I have been a fan of Charles Fernyhough since reading his lovely description of the first 3 years of his daughter’s life in ‘The Baby in the Mirror’: ‘ a painstakingly observed, exquisitely written voyage of discovery’, in the words of one reviewer. 

So I pricked up my ears when driving from one school to another to hear him on ‘Woman’s Hour’ last week. He was talking about new research finding that babies are capable not just of picking up information in the womb, but of using that knowledge to shape their own behaviour. They don’t just learn the accent; they reproduce it for themselves.

In a post here  he describes the research:

German researchers recorded and analysed the cries of some very young babies—between 2 and 5 days old—born into two language groups, French and German. There were 30 babies in each group. The analysis of the recordings involved examination of the cries’ ‘melody contours’, which makes use of the fact that the cry of a baby follows a distinctive pattern: first rising in pitch, and then falling, in a single arc. 

The results of the analyses showed clear differences between the language groups. The French babies’ cries spent longer on the rising part of the arc, and the German cries were skewed towards the falling part. These patterns match up to the particular prosodic patterns of the French and German languages, as demonstrated in other studies (and fully evident to listeners to those spoken languages).
What is striking about this new study is that babies aren’t just learning patterns in the womb, but they are also showing an ability to mimic them—which must call for some very sophisticated control over the articulatory system (the system of muscles that allows us to produce speech).

 I keep plumbing the depths of my ignorance about how language develops. Fascinating.