Spanish, Italian and Portuguese word overlap

Ask how close Spanish and Portuguese are and the answer usually arrives as a lexical similarity percentage from the linguistics literature. We cannot reproduce that figure from anything we own, so we do not print it. What we can do is open our own packs and count. Our A1 list carries the same 617 English headwords into Spanish, Italian, Portuguese and French, and for 303 of them the Spanish and Portuguese words sit within a small spelling distance of each other. That is 49.1 percent. Spanish and Italian manage 178. French against Spanish manages 79. This measures spelling similarity within our selected words. It does not measure how much of another language a learner will understand.

What we counted and what we did not

Two different things get called overlap here, and only one of them says anything about the languages.

The first is curation overlap: how many English headwords two of our lists share at the same CEFR level. At A1 that is 617 out of 617 for every pair among Spanish, Italian, Portuguese and French, and at A2 it is 625 out of 625. We built it that way. One graded headword list was carried into all six languages we ship, so the Spanish A1 pack and the Portuguese A1 pack teach the same concepts in the same order. That figure describes our curation, and we print it below so nobody reads it as a finding about Romance vocabulary.

The second is spelling proximity. It describes the word forms we selected; it does not establish that they share an origin. For every shared headword we take the word each language teaches, lower-case it, strip the accents and the punctuation, then measure the Levenshtein edit distance between the two forms and divide by the length of the longer one. At or below 0.25 counts as close, which is roughly one character in four allowed to differ. Spanish verde against Italian verde scores 0.0. A pair like vida and vita scores exactly 0.25 and squeaks in. Put tienda beside negozio and you get 1.000, the ceiling.

The source is the packs the app ships: every pack id in our registry that has a manifest on disk, English to Spanish, Italian, Portuguese and French, at A1 and A2. Headwords come from the New General Service List, which Browne, Culligan and Phillips publish under CC BY-SA 4.0. Grading, translations and the arithmetic are ours; the methodology page covers the grading.

The same 617 concepts in all four lists

Every cell below counts the English headwords two lists share at that level. The percentage is that count against the shorter of the two.

A table of identical numbers is worth printing once, because it is the disclosure that makes the rest of the page readable. Nothing in it was discovered. Had we graded each language separately these cells would differ, and the differences would still be a fact about the curators.

The entry counts behind those headwords are less tidy. Spanish ships 625 A1 entries for 617 headwords, because 8 headwords carry two entries each: help is both Ayuda and ayudar, work is both trabajo and trabajar. French ships 624. Italian shows 1,248 entries and Portuguese 1,250, because those two ship A1 twice in this direction, once as a single starter pack and once split across twelve themed packs. The words are the same words. Counting headwords takes both effects out.

How close the words actually look

Here are the spelling-distance results. For the 617 headwords all four lists carry at A1 and the 625 at A2, here is how often two languages' words fall within a quarter of their length of each other, and how often they match exactly once accents come off.

Spanish and Portuguese sit apart from everything else on the page. About half the A1 headwords in this comparison have Spanish and Portuguese forms within our chosen spelling-distance threshold, and 179 of the 617 A1 words match once accents come off, which is how this page counts a same spelling: about a ninth of those 179 still differ on the page, día against dia and mes against mês, and the marks the comparison drops are the ones carrying vowel quality and nasality. The rest are the same word on the page: verde, grande, cliente, falso, euro. Spanish and Italian come in at 28.8 percent at A1 and 31.2 at A2. Italian and Portuguese track just below that pair at both levels.

The mean distances say it from the other side. Across the A1 list, Spanish to Portuguese averages 0.352, Spanish to Italian 0.489, Italian to Portuguese 0.523. Every word that shares nothing is inside those averages, which is why none looks small.

Split A1 by our twelve themes and the overlap moves around. Numbers & Time is 53 of 87 words close and Everyday Verbs 62 of 104. At the bottom sit People & Family, 10 of 35, and Greetings & Essentials, 15 of 51. The vocabulary you need in the first hour is the part these two languages agree on least.

The ten closest and the ten furthest at A1

Ranked on the mean of the three pairwise distances and quoted from the manifests exactly as the packs carry them. Ties break on the English headword, so every run returns the same twenty.

A month, a colour, a size, two job words, the currency, and a run of Latinate adjectives. Two of the rows are the same three words, because our A1 list gives both cool and fresh to fresco in all three languages. Similar spelling can help recognition, but these words may still need practice for pronunciation, grammar or use in a sentence.

The other end is where a beginner lives: shop, street, face, grandmother, class. Everyday concrete nouns, and the translations selected in our packs look quite different. The headword their is the only one at A1 whose three words hit the ceiling in all three pairs, because su, loro and deles put no letter in the same place. Portuguese envelope is closer to the English headword than to Spanish sobre.

French sits apart from the other three

French carries the same 617 headwords at A1 and matches the other three about a third as often as Spanish matches Portuguese: 79 close against Spanish, 89 against Italian, 89 against Portuguese. It is close to equidistant from all three. Nothing in these lists shows a French and Italian axis to set against the Spanish and Portuguese one.

