From Phys.org: How many languages can babies learn? Study shows how Ghanaian babies grow up speaking two to six languages.
Africa is a multilingual continent and many adults speak several languages fluently. An empirical study by a research team led by the Potsdam psycholinguists Prof. Dr. Natalie Boll-Avetisyan and Paul O. Omane now shows that the roots of this multilingualism can be found in infancy: In Ghana, most babies grow up multilingually, with most of them coming into contact with two to six languages and just as many regular speakers of each language. The researchers also showed that the babies heard some languages primarily indirectly—i.e. via radio, television or background conversations—while other languages were used by their caregivers to directly communicate with them. The results of the study have now been published in the journal Cognitive Development.
The study, which examined 121 babies aged three to twelve months in Accra, the capital of Ghana, demonstrates a remarkable variety of language input in the early months of life. The children are regularly exposed to two to six languages. Strikingly, the number of caregivers the children have also ranges between two and six, and babies who have more adults in their daily lives who regularly take care of them also hear more different languages. In Ghana, families often live in so-called “compound buildings,” where many everyday interactions take place in the courtyard, where family, neighbors and other relatives play an important role in the lives of children.
“The idea that a child learns only one particular language from a single caregiver, as is often assumed in Western cultures, does not apply to these communities. Rather, children are surrounded by a rich spectrum of linguistic inputs from the very beginning,” says O. Omane, the first author of the study. […]
A key finding of the study is the distinction between direct and indirect language input. While English is primarily acquired through indirect channels such as television and official communication, children receive most of the local languages (such as Akan, Ga and Ewe) through direct contact with their caregivers. Accordingly, the proportion of direct input is higher in the local languages than in English, which is predominantly present as indirect input. […] As a result of their empirical study, the researchers call for a broader view in language research. The common assumptions do not reflect the diversity and complexity found in other cultural contexts such as Ghana. The study makes it clear that it is not only the number of languages a child hears, but also the diversity of people and the different forms of input that have a decisive influence on language acquisition.
Thanks, Bathrobe!
To make the numbers more precise, the distribution of the 121 subjects is approximately: 3 languages, 50%; 2 or 4 languages, 20% each; 5 languages, 9%; 6 languages, 1% (1 subject). The specific language combinations are listed on table 2 of the paper. All include Akan, and most Ghanaian English.
Also interesting: “For 14% of infants with an older sibling in our data, the older sibling introduced a language not spoken by parents and other caregivers.”
That is interesting! And once again, I’m grateful to have grown up in several multilingual environments.
*double-take*
*running out of room*
*calculating for 20 minutes*
*coming back*
Yes, it is obvious.
Not surprised, obviously …
Having said that. Accra may be a bit of a special case, given that the traditional local language is Ga, but Akan is spoken by most Ghanaians in the south, including in Accra, where it’s non-native; moreover, Accra is not very far away from the predominently Ewe area. If they ran the same study in Kumasi, I suspect you’d just get Akan and English predominantly (and Hausa if you included the zongo.)
Still, the overall point is entirely valid: multilingualism is normal in Ghana, including in the towns and cities
Yes, it is obvious.
Not to me. Does it come from the sibling shifting to a language different than the parents’ household language or languages? Does it imply that the sibling uses the language learned from, say, a different mother, like a previous wife or a co-wife?
There are obviously differences in approach. When each of my three sons was a not-yet-anything-speaking baby, his maternal grandparents would address him exclusively in (baby-talk-intonation) English, even while carrying on conversations with my wife in a complicated macaronic/code-switching mixture of English and Sinitic (substantially more Taiwanese than Mandarin but my untutored ear can’t reliably tell when that second-order code-switching between the two Sinitic options happens). As if they were acknowledging that these not-yet-verbal babies were Anglophones-by-destiny. The youngest boy (aged 3) is currently very interested in practicing his counting-to-ten skills but has been indifferent to my offers/attempts to model for him how to count to ten in various languages other than English. (I know I tried German and Japanese and maybe something else.)
Or from a different uncle or third-cousin-twice-removed or friend of the family or whoever happened to be present and is no longer.
Is there any substance to my (entirely anecdotal) impression that babies growing up in a multilingual environment take longer to begin to speak than in a monolingual, but then something ‘clicks’ in their brain and they catch up/get ahead very fast, including figuring out which utterances belong in which language?
I mean balanced input from “direct contact with their caregivers” rather than ‘indirect language input’, or later input from school/nursery.
I’m not sure if Esmeralda Bermudez’s infuriating and inspiring story has come up here before, about her experience raising a multilingual child in the United States.
, the older sibling introduced a language not spoken by parents and other caregivers.”
This is quite common in Europe and the U.S. among immigrant families.
I think it’s a different situation. In Europe and the U.S. the older sibling might introduce the prevalent majority language. But here the prevalent majority languages (Akan, Ghanaian English) are heard by virtually all babies. So those 14% sibling languages would be something else.
On the elder-sibling thing: a lot of Ghanaians pick up both English and Akan properly for the first time when they go to school; in fact, in the Bawku area when I was there, whether people could speak Twi was largely predictable from whether they’d been to school. (I had one colleague who spoke only Kusaal and English, but he was unusual.) Bawku is very different from Accra in terms of the local sociolinguistics, but I imagine there are similarities there.
