There is a lot involved in learning words! We assume that when someone “knows a word” they (1) comprehend its meaning, (2) could use/produce the word, and (3) use it appropriately. When babies says their first words, we are so proud because we assume they have conquered (1) and (2).
But as babies continue to use these first words, observers quickly realize they might not be doing (3) which then brings (1) back into question. Cogsci-baby #2 has always loved dogs and makes it clear that he would like to see our neighbor’s dog as soon as he wakes up in the morning. “Mung-mung” (the sound a dog makes in Korean is often how Korean babies say dog early on) was one of the first words cogsci-baby2 said (~10 mo) but he often overgeneralized this word. He used to refer to cats, bears, deer, oxen, and foxes as “mung-mung” with little concern for accuracy. Even though overgeneralization is technically incorrect, it’s actually a wonderful behavior because it suggests that he is trying to communicate with words despite his limited vocabulary! By golly, he’s not going to let the fact that he doesn’t know more than 5 words get in the way of connecting with people!
Overgeneralization speaks to one of the classic philosophical questions in language acquisition: how do we demonstrate to babies exactly what a word means when all we do is point to stuff and say “hey that’s a ____!” The philosopher W.V. Quine (1960) pointed out this “indeterminacy of reference” with a now famous example. Imagine we were looking at this scene (below) and we did not speak the same language. And then I pointed to this scene and shouted, “Gavagai!”
What could I possibly mean? Most English-speaking adults would assume that gavagai meant “fish” but it doesn’t necessarily have to be the case. I could have been trying to communicate any number of ideas (e.g., “look over there!”, “swimming!”, “a reef!”, “vertebrates!”, “dinner!”, “beautiful!”). But culturally we differ in the way that we describe scenes. Shown the same animation of a scene, Japanese students are much more likely to describe it holistically and describe the background (“a coral reef”) than American students who are more likely to mention only the foregrounded creatures (“fish”) (Masuda & Nisbett, 2001).
So when I pointed into a tank at the aquarium and said “fish!” how did Nathan eventually figure out what I meant? I could have meant “glass!”, “pay attention!”, “tank!”, “swimming!”, “coral reef!” Amazingly babies everywhere overcome this indeterminacy issue and come to use words appropriately.
Some philosophers/psychologists (e.g., Noam Chomsky) are so convinced of this indeterminacy of reference problem that they conclude that language must be innate. As for me, I’ve never quite bought the nativist hypothesis; the conclusion seems a little premature. Most of the non-nativists believe that somehow, in the richness of life, this indeterminacy problem is tractable even though we may not have all the answers. There are lots and lots of theories about how this process occurs.
But here is one potential mechanism: even though you may not understand “gavagai” the first time I use it, if I always say “gavagai” whenever we pass any tank (even when there are no fish), you could rule out fish. If I always say “gavagai” whenever I point at something, you could rule out tank. This isn’t the only mechanism but learning across time, across situations, and across examples gives language learners a leg up.
So I bring you my newest idea to use technology to kick start accurate word learning. Because Quine is right about the indeterminacy of learning the meaning of a word from one image. But Quine never had google images.
It may be hard to know what the sounds “fish” refer to with one image of a fish… but what about a bunch of simultaneous separate images that all include fish in their various incarnations?! It helps constrain the reference! It’s not a full fix — after all there are non-fish that have the same perceptual components of fish (e.g., dolphins, orcas) — but it’s a step in the right direction!
Complete Nathan productive vocabulary (16.5 months): more*, eat*, all done*, milk*, apple, rice^, bread^, fruit, up, mama, grandpa^, dada, Elmo, fish, hippo^, dog^, bird, no, little, go, fly^, hat, ball, cat, water*^, bottle^, please, bye, amen, open, pouch, sleep^.
Comparison, Complete Amos productive vocabulary (16.5 months): more*, all done*, milk*, up, go, mama, dada, shoe, fruit, please, bye, ball, open, amen, down, boro (Japanese cookie), dog^.
*baby sign, ^Korean