So in the episode, we talk about how arbitrary the connection between words and their meanings really is. Generally, there’s not really a deep link between how a word sounds and what meaning it has. But there are a number of cases where this isn’t true: in languages like Japanese, there’s a lot of ideophones, words that have sound symbolism even if the actions or states they describe aren’t immediately related to actual sound effects in the world. I mean, on the face of it, the sounds in something like ピカピカ [pikapika], which mean sparkling or a flash of light, don’t really seem associated with the actual visual event.
But when we dig into the sounds beneath, we can see some patterns bubbling up. Like, combining velar stops like [k] or [g] with the r-like [ɾ] gets you ideophones related to rotation. If it’s a light object, use a voiceless velar, like [kɯɾɯ] for a light object spinning once around an axis, or [koɾo] for a light object rolling a little. A heavier object? Go for the voiced [g], like in [gɯɾɯ] or [goɾo]. And if it’s a repeated movement, just say it twice, so a continually heavy spinning object would go [gɯɾɯgɯɾɯ]. And we can find a lot of other Japanese ideophones we can break down along those lines.
This connection raises an interesting question: is it easier for kids to learn words that have that sound connection? Or are they just as good at learning words where there’s no link? Generally, there are so many arbitrary words that it seems like kids might be pretty streamlined for word-learning regardless. But it’s also fairly easy to picture words that match in some way helping kids out. So to get at this question, Mutsumi Imai and colleagues ran through a series of experiments.
The first step is to come up with a good set of words to have the kids play with. But they couldn’t use real Japanese words, because then the kids might have had exposure in advance; you can’t run an experiment on how kids learn new words when they know your words already. So the researchers came up with a new set of ideophones to describe ways you could walk - slow or fast, steady or unsteady, heavy or light, etc. - that matched the general Japanese rule scheme. Like, [nosɯnosɯ] would be slow walking with heavy steps; [n] is sluggish, [s] for friction. Then, they tested them on Japanese adults, and found that they could get them to match the new words to the right videos of people walking 100% of the time. (Interestingly, English speakers were also tested with these novel words, and scored above chance, which is in line with the results reported for Dutch speakers in the paper we talked about in the episode. That’s a ripe topic for future discussion.)
So then it was time to move onto the kids. The researchers tested 18 2-year-old and 17 3-year-old kids, and gave them a variation on the task, using a live speaker to ask the little Japanese speakers to match the videos to the new words. And already, they showed they could do it above chance: the 2-year-olds scored 65.7%, and the 3-year-olds at 75%. Which is already amazing - by that point, they know enough about sound-symbolism to work this out.
But now we want to know whether there’s an advantage for these words. And so the researchers got a new group of 34 3-year-olds to check on. And though they used their ideophones that they’d been testing so far, they also got a bunch of new words: ones that were made-up non-sound-symbolic verbs. They came up with 6 verbs, and gave them meanings that had nothing to do with their sounds. They split the kids into sound symbolism vs. non-sound symbolism groups, and then showed them videos to train them on each of the words.
The big step was this, though: we want to know whether they’re actually learning the verb. And that means like associating them with the actual actions beyond the one video. So after the kids were trained, they showed them new videos to choose from: one with the same walking actor from the first video, but walking differently; and one with a different actor, but with the same walking style. And kids had to pick out which was correct.
So what happened? The kids from the non-sound-symbolism group - so the one who got regular verbs - were basically just guessing about how the words worked; they got it right 54% of the time. But the sound-symbolism kids? They scored at 82% - way above chance, and way above the other group. They were really using the sounds to learn the words more easily.
So here’s what we find: more sound symbolism gives young kids a better chance at picking up new words. Languages that use a lot of these words, like Japanese, give kids an extra tool to grasp at meaning. And those kids aren’t just sensitive to how those sound patterns work, they really use it to learn. Just another way that kids are awesome at picking up language!
So how about it? What do you all think? Let us know below, and we’ll be happy to talk with you about sound symbolism, and how it could help in learning new words and connecting new concepts. There’s a lot of interesting stuff to say, and we want to hear what interests you!
Previous Topic: Next Topic:
Desert Island Words Coming soon!