What changes in the Portuguese spoken by the inhabitants of São Paulo state have been ascribed to Italian immigration? Just as the transfer of the Portuguese Court to Brazil to the escape Napoleonic forces determined permanent changes in the Portuguese spoken by the inhabitants of Rio de Janeiro, Italian immigration to Brasil is also said to have produced changes in the Portuguese spoken by Paulistas. One single example I have read is the way the word "carroça" is pronounced: while it's pronounced with a double "r" sound in most of Brazil, a large number of Paulistas pronounce it as if it had only one "r". How exactly has the Italian presence influenced and changed pronunciation and vocabulary in São Paulo state?


Da mesma forma que a vinda da Corte de Lisboa para o Brasil em 1808 determinou mudanças permanentes no modo de falar dos cariocas (e, notem bem, não foram 200 e nem 500 nobres portugueses que vieram. O número total é calculado em uma leva incial de 15.000 almas em uma época em que a população da cidade não passava de 50.000) a imigração italiana é dita ser responsável por diversos aspectos do modo de falar dos paulistas atualmente. Um único exemplo que vi citado é a pronuncia da palavra "carroça" que é pronunciada como tendo apenas um único "r", principalmente em comunidades italianas do interior. Quais exatamentes foram as mudanças, tanto em termos de pronúncia quanto de vocabulário, que a imigração italiana trouxe para o estado de São Paulo?

EDIT - I didn't realize the subject was so extensive.


Throughout its history, Brazil has enjoyed immigrants from the Italian peninsula, dating back to Genoese sailors and merchants during the first half of the sixteenth century. Large waves emigrated from Italy between 1870 and 1920, so many that in Brazil’s 1940 census, more children had been born up to that point to foreign-born fathers of Italian origin (1.26M) than to fathers of Iberian origin (1.08M), even when the figures from Portugal (0.74M) and Spain (0.34M) are combined.

Some 70% of these immigrants settled in São Paulo State, and many children were born to households where one or another Italian dialect was the main language, or which were at least bilingual. These dialects came to be heard everywhere in the streets of São Paulo, and this produced several observable changes in the Portuguese spoken there.

A similar phenomenon can be seen in the Spanish spoken in Argentina, and for exactly the same reason: massive numbers of Italian immigrants. In contrast, even though Brazil and Argentina both had many German immigrants, those particular immigrants were responsible for very little linguistic change. Italian is something of a cousin language, and the Italian culture is not so far removed from that of the existing Iberian colonists compared to the language and culture of the German immigrants. It was therefore much easier for Italian linguistic influences to enter the dominant spoken languages.

Speaking broadly, these influences can be grouped according to phonology, lexicon, and grammar, where grammar comprises syntax and morphology.


The most extensive treatment of this subject which I have come across is the 2010 master's thesis Para um estudo das influências fonológicas do italiano no português falado na cidade de São Paulo by Marcílio Melo Vieira. This study outlines the following influences:

The pronunciation of /r/ at the end of a syllable.

The exact realization of the /r/ phoneme at the end of the syllable does not matter much, because there exists no minimal pair contrasting the two types of rhotic phoneme in that position. Italian has only an apical rhotic [ɾ] for carta [ˈkaɾta], not the velar fricative [ʁ] heard in [ˈkaʁta]. The southern region is more likely to have a simple apical rhotic here compared with the velar fricative heard in the north.

The pronunciations of /t/ and /d/.

When /t/ or /d/ is followed by /i/, these are usually palatalized into affricates in Brazilian Portuguese, so /t/ > [ʧ] and /d/ > [ʤ] in those positions. Italian does not do that, however, and so in São Paulo this palatalization does not always occur, making the pronunciation of those sequences closer to Iberian Portuguese.

São Paulo also has a greater occurrence of the dental variants [t̪] and [d̪] as heard in both the Italian and Iberian peninsulas than the alveolar pronunciations of other Brazilian regions. If that distinction doesn’t make sense, just think of the difference between how Italian and English normally pronounce their /t/ and /d/: the first is dental and the second alveolar.

