Showing posts with label historical linguistics. Show all posts
Showing posts with label historical linguistics. Show all posts

Wednesday, September 16, 2009

Pronoun borrowing in Portuguese

While some linguists seem to be especially talented at finding exotic features in little-known languages, I am more likely to find intellectual reward in uncovering unusual facts in familiar places. For instance, in my native Caipira dialect of Brazilian Portuguese, the plural marker -s presents a peculiar distribution (when compared with other Romance languages), occurring with interjections and other words traditionally regarded as "invariable." Its distribution is, in all relevant syntactic aspects, that of a second-position clitic (as I analyzed it back in 2001). A little-known fact, such evolution of a suffix into a second-position clitic may have interesting implications for grammaticalization studies.

The list of "exotic" features which can be found at home can be further extended. Although Indo-European languages are hardly mentioned in association with the topic "pronoun borrowing," Portuguese provides a rather illustrative example. In (Brazilian) Portuguese, French moi is commonly used informally, mainly in a sort of tongue-in-cheek "style." Notice that such restriction in usage--to informal, humorous situations--is not uncommon in other instances of pronoun borrowing; in the oft-mentioned examples from Southeast Asian languages, borrowed pronouns may convey different degrees of formality.

Male vs. female speech in Portuguese

I suspect this usage of moi in Portuguese can be considered as typical of female speech, an impression that seems to be corroborated by a quick internet search (look up "pra cima de moi", "para moi", etc.). Notice that diminutive moizinha 'little me (fem.)' is also common, unlike moizinho "little me (masc.)'. Again, this "genderlectal" nature of the borrowing turns out to be rather reminiscent of well-known cases of pronoun borrowing. I'm aware of at least two clear South American cases in which male/female speech distinctions were a result of language contact. The classical example is Island Carib, but another example closer to home (and also involving pronouns) is Cocama, such as described by Ana Suelly Cabral in her PhD dissertation. In both cases, the role of language contact in the development of male/female speech distinctions is clear, since the forms used by males and females are from different genetic sources. The fact that females and males tend to respond differently to language change (including contact-induced change), besides being thoroughly demonstrated by Labov and others, is further illustrated here by the interesting use of French moi in Portuguese. [By the way, if you're a speaker of Portuguese who uses this word, please fill out the form available here.]

Friday, October 12, 2007

"Palavras muito usadas não têm som alterado" [sic]

Este é o título de uma nota publicada ontem no Estadão (divulgada no Jornal da Ciência). A começar pelo título, a matéria dá a parecer que um estudo publicado agora na revista Nature teria concluído que "quanto mais uma palavra é usada, menos o som que a representa muda". Para qualquer um que aprendeu na escola que vossa mercê veio dar no português você (e, no português coloquial, simplesmente ), o texto do Estadão seria de encabular.

Lendo o resto da nota, porém, percebe-se que parece haver uma confusão entre a alteração da forma fonológica de uma palavra e a sua substituição por outra. Palavras de alta freqüência tendem a ser substituídas com menos freqüência. Mas mudam sim, e muito -- a ponto de se tornarem irreconhecíveis. Pode-se dizer, aliás, que quanto mais uma palavra é usada, mais sujeita é a simplificação (ou erosão) fonológica, como ilustrado pelo exemplo mencionado acima (vossa mercê > você > ).

Um exemplo clássico nos estudos do Indo-Europeu (para citar uma das palavras mencionadas na matéria) é a palavra armênia para "dois", erku, que é -- ainda que não pareça -- cognata com duo em latim e, claro, dois em português (vide Antoine Meillet, The Comparative Method in Historical Linguistics, 1967, p. 18). E, no tronco Macro-Jê, um morfema derivacional extremamente produtivo, que deriva nomes "de agente", é preservado em pelo menos duas famílias (Jê e Karajá), mas as semelhanças fonológicas são, à primeira vista, difíceis de se detectar (Karajá -du, Xerente -kwa).

Se isto é tudo, o estudo, então, não chega a nenhuma novidade bombástica. A idéia de que há partes do vocabulário que são mais estáveis (ou seja, são substituídas menos rapidamente que outras) é a razão que subjaz, entre outras, à lista de Swadesh, na qual muitos se baseiam para diagnósticos (preliminares que sejam) de parentesco e de profundidade temporal de famílias lingüísticas.

"Milestone"


An Associated Press article about this study doesn't do much better in terms of explaining the findings, and it seems to exaggerate their significance even more:
"Scientists have uncovered what might be called the law of language evolution: the more a word is used, the less likely it is to change over time. [...] [I]f [the study] also holds for other languages, it would be a milestone in understanding one of humanity's defining attributes."

Milestone? Well, for linguists, this seems to have been textbook wisdom for a while now. As a (former) journalist, I understand the media's attempts to make Sci & Tech news more exciting to the general public. But, as a linguist, I find the truth of the facts exciting enough.