Current Trends and Research in Language Acquisition

1A teacher on a screen is not a teacher

In 2003 Patricia Kuhl, Feng-Ming Tsao and Huei-Mei Liu gave 9-month-old American babies from English-speaking homes twelve short sessions of Mandarin over four weeks. In one group, Mandarin speakers sat with the babies, read books and played with toys. Other groups saw and heard exactly the same kind of sessions on a video screen, or heard the sound only. Afterwards, the babies were tested on a Mandarin sound contrast that English does not use.

Babies who had met live speakers discriminated the contrast about as well as babies raised in Taiwan hearing Mandarin every day. Babies who had watched the screen or heard the recordings showed no learning at all; they did no better than babies who had heard only English. The same language, delivered without a person to interact with, taught nothing.

21. Interaction and screens

Kuhl’s study and others suggest that, for infants, language learning is social: a live partner directs attention, responds and shares focus. Later work found that toddlers could learn new words through a live video chat with a responsive adult but not from a recording of the same adult (Roseberry, Hirsh-Pasek and Golinkoff, 2014), so what matters seems to be responsiveness, not the screen as such. The World Health Organization’s 2019 guidelines advise no sedentary screen time for children under 2 and no more than one hour a day for children aged 2 to 4.

32. The “word gap” debate

In 1995 Betty Hart and Todd Risley recorded talk in 42 American families once a month for over two years. They estimated that by age 3 children in the poorest families had heard about 30 million fewer words than children of professional parents, and that vocabulary differences followed. The “30-million-word gap” shaped policy and parenting campaigns.

It has also been strongly challenged. The sample was small and from one city; the estimate was extrapolated from one hour a month. Douglas Sperry, Linda Sperry and Peggy Miller (2019) recorded families in five American communities and, when they counted all the talk children could hear, including talk among other family members, found no consistent gap by income. Other researchers dispute that reanalysis. A point on which many now agree comes from studies using all-day recorders: the number of conversational turns may matter more than the number of words heard. Rachel Romeo and colleagues (2018) found that children who experienced more turns showed stronger activity in left-hemisphere language areas and better language scores.

43. Does bilingualism make you mentally sharper?

Studies from the early 2000s, such as Ellen Bialystok and colleagues (2004), reported that bilinguals were faster on tasks requiring them to ignore distracting information, which suggested that constantly managing two languages trains mental control. Later large studies (for example Paap and Greenberg, 2013) often found no such advantage. Angela de Bruin, Barbara Treccani and Sergio Della Sala (2015) then traced conference presentations on the topic to see which were later published in journals.

Figure 1. Studies that supported the bilingual advantage were far more likely to be published (de Bruin and colleagues, 2015).

A 2018 meta-analysis by Minna Lehtonen and colleagues, correcting for publication bias, found no reliable advantage in adults’ mental control. This does not mean bilingualism has no value: speaking several languages brings obvious benefits for communication, culture, travel, work and identity. It means the claim of a general “brain-training” effect is, at best, uncertain.

54. Language of learning

Organisations such as UNESCO recommend that children first learn to read in a language they already speak well, then add other languages, because reading skills transfer and comprehension is easier. For multilingual countries like Rwanda this raises practical questions studied in psycholinguistics: when to introduce a second language of instruction, how to support teachers and learners in the switch, and how to keep all languages growing.

65. Children and machines

AI language models now produce fluent text, but they are trained on hundreds of billions of words or more, while a child hears perhaps tens of millions by school age. Researchers such as Michael Frank (2023) use this gap as a research question: what do children bring, such as social understanding, a physical world to talk about and the ability to ask questions, that lets them learn so much from so little?