Interference of the Italian vocalic system upon the Portuguese one

Italian does not have nasalized vowels. Where Rio de Janeiro pronounces homen and fome as [ˈõmẽj] and [ˈfõmɪ], São Paulo has [ˈɔmẽj] and [ˈfɔmɪ].

Post-tonic vowel reduction

Italian does not have the reduced vowels that Portuguese has in unstressed vowels at the ends of words like gota, fala where /a/ > [ə], mato, como where /o/ > [ʊ], and jure, come where /e/ > [ɪ] in those words’ most common Brazilian realization.

For this reason, those vowels often retain their non-reduced values even in that position when spoken in São Paulo. Where Marta and Roberto are [ˈmaʀ:tə] and [ʀɔˈbɛʀ:tʊ] in Rio, in paulistano mouths they can be [ˈmaɾta] or [ˈmaɹta] and [ʀoˈbɛɾto] or [ɾoˈbɛɾto]. This absence of customary reduction makes those words sound a bit more like they were Italian or Spanish.

Hiatus versus diphthongs in combinations with semi-vocalic glides

When sounds like [i] and [u] come into contact with other vowels, they can combine with the other vowel to form a diphthong of a single syllable where they become semi-vocalic glides [j] and [w]. However, they can also remain in hiatus occurring in two separate syllables. Where for italiano, the typical Brazilian pronunciation has a diphthong and four syllables [i ta ˈlja nʊ] but the pronunciation influenced by Italian has a hiatus and five syllables [i ta li ˈa nʊ].

Retention of final /l/

Italian does not change a final /l/ into a /u/ or /w/ as so often occurs in Brazilian Portuguese, and so in São Paulo words like Brasil are often heard with an actual [l] at their end.


Like all European languages, Portuguese has many words of Italian origin, especially in areas such as art, music, architecture, cooking, and military terms. It is more difficult to find words or expressions of Italian origin that mainly occur in Brazilian Portuguese rather than those in common use amongst all Lusophones everywhere.

I searched for substitutions like Italian nono for avô as occurs in Argentina where it can substitute for abuelo, but could not find words of Italian origin which I could positively identify as being specifically Paulistano in usage.

I’ll therefore leave this section open for now in the hope that area residents with direct knowledge of the spoken language might offer some suggestions.


Singular used for plural nouns

One area consistently pointed out as being a possible Italian origin is the frequently heard use of singular nouns as though they were plural. Italian plural nouns do not have a final -s as Portuguese nouns do, and so it has been suggested by many sources that this habit may derive from there.

Because articles and verbs that accompany those nouns still have their normal plural inflections, one can still make sense of what is being said. Nonetheless this is not well looked upon, and so you seldom find it in printed material of more permanence than casual text messages.

Enclitic pronouns instead of proclitic ones

Some sources also suggest that the Brazilian habit of placing clitic pronouns before all finite verbs (for example, proclitic se tornou instead of enclitic tornou-se) may be explained by how Italian does these things.

However, Italian is not alone in this, because French and Spanish and various other Romance tongues also use proclitic pronouns in these situations, and so I am unconvinced that Italian is responsible for it. A thousand years ago in Iberia all Romance languages worked the way modern Portuguese still does in Portugal, but Catalan and Castilian have evolved away from that during the past millennium.

I was unable to discover evidence that this change occurred in Brazil during the period of greatest Italian migration. It may simply be common trait of Western Romance. While absence of evidence should not be construed as evidence of absence, nonetheless it makes me suspicious.

Gerunds for progressive constructions

The same can be said for the Brazilian style of forming present progressives using the gerund, so estou falando instead of estou a falar as heard in Portugal. One reference cited below nominates these progressives as a grammatical structure originating in Italian and therefore due to the Italian immigrants in Brazil.

While it is true that Italian uses the gerund there, again so too does Spanish, a language spoken throughout South America outside of Brazil. Therefore to attribute this grammatical change of the Brazilian construction to Italian influence would require evidence that it took place during periods of high immigration specifically in those areas of the country where Italians settled.

I was unable to discover any such evidence, and so for now I must consider such hypotheses to be suspect.

