Regarding ducks

I have been lately discussing the verbal development of toddlers with a friend of mine. Of course the amount of words a very young child uses is in no way a direct representation of their understanding, but nevertheless it gives us a little window into what's going on in the human brain at that age. I am mostly interested in how images and words make sense in a young child's mind. For example most toddlers begin by using a word to refer to a wider group of things, such as "mummy" for all female people, including elderly ladies and young girls. Or the word "horse" for horses, donkeys and zebras. Some then tend to be more specific and say "brown horse" or "baby horse" before they start using the actual words for different animals.

So far so good. My big amazement started when I realised that toddlers can recognise actual living animals even if they have never seen a photograph or a film of these animals, but only an illustration of them in a book. For example my daughter pointed at a duck in the park and said "dack". She had never seen a live duck before, but I went around the house finding which illustrations of ducks she had come across. The illustrations of ducks in the books are only the personal artistic interpretation of different illustrators and quite different from each other. So can the human mind recognise an overall "duck" shape in these illustrations that is instantly recognisable as that of an actual duck? Or is the repetition of the word "duck" by an adult each time the child looks at the illustration of a duck that links together in the brain all the different illustrative versions of a duck?

My friend's baby pointed and named an actual elephant at the zoo at age 18 months and the only one he had seen was Elmer who (for those poor of you who are enduring this although you have no kids) is a colourful elephant from small square books, I don't know how many times smaller than an actual elephant!

The different ducks my daughter had
access to around the house

Elmer the lovable elephant my friend's
baby  was exposed to

And then to top it all off I happened to watch a program on television about the capacity of the human brain for facial recognition. Basically if any toddler was exposed to, for a example, a flock of 30 ducks for a substantial period of time, they would be able to tell apart every single one of those ducks. And also pick them out amongst thousands of ducks. That applies to any species they were in contact with in their very early years. As adults, we are of course able to pick out a photo of our partner's face for example amongst thousands of images of people's faces. And we don't think much about this ability of our brain. Because it was programmed to identify even the most minuscule facial differences and got exposed to so many numerous different faces and facial expressions. This development happens mainly before age 5, so humans that were not exposed to other human contact by that age, are almost unable to regain this ability. But if you were for example to have grown up amongst chimpanzees (this program was about cases of children who did) you would be able to tell each one of them apart and pick them out amongst thousands of individuals.
So where do words come into this? And my confusion over the duck issue remains. Is it a visual learning of the word "duck" or a verbal  trigger that prompts the mind to learn what an image of a duck can look like?