Church, too, is much more important in these things in Ghana than in Europe or America. The church attached to the Presbyterian Hospital supported Kusaal in theory, but in practice it worked in English and Twi; though this was also because most Presbyterians were southerners: Kusaasi Christians were then relatively few, and were mostly either Catholics or Assemblies of God.
In general, what you get is a sort of stable polyglossia where people know several languages but tend to use each one in a particular niche. Talking to preschool children directly in most families is going to be in the most-local language, but the children are going to overhear other languages regularly, even at home.
Vanya, Y – perhaps a common thread between both scenarios could involve the older sibling acquiring a new tongue in the schoolroom/yard, and subsequently passing their new linguistic skills onto the younger one? While waiting in a doctor’s reception in December last year, witnessed a mother speaking to her primary school age children in an Eastern European language (not being au fait with any, couldn’t tell you which exactly), yet once she left the room, the latter began conversing with each other in English, so code-switching evidently becomes a natural practice at an early age.
But here the prevalent majority languages (Akan, Ghanaian English) are heard by virtually all babies. So those 14% sibling languages would be something else.
What’s the argument here? That if there were parents not using those languages with the infant, you would expect some of the sample to not have contact with those languages at all? It doesn’t look like there were too many infants with parents as the only inputs, and while the authors do point out how different the picture painted by the study is from the deliberate one parent-one language approach popular in some Western contexts, I don’t think that rules out the 14% being bumped up by parents deliberately leaving to the older sibling – it’s not quite the same as splitting languages between parents.
The situation in West Africa called to mind for me the situation in mainland Southeast Asia. To quote Enfield’s Areal linguistics and Mainland Southeast Asia (2005) (see Researchgate):
Mainland Southeast Asia provides a dramatic demonstration of the areal phenomenon in linguistics: When languages are spoken historically in the same location they often show significant parallels in the organization of a wide range of structural domains, whether the languages descend from the same historical source. The effects of areal diffusion raise fundamental questions for the traditional essentialist vision of languages as entities with offspring that diverge, with shared innovations marking divergent branches and internal processes of evolution accounting for diversity among modern languages. Recent theoretical and empirical research on linguistic diversity, language change, and social diffusion of innovation argues for a unit-based approach to language change and relatedness, where the units of analysis are individual speakers and individual linguistic items. This review begins with discussion of the language situation in Mainland Southeast Asia, where the language “genealogies” have been dramatically permeated by socio-historical contact, then explores theoretical and methodological implications for research on language both generally and in its areal context.
I am also reminded of what I’ve read in the distant past (sorry no references), to the effect that multilingualism in inland Southeast Asia is to a considerable extent a matter of switching vocabulary, since the languages themselves are quite similar grammatically.
I’m not at all familiar with the language genealogy of West Africa, or Southeast Asia, for that matter, but I’m wondering to what extent multilingualism in that area is aided by typological similarities among languages. Is picking up a different language simply a matter of inserting different vocabulary and morphology into similar grammatical frameworks, with little need to reconfigure the syntax? Or is it more like the difference between, say, Navajo, Arabic, and Chinese?
All the indigenous languages of Ghana are “Niger-Congo”, and nearly all are Volta-Congo, and thus really genetically related, not just Related According to Greenberg.
On the other hand, Volta-Congo is pretty diverse internally, and in terms of vocabulary and morphology you would be hard put to detect any real connection unless you were into comparative historical linguistics. (Akan soma “send” and Kusaal tʋm “send” are actually exact cognates, for example, but that really wouldn’t help much in practice: and that’s one of the more obvious ones.)
However, though especially the Kwa languages in the south and the Gur languages in the north have little or nothing in common in the mechanisms of their flexion and derivation, there’s a lot of similarity in the kinds of grammatical distinctions that they make – a primary distinction in verb flexion between perfective and imperfective aspect, for example, with tense marking a much more peripheral part of the system.
Similarly, though the languages differ quite a lot phonologically, they all have lexical tone, and the great majority actually have quite similar tone systems (terracing H versus L with emic downsteps.)
Similarities become much more apparent when you start looking at semantics. It’s quite common to find words with pretty much identical semantic ranges in only distantly related Ghanaian languages, where English doesn’t match at all (for example, in both Kusaal and Twi, the “mouth” of a river is in fact its bank.)
This sort of similarity is still there even when you start looking at Ghanaian Hausa, which is completely unrelated genetically to any indigenous Ghanaian language. And I’ve even found that accounts of things like focus in Hausa can shed light on analogous things in Kusaal – some of this is surely just widespread cross-linguistic patterns, but a lot of it is definitely regional Sprachbund phenomena.
And the similarity even extends to Nigerian Pidgin English, which struck me as quite clearly a West African language when I encountered it, bioprogram be damned.
It’s not simply a matter of inserting different vocabulary and morphology into similar grammatical frameworks; although almost all the languages are SVO, and all the Ghanaian ones use specialised possessed nouns as pre- or postpositions, they differ considerably in things like word order within phrases. But the semantics are often strikingly parallel, and in general it doesn’t even begin to approach the divergences between Navajo, Arabic, and Chinese.
I once asked my chief outpatient nurse, who spoke five languages pretty much perfectly as far as I could tell (including Hausa and English), which was the most difficult language. He immediately replied: “English, because it’s the most different.”