As pointed out by Jacinto, the adoption of the estou a falar style of forming progressives occurred in Portugal during the 19th century, but not in Brazil. So this theory does not hold water.


  1. Wikipedia: Imigração italiana no Brasil, Idioma
  2. Para um estudo das influências fonológicas do italiano no português falado na cidade de São Paulo
  3. Influências da Língua Italiana na Língua Portuguesa do Brasil
| improve this answer | |
  • 3
    I think, and most of you will agree, that excellence should be called out when we see it. I doff my cap to you tchrist. Magna cum laude – Centaurus Dec 5 '15 at 19:32
  • 2
    Excellent work. As to estou falando versus estou a falar there was no change in Brazil. It was Portgual that changed. Estou falando was the standard form in Portugal up to the 19th century. It is still used by some people in the south. There's this question about this topic – Jacinto Dec 6 '15 at 9:12
  • Perhaps the expression à paisana is an example of lexical influence? – bfavaretto Dec 11 '15 at 0:54
  • @bfavaretto Bernardo Guimarães (1825-84) uses à paisana in O Garimpeiro, 1872. Perhaps a tad too early for Italian influence. – Jacinto Dec 16 '15 at 22:41
  • Is "fome" pronounced /fɔmɪ/ in São Paulo city? I never lived there, so I am no expert, but I have always heard them pronounce /fow/ or /fo/ as the first syllable. On the other hand, I am also not sure that cariocas pronounce /fõmɪ/ either. – Luís Henrique Jan 30 '18 at 10:52

A discussion of this subject should map the variants spoken in São Paulo state versus the density of Italian immigration. Do the areas where Italian immigration was heavier speak differently from the areas were it was lighter?

I haven't seen a study that does this. So, with no systematical data at hand, I would contrast the Portuguese spoken in São Paulo city to that spoken in Campinas (the state's second city by population). Both cities received huge numbers of Italian immigrants. But Campinas' Portuguese is similar to that of other cities in the hinterland (for instance, in the southwestern cities, where Italian immigration was smaller, and indirect, and where Japanese immigration was also important). São Paulo's Portuguese, in contrast, is remarkably different from that of the hinterland. And so I would treat the claim of Italian influences in the Portuguese spoken in São Paulo with some caution. Below, I will underline the reasons I take the influences outlined by Marcílio Melo Vieira's thesis with some salt.

The pronunciation of postvocalic "r"

Brazilian Portuguese can be divided roughly into two variants, concerning the pronunciation of postvocalic "r":

A. In the litoral, from somewhere in between Santos and Rio de Janeiro to the North, the "r" in "carta" is pronounced similar to "rr" as in "carro".

B. To the South, and then again to the North through the hinterland, as far North as the Amazon, it is pronounced similar to "r" as in "caro".

Variant A is spoken in Rio de Janeiro; in São Paulo - all of the state, not just the city - B is widely predominant.

But then those variants can be subdivided, regarding what exactly is an "r" or an "rr". In Rio de Janeiro, "carta" is pronounced [ˈkaʁta], as "carro" is pronounced ['kaʁu]]; but in other places we can hear [ˈkaxta] or [ˈkahta], and I fear a few other pronunciations. The other variant, within which São Paulo is placed, is less varied: it is either [ˈkaɾta], as in São Paulo city - and then in Paraná, Santa Catarina, Rio Grande do Sul, the Center-West, and the Amazon - or it is a retroflex "r" like in English: ['kaɹta], as in most of São Paulo hinterland, and parts of southern Minas Gerais and Goiás.

So, it seems that the specific pronunciation of post-vocalic "r" must be explained within the context of the whole Southern, Center-Western, and Northern regions, not exclusively of the city of São Paulo. And then we have to take into account not only Italian, but also Spanish and Tupi-Guarani, and Portuguese as it was at the time of expanding Portuguese colonial presence in Brazil. And, as far as I know, the uvular pronunciation is an innovation from the 19th century in Portugal. So the main hypothesis would seem to be, variant B, which is spoken in São Paulo state, is more conservative regarding the pronunciation of post-vocalic "r", not that such pronunciation is due to a foreign influence.

