación, nação, nazione, națiune. This is the single biggest reason the languages look alike today, and why an English speaker can read a Spanish newspaper headline more easily than a Spanish novel.

It also produced doublets: the same Latin word twice in 1 language, once worn down by speech and once fresh from the book. Spanish has llave (a key) and clave (a code), both from clāvem; French has hôtel and hôpital, both from hospitālem.

clāvemllave (spoken) · clave (book)
hospitālemhôtel (spoken) · hôpital (book)
8

The languages are written down and fixed

The vernaculars appear in writing slowly: French in the Strasbourg Oaths of 842, Italian in the 960s, Portuguese by 1214, Catalan in the late 12th century, Romanian not until a letter of 1521. For centuries the written languages were patchworks of local habits.

Printing, grammars and academies then fixed the standards. Antonio de Nebrija published the first grammar of Spanish in 1492, the year Granada fell and Columbus sailed; he told Queen Isabella that language was the companion of empire. The Crusca in Florence followed in 1583, the Académie française in 1635, the Real Academia in 1713.

Standardisation chose 1 dialect and made it the language. Standard Italian is Florentine, thanks to Dante, Petrarch and Boccaccio. Standard Spanish is Castilian. Standard French is the speech of Paris and the court. The other dialects did not disappear; they became what people now call accents, or, in Italy, dialects that are languages in their own right.

842Strasbourg Oaths, French
1492Nebrija's grammar, Spanish
1521Letter of Neacșu, Romanian
9

The ones that disappeared

Not every branch survived. Dalmatian, spoken on the Croatian coast, faded as Venetian and Croatian took over; its last known speaker, Tuone Udaina, died in an explosion in 1898, and a scholar had recorded his speech just before. Mozarabic, the Romance of Muslim Iberia, vanished as Castilian moved south, leaving fragments in Arabic and Hebrew script at the ends of poems. Check the source

Romance was also spoken in North Africa until the Middle Ages, along the Moselle and in Pannonia, and left only traces. Others are alive but small: Sardinian, often called the most conservative of all, still says kentu for a hundred, with Latin's hard k; Romansh in Switzerland; Occitan in southern France; Galician, close kin of Portuguese, and Sicilian, both now included on this page.

Most of the 8 on this page have states, armies or academies behind them; Galician and Sicilian are the exceptions, regional languages with millions of speakers. A famous line, usually credited to the linguist Max Weinreich, says a language is a dialect with an army and a navy.

1898last speaker of Dalmatian
kentuSardinian, still a hard k
Galiciansister of Portuguese
10

English: Germanic core, Romance coat

English is not a Romance language, but more than half its dictionary is Romance in origin: first from Norman French after 1066, then straight from Latin in the Renaissance, then from French again. Its everyday core stayed Germanic: be, have, do, man, house, water.

The result is a lever no Romance language has: 2 or 3 words for many things, 1 plain and 1 formal. Ask, enquire, interrogate. Kingly, royal, regal. The cow in the field and the beef on the plate, because the English herded the animals and the Normans ate them.

English also froze some French forms that French itself later changed. English forest, feast and hostel keep an s that French lost, leaving only a circumflex: forêt, fête, hôtel.

ask / enquire / interrogateGermanic / French / Latin
forest / forêtthe s English kept
11

What is still argued

Not everything here is settled. Why Spanish turned f into a silent h (hijo, hacer) is disputed: a Basque substrate, or a change from inside the language. Which text is the first in each language depends on where you draw the line between late Latin and early Romance. How Romanian survived north of the Danube after Rome left is still debated, and the debate has been political as well as scholarly.

The exact count of Arabic words in Spanish, of Frankish words in French and of Slavic words in Romanian all vary by source, depending on whether you count place names, rare words and derivatives.

This page marks those points as disputed rather than picking an answer, and tags anything not yet checked against a source with a green label.

f- → h-cause disputed
first textsconventional dates
Daciacontinuity debated