When my first child starting talking, the way she said “baby” was comical: it was basically two grunts, a high-pitch followed by a low-pitch, which when concatenated sounded roughly like the word itself.
As time went on, she learned how to speak with greater precision, but the funny bit was that for a while, whenever she would say “baby” she would revert back to her earliest versions of the word. So you’d hear fair renditions of words and then the two grunts.
Eventually, she worked it out.
Tonight as I spent time with my fourth child, I noticed the same behavior with a different word. And then it occurred to me. We don’t reproduce the sounds that we hear. Instead, we learn how to produce sounds and, independently, map the sounds we hear to the sounds we’ve learned how to produce. Hence, most of us have really bad accents when we learn new languages–we are mapping a new set of different sounds to the set of existing sounds that we already know how to make. Mismatches are inevitable.
I doubt this will be a revelation to anyone but me, but it did get me wondering: are there some people who can reproduce the sounds they hear with anything near fidelity? Such people would be amazing at learning new languages, among other things.
Is this a skill I can learn? It seems so much more efficient than having to map what I hear to what I know how to say.
Does anyone know?