The pronunciations of /t/ and /d/

While the palatalisation of /t/ and /d/ when followed by /i/ is less common in São Paulo than in Rio de Janeiro, it is hardly an exception; indeed, it seems to be the most widespread pronunciation. The pronunciation of /ti/ and /di/ is very difficult for Brazilians(1); the tongue slips "naturally" from /t/ to /i/, due to the proximity of the point of articulation, resulting in [ʧi]. The other option is to change the articulation of the /t/ into a dental, thus setting the articulation point of consonant and vowel farther apart, which is characteristic of some Northeastearn pronunciations. This is present in São Paulo, too, and is popularly attributed to "Italian influence" - but given the equally huge, and more recent, internal migration from the Northeast to São Paulo, it would be necessary to discuss the probability that the "influence" in the case is Northeastern, not, or not only, Italian. Moreover, such dental pronunciation is common in Portugal, too, though I am not sure when it superseded the previous [ti] (or [te], as in the case of final syllables), and to what extent, and in which regions, it precluded the palatalisation. So, again, I wouldn't exclude the possibility of mere conservatism regarding those pronunciations.

(1) What is discussed about /t/ in the following applies to /d/ the same.

Interference of the Italian vocalic system

Regarding the pronunciation of stressed /o/ before /m/ and /n/ syllable onsets, I think the idea that there is nasalisation in the rest of Brazil, but not in São Paulo is mistaken. The pronounce of /o/ in such positions is non-nasal in every Brazilian variant that I am aware of - and I fear the same is true of European variants either. The vowel that notably changes its pronunciation before /m/ and /n/ syllable onsets, when in a stressed position, is /a/, and even there it is hardly a nasalisation; rather the pronunciation is [ɐ] instead of [a]. And this is true of São Paulo, both city and state, no less than of Rio de Janeiro or the rest of Brazil; it is [ba.nɐ.na] in both São Paulo and Rio, not [ba.na.na] in São Paulo and [ba.'nã;na] in Rio.

What could be discussed here is the denasalisation of unstressed nasal vowels, such as /'ɔr.gu/ instead of /'ɔr.gɐ̃w/ for "órgão"; but I would say that this is rather an innovation of Portuguese Brazilian, more or less as a whole, than a phenomenon focused in São Paulo.

Conversely, I do not hear paulistas or paulistanos failing to reduce final vowels, unless for emphasis (/'bɛ.tu/ if you are mentioning Beto, but /'bɛ.to/ if you are calling him). As far as I know, such pronunciations are characteristic of rural Rio Grande do Sul, western Santa Catarina, and Paraná, and they seem either conservative or due to proximity to the Argentinian/Uruguayan borders, rather than influenced by Italian immigrants.

Hiatus versus diphthongs

A better argument seems to be the hiatus in words like "italiano" ([i.ta.li.ˈa.nu] instead of [i.ta ˈlja nu]. This, at least, is characteristic of São Paulo, not of the whole Southern region, and seems to be innovative, rather than conservative.

Retention of final /l/

This is an obviously conservative trait, and I am not sure that it is present in São Paulo at all. It is more characteristic of Rio Grande do Sul, and even there it seems to be quickly fading; I still say /a.'zuł/, and even /a.zu.li.brã.cu/, but to my daughter the difference is inaudible. An "actual l" such as in /mal/ instead of /maw/ or /mał/ is something that cannot be heard in Brazil, unless it is in some very small and isolated area (or, of course, if the following word starts with a vowel).


There are surprisingly few words that passed from Italian (or Venetian, Sardininian, Calabrian, Sicilian, etc.) through oral transmission, as opposed to literary imports. They are not always easy to distinguish, but if they have a "c" and are pronounced with an /s/, they have not been brought into Brazil by Italian immigrants. "Cicerone" is pronounced /si.se.'ro.ni/, even in São Paulo, so Brazilians read it and started pronouncing it according to Portuguese orthographic rules. "Ciao", conversely, is /ʧjaw/, and so it was either brought here by Italian immigrants (or imitated from Italian movies).

