How good are babies at differentiating the languages they hear?

In an increasingly bilingual world, parents often occupy themselves with the age-old question of how to (effectively) raise a child to be bilingual. Beyond that, perhaps a more theoretically fundamental question to unpack is how infants come to differentiate between languages in the first place. What are the mechanisms that account for this cognitive differentiation in a baby's brain? How early does this happen, and does bilingual learning take place as early as this differentiation comes to form?

An understanding among psycholinguists is that infants adapt to their dominant language (usually a 'native language') as early as birth, and up to age two there tends to be a decrease in the active processing of sounds labelled as 'non-native'. This filtering process suggests that the infant brain has the capacity to 'filter' languages. Further, another idea is that regardless of the number of languages one knows, languages in the brain are not separate, but rather, represented conceptually under one unified system. This finding was first revealed in adult populations, and later the same was seen to be applied to infants as well.

Delving deeper into the properties of language, some may question whether the similarities between languages may play a role in infants' differentiation of languages. While there is experimental research to support the idea that infants of age 1 and 8 months can indeed process and acquire a linguistically distant second language (e.g., the case of babies of French backgrounds responding to Cantonese), linguistically similar language pairs may be trickier to differentiate. Given how listening is typically the initial means of language exposure in infancy, the question most likely depends on the dimensions of phonetic and accentual similarity between the babies' two languages. To illustrate, a speculation is that infants may be more able to tell Spanish and French apart than to tell Spanish and Catalan apart, even though all three languages belong to the same linguistic family. Whether or not extremely high levels of linguistic similarity interfere strongly with the acquisition of both the first and second languages is by far not yet thoroughly researched.