Babies pick up complex grammar rules at 4 months, study shows
Babies as young as 4 months are able to identify grammar rules in an unfamiliar language, new EU-funded research reveals. The study, published in the journal PLoS ONE, significantly adds to our understanding of how babies learn languages.
The study was funded by the EU through the CALACEI (‘Universal and specific properties of a uniquely human competence. Tools to study language acquisition in early infancy: Brain and behavioural studies’) project, which received EUR 15 million under the ‘New and emerging science and technologies’ (NEST) budget line of the Sixth Framework Programme (FP6).
Infants pick up languages with a speed and apparent ease that appears startling to parents and researchers alike. Studies on babies have shown that even newborns are able to discriminate between different phonemes within syllables. In addition, very young babies appear to be able to recognise the relationships between adjacent syllables that often occur together.
However, very often grammatically linked elements do not occur next to each other in a sentence; for example, in the phrase ‘is singing’, ‘is’ is separated from ‘ing’ by the stem ‘sing’. ‘The learning of these non-adjacent dependencies is much more difficult than learning adjacent dependences,’ the researchers write. In fact, previous studies have suggested that children may not understand these kinds of rules until the age of 17 or 18 months. ‘That seemed to be very late,’ commented Professor Angela Friederici, Director of the Department of Neuropsychology at the Max Planck Institute for Human Cognitive and Brain Sciences in Germany.
In this study, Professor Friederici and her colleagues sought to determine if 4-month old babies could pick up on these non-adjacent dependences. They picked 4-month olds because research has suggested that babies of this age already have some verbal memory and phonological discrimination skills.
The researchers started by effectively teaching 4-month old German babies some basic Italian. For a little over three minutes, the babies listened to sentences in Italian featuring two simple constructions. One kind of phrase involved the construction ‘sta X-ando’ (i.e. ‘is X-ing’); an example of this type of phrase is ‘[il fratello] sta cantando’, which means ‘[the brother] is singing’. The second kind of phrase included the construction ‘può X-are’ (i.e. ‘can X’); an example here is ‘[la sorella] può cantare’, which means ‘[the sister] can sing’.
After listening to correctly formed Italian sentences like these for three minutes, the infants underwent a little test, during which they heard both correct and incorrect sentences. Incorrect sentences muddled up the constructions, saying things like ‘la sorella può cantando’ (‘the sister can singing’) or ‘il fratello sta cantare’ (‘the brother is sing’). In total, the infants underwent four learning phases, during which they listened to a total of 256 correct sentences; their total learning time was only just over 13 minutes.
During the experiment, the babies’ brain activity was measured. In the early test phases, there was little difference in the pattern of activity registered when the babies heard incorrect Italian sentences. However, by the fourth test period, very different brain activation patterns emerged, indicating that the infants had learnt that ‘sta’ goes with ‘-ando’ and ‘può’ goes with ‘-are’.
‘The present data demonstrate that 4-month-old infants can extract dependencies between non-adjacent elements in sentences from brief exposure to a natural, non-native language,’ the researchers conclude. ‘The emergence of the sensitivity to the grammatical regularities indicates that infants extracted the dependencies within the two pairs of non-adjacent elements (i.e. the auxiliaries and the respective verb suffixes) from correct sentences they had heard during the training phases.’ ’Naturally, at this age infants do not notice content-related errors,’ said Professor Friederici. ‘Long before they comprehend meaning, babies recognise and generalise regularities from the sound of language.’
According to the researchers, the brain activity patterns of the 4-month old German babies when exposed to errors looked more like those of adult native Italian speakers. Native German speakers who were learning Italian as a second language did not respond in the same way. According to the researchers, ‘[this] suggests that native learning may be restricted to a sensitive time window during development.’
Babies and language? It all starts in the womb
A mother’s womb is the ideal place for a baby to pick up elements of what their first language will be, according to new research published in the Current Biology journal. The researchers from Germany and France said the cries newborns make are along the melody patterns typical of the language they heard before they were born.
‘The dramatic finding of this study is that not only are human neonates capable of producing different cry melodies, but they prefer to produce those melody patterns that are typical for the ambient language they have heard during their foetal life, within the last trimester of gestation,’ explained Dr Kathleen Wermke of the University of Würzburg in Germany. ‘Contrary to orthodox interpretations, these data support the importance of human infants’ crying for seeding language development.’