Most words that are actually due to Italian influence (via immigrants, not via Italian literature or high culture) are gastronomical terms. Pizza, brachola, risoto, ravióli, talharim, espaguete , etc. (And even here, we say /ʁi.'zo.tu/, not /ri.'zo.to/, so it is the Portuguese phonology that actually predominates over the Italian, not the other way round.) And then those words are all common throughout Brazil; none is specific of São Paulo.

Grammar Singular used for plural nouns

This is widespread in all Brazil. See for instance http://www.avozdapoesia.com.br/obras_ler.php?obra_id=4432 for a poem in popular Portuguese, full of instances of that ("Venho dos brêdo", "vivo naquelas mata", "os cumpanhêro", etc). And it is Northeastern, not paulista, and I would dare say absolutely virgin of any Italian influence.

Enclitic pronouns instead of proclitic ones

This seems to be generalised in Brasil. It is not like people say "me dá" in São Paulo but "dá-me" in Recife. And so it seems to be wholly autochtonous - possibly related to prosodic issues ("dá-me" is more difficult to pronounce if we don't delete the final "e" as the Portuguese do; conversely, "me dá" is less natural when pronounced /m'da/.

Gerunds for progressive constructions

As Jacinto says, this is a conservative trait.

If we look to a massive case of actual influence of a language upon another - for instance, the undeniable influence of Norman French over English, in the transition from Old English to Middle English - we will see that such influence is predominantly lexical. And enormous amount of French words were borrowed into English, from "beef" to "lieutenant". But the phonological impact seems to have been much more limited; the most impactful change in English phonology, the Great Vowel Shift, seems unrelated to French. And if there were grammatical impacts, they did not go in the way of substituting French grammar for Old English grammar; the change from Old English to Middle English was an enormous simplification of grammar, with the demise of cases, most of gender, most of verbal tenses, etc. The complex grammar of Old English was abandoned, but the complications of Old Norman French didn't replace it, and were instead lost as well.

And so, though this is only one example, I would dare say that the influences of any language over another one are mainly lexical, only marginally phonetic, and, in terms of grammar, if any, mostly destructive, not constructive.

Do we have other examples? The influence of Arabic on Iberian Romance, or Tupi-Guarani on Brazilian Portuguese, of Germanic languages over Vulgar Latin, perhaps? If so, they seem to point in the same direction. Portuguese phonology seems to remain basically Romanic, evolving of itself, not of out of contact to other languages. It gained phonemes, but not from Arabic, Sueve, Visigoth, or Tupi. Or, more modernly, from French or English. It gained words, lots of them, from all those languages, and Quimbundo, Nahuatl, Japanese, or Tagalog. But it seems to have imposed its phonology over all those words, rather than having its phonology altered to adapt to such imports. And it doesn't seem to have gained syntactical features from any such contacts.

| improve this answer | |
  • Isn't this a thesis ? – Centaurus Jan 30 '18 at 15:29
  • 1
    Good job, Luis. I have one doubt, though. I'm not sure the following sentence conveys what you meant it to "While the palatalisation of /t/ and /d/ when followed by /i/ is less common in São Paulo than in Rio de Janeiro, it is hardly an exception; indeed, it seems to be the most widespread pronounciation." less common than in Rio(?) But is widespread(?) – Centaurus Jan 30 '18 at 15:33
  • @Centaurus - yup. In Rio it is always palatalised. In São Paulo, I think, most of the time, but not really 100% – Luís Henrique Jan 30 '18 at 17:51
  • 1
    Excellent job, very instructive. As to the enclitic vs. proclitic thing, it appears to have varied quite a lot over the centuries. Some writers from earlier times in Portugal resemble more modern Brazilian than modern European Portuguese. Not to mention things like “ele me não dá”. – Jacinto Feb 7 '18 at 10:17
  • Still regarding enclisis and proclisis, "me dá" may be less natural when pronounced /m'da/, but the Portuguese say "não me dá" (never "não dá-me"), "já me dá", "que me dá", etc. – Jacinto Feb 9 '18 at 10:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.