At A2 French gains ground: 112 against Spanish, 108 against Italian and 97 against Portuguese, so 17.9, 17.3 and 15.5 percent against 12.8, 14.4 and 14.4 one level down. The exact matches move the other way. French shares an identical spelling with Spanish 26 times at A1 and 15 times at A2, and with Italian 20 times at A1 and 9 times at A2. More French words drift into the near-miss band as the vocabulary gets more abstract, and fewer of them land on the nose.

Silent letters are ordinary in written French, and this measure reads letters. Whether a French word sounds nearer to Italian than it looks is a question these numbers cannot reach.

If you are picking a second Romance language

The numbers offer a limited clue if you are choosing another language.

Coming from Spanish, Portuguese is the shortest hop on the page: 612 of the 1,242 headwords in our A1 and A2 lists sit one small spelling step from their Spanish forms, and 344 of those are spelled the same once accents come off. Italian gives you 373 of 1,242. French gives you 191. These are counts of similar spellings. Whether a reader recognises a word on sight is a different question, and this page does not test it. Pronunciation and listening are outside the measure.

What the numbers refuse to say: anything about listening, gender, verb endings, or how fast you will be understood in a shop. A word you can recognise is a long way from a word you can produce. Our own A1 lists put tienda, negozio and loja against one English headword, and Spanish does nothing for the other two.

Whichever pair you pick, once charger playback is set up and background settings allow it, the app plays one word when the phone goes on charge, weighted towards the entries you have gone longest without hearing. One word per charge explains the rotation.

Where the measure breaks down

Edit distance counts characters. It cannot tell a cognate from a coincidence, and it scores a false friend as a perfect match.

Our own packs prove that. Across A1 and A2 there are 23 cases, over 18 distinct forms, where two of these three languages teach the same spelling for different English headwords. Reading them one by one, a third are one idea wearing two English labels (barco as ship and as boat) rather than a trap, so the number of real traps is nearer fifteen. Spanish largo is long, while Italian and Portuguese use largo for wide. A primo is a cousin in Spanish and Portuguese, and first in Italian. Italian palestra is a gym; the Portuguese palestra is a lecture. Portuguese teaches aula at A1 for the class, and Spanish teaches the same spelling at A2 for the room the class happens in. A Spanish escritorio is a desk; a Portuguese escritório is an office. One accent separates Spanish más from Portuguese mas, and it separates more from but. All of them score 0.0 on this page's measure.

That detector over-reports, and we would rather say so than quietly round it down. Some of the 23 are one idea under two English labels: Spanish barco is graded as ship and Portuguese barco as boat, which is a grading choice showing up as a trap. We have not manually classified the full set, so we cannot give a verified count of false friends from these matches.

Four more limits. The threshold is a choice, and it is unkind at the edges: Spanish restaurante against Italian ristorante scores 0.273 and falls outside, though no reader would miss the connection. Move the line off 0.25 and every percentage here moves with it. Our A1 and A2 lists are 617 and 625 headwords out of languages that hold far more, and the New General Service List picks the common frequency core, where cognates may be over-represented or under-represented; we do not know which way that pushes these figures. Portuguese here is the Brazilian voice throughout, so read that column as Brazilian Portuguese. And a headword carrying two entries keeps the target with the lowest entry id, which affects 8 Spanish A1 words, 6 Italian, 7 Portuguese and 6 French, and none at A2.

Where the lists live

Every number above came out of packs you can open in the app. Corpus statistics holds the whole 33,233-entry corpus by language and level, and the per-language pages carry the level counts, the theme split and the entries-against-distinct disclosure: Spanish, Italian and Portuguese.

To run the comparison yourself, download A1 in two of the four languages and read the same theme in both. Our Spanish and Portuguese packs show higher spelling proximity in Numbers & Time and Everyday Verbs, and lower proximity in People & Family and Greetings & Essentials. Comparing those themes gives you concrete examples of both the similarities and the gaps.

Common questions

Is Italian easier after Spanish?

In our packs, Spanish and Italian forms fall within our spelling-distance threshold for 178 of 617 A1 headwords (28.8 percent) and 195 of 625 at A2. Spanish and Portuguese reach 303 of 617 at A1. Similar spelling may help recognition, but these figures do not measure how easy a language is to learn.

How similar are Spanish and Portuguese?

We do not publish a lexical similarity figure, because we cannot reproduce one. In our own A1 packs, 303 of the 617 shared English headwords have Spanish and Portuguese words within a small spelling distance of each other, and 179 of them are the identical string once accents come off.

Which Romance language is closest to Spanish?

Among Portuguese, Italian and French in our selected packs, Portuguese has the highest spelling proximity to Spanish: 49.1 percent at A1, against 28.8 percent for Italian and 12.8 percent for French. This is not a comparison of every Romance language or of their grammar and pronunciation.

Does knowing French help with Spanish?

Less than the other pairings, on spelling. Our A1 lists put 79 of 617 headwords close, which is 12.8 percent, rising to 17.9 percent at A2. French is close to equidistant from Spanish, Italian and Portuguese in these lists, and it reaches 14.4 percent against both Italian and Portuguese at A1.