Past studies found that human foetuses have the capacity to memorise external sounds by the last trimester of pregnancy, with a particular sensitivity to melody contour in music and language. A comparison of a mother’s voice with another voice shows that newborns favour their mom’s voice. The babies also perceive the emotional content of messages conveyed via intonation contours in maternal speech – what is commonly referred to as ‘motherese’, a non-standard form of speech used by adults when talking to babies.
Melody plays a central role in a baby’s perceptual preference for the surrounding language and its ability to distinguish between different languages and changes in pitch. Researchers had long speculated that the surrounding language impacted sound production in a baby’s life later rather than sooner. This latest study found the opposite.
By recording and assessing the cries of 60 healthy newborns aged 3 to 5 days old (30 German-speaking and 30 French-speaking), the researchers discovered obvious differences in the shape of the newborns’ cry melodies, based on their native language.
The researchers found that German newborns tend to favour a falling melody contour in their crying, and French newborns tend to prefer a rising melody contour. Dr Wermke pointed out that those patterns are consistent with characteristic differences between the two languages.
According to the team, the new data indicate a significant early impact of native language. Past studies of vocal imitation had demonstrated that babies are able to match vowel sounds presented to them by adult speakers, but only from 12 weeks on. That skill is contingent on vocal control that is impossible to obtain before three months.
‘Imitation of melody contour, in contrast, is merely predicted upon well-coordinated respiratory-laryngeal mechanisms and is not constrained by articulatory immaturity,’ the research showed. ‘Newborns are probably highly motivated to imitate their mother’s behaviour in order to attract her and hence to foster bonding,’ the authors wrote.
‘Because melody contour may be the only aspect of their mother’s speech that newborns are able to imitate, this might explain why we found melody contour imitation at that early age.’
Also participating in this study were researchers from the Max Planck Institute for Human Cognitive and Brain Sciences in Germany, and the École Normale Supérieure/National Centre for Scientific Research (CNRS) in France.
Babies can understand people’s perspective!
A new EU-funded study suggests that tiny tots, even babies, have the ability to understand and consider others’ perspectives. But what is even more intriguing is that it seems to be an automatic response for them, one that is made with virtually no effort. And all this kicks in way before the tot turns a year old. The results are presented in the journal Science.
The research was funded in part by the CALACEI (‘Universal and specific properties of a uniquely human competence. Tools to study language acquisition in early infancy: Brain and behavioural studies’) and DISCOS (‘Disorders and coherence of the embodied self’) projects, both funded under the EU’s Sixth Framework Programme (FP6). CALACEI received EUR 1.5 million under the ‘New and emerging science and technology’ (NEST) Activity and DISCOS is supported by a Marie Curie Research Training Network grant worth almost EUR 3 million.
Researchers call a person’s ability to infer others’ intentions and beliefs the ‘theory of mind’. This ability plays a crucial role in effective social interactions and may have been a central condition in building cooperative human societies, experts say.
Before this study, most researchers believed that children younger than ages three or four did not possess the theory of mind ability. But the results indicate that there is a strong possibility that they actually do. Marie Curie Research Fellow Dr Ágnes Melinda Kovács from the Institute for Psychology at the Hungarian Academy of Sciences and her colleagues investigated this phenomenon by testing adults and seven-month-old infants with a series of animated videos. In these videos, a ball first rolls behind a small wall, and then either stays put, rolls away and returns, or rolls out of view.
The reaction times of both the adults and infants were faster when the animated character’s ‘belief’ about the ball’s location matched the ball’s actual whereabouts, according to the researchers. This was also the case when the animated character had left the screen once the video ended.
The team speculates that even if the people are no longer present, regardless of whether they are young or old, people will still remember their beliefs as alternative representations of the world.
‘Developing tasks that can be used with very young infants will significantly contribute to current efforts to achieve early detection and diagnosis of autism, and will pave the ways towards early intervention techniques,’ Live-Science quoted Dr Kovács as saying.
‘Human social interactions crucially depend on the ability to represent other agents’ beliefs even when these contradict our own beliefs, leading to the potentially complex problem of simultaneously holding two conflicting representations in mind,’ the authors of the study write. ‘Here, we show that adults and seven-month-olds automatically encode others’ beliefs, and that, surprisingly, others’ beliefs have similar effects as the participants’ own beliefs.’
The results of this study will further boost awareness of how important a role the theory of mind ability is, particularly for one’s capacity to infer others’ mental states.
Researchers from Italy and the US contributed to this study.