Source: TW
In a new study, out in Attention Perception and Psychophysics, @ananyapassi and I have cool insights to share about a fascinating cross-modal association called the Bouba-Kiki effect. Consider the shapes below: which shape would you call a Bouba? 1/n https://t.co/jt0iwYW2ky Vote on it before you go any further!! Which shape would you call a Bouba?
If you’re like most of our participants, you would have call the rounded shape on the left as a “Bouba” and conversely, the angular shape on the right as a “Kiki”. But why does this strange association occur between shapes and words? Such effects may have driven the evolution of language, and of course are interesting aspects of perception and the brain in their own right, so scientists have always been curious about how these effects arose. This effect was first noticed by Kohler in 1965, and called the baluba/takete effect and was later popularized by Ramachandran and Hubbard in 2001, who claimed that this effect arises because our mouth makes an rounded shape when we utter “Bouba” and an angular shape when we utter “Kiki”. So the widely accepted explanation for why humans show a Bouba-Kiki effect is that it is due to the sensorimotor association between mouth shape and the associated sounds made by our mouths.
While its true that our mouths make angular/rounded shapes while uttering Bouba or Kiki, could there be other explanations for this effect? We thought, maybe this effect arises due to systematic associations between object shapes and the sounds they make. For instance, rounded objects make low-frequency sounds and sharp objects make high-frequency sounds. But, as it often happens in science, the hard part is not in thinking about explanations but coming up with the right experiment to prove/disprove them!
We realized that if you play a spoken word backwards in time, it becomes much less pronounceable, e.g. the word “chitiki” (https://t.co/AIpwYHySDb) and and its reversed version (https://t.co/tqrrQCGiDe). Importantly, both sounds have the exact same frequencies! If the Bouba-Kiki effect is due to mouth shape or speech properties, then it should be greatly reduced or even absent for reversed words since they cannot be pronounced. However if it is due to shape-sound associations, it should remain. This was our key prediction. And voila! Not only did we see a Bouba-Kiki effect for reversed versions of spoken words, we also saw it for sounds made by objects when they are struck, such as a pillow or metal. This cannot be explained by the mouth-shape explanation!
If the Bouba-Kiki effect is really due to sound properties rather than speech properties, then some simple sound property should predict the size of the effect. And indeed, we found that the effect strength is highly correlated with the mean frequency of the sound! We performed a bunch of extra analyses to show that the effect cannot be predicted by how pronounceable a word is, or even by the aspect ratio of the mouth while speaking.
We cannot conclusively rule out the mouth-shape or speech explanation, since there could always be an untested aspect of mouth shape or speech that could explain the Bouba-Kiki effect, but that is also the nature of science. As Bertolt Brecht says, “The aim of science is not to open the door to infinite wisdom, but to set a limit to infinite error”+++(5)+++. We can only limit how much speech properties might contribute to the effect but we can never disprove their contribution, because there could be some untested property that really works, or they could have some tiny contribution too hard to measure.
Overall, we think our results weigh strongly in favor of the Bouba-Kiki effect arising due to sound properties and NOT due to speech properties - which is pretty much the title of our study🙂. Kudos to @ananyapassi for persevering through it despite coursework and even the pandemic! And thanks as always to the @India_Alliance for their funding support (although it has been scarily erratic lately).