The Bouba/Kiki Effect And Sound Symbolism In CLIP

The bouba/kiki effect is the phenomenon where humans show a preference for certain mappings between shapes and their corresponding labels/sounds.

One of the above objects shall be called a bouba, and the other a kiki

The above image of 2 theoretical objects is shown to a participant who is then asked which one is called a ‘bouba’ and which is called a ‘kiki’. The results generally show a strong preference (often as high as 90%) for the sharply-pointed object on the left to be called a kiki, with the more rounded object on the right to be called a bouba. This effect is relatively universal (in languages that commonly use the phonemes in question), having been noted across many languages, cultures, countries, and age groups (including infants that have not yet learned language very well!), although is diminished in autistic individuals and seemingly absent in those who are congenitally blind.

What makes this effect particularly interesting is less so this specific example, but that it appears to be a general phenomenon referred to as sound symbolism: the idea that phonemes (the sounds that make up words) are sometimes inherently meaningful rather than having been arbitrarily selected. Although we can map the above two shapes to their ‘proper’ labels consistently, we can go much further than just that if desired.

Which is a takete, and which is a maluma? Only you can decide.

We could, of course, re-draw the shapes a bit differently as well as re-name them: the above image is a picture of a ‘maluma’ and a ‘takete’. If you conformed to the expectations in the first image of this section, it’s likely that you feel the maluma is the left shape in this image as well.

We can ask questions about these shapes that go far beyond their names too; which of these shapes is more likely to be calm, relaxing, positive, or explanatory? I would certainly think the bouba and maluma are all four of those, whereas the kiki and takete seem more sharp, quick, negative, or perhaps even violent. If I was told that the above two shapes were both edible, I can easily imagine the left shape tasting like sweet and fluffy bread or candy, while the right may taste much more acidic or spicy and possibly have a denser and rougher texture.

Sound symbolism

The idea that large sections of our languages have subtle mappings of phonemes to meaning has been explored extensively over time, from Plato, Locke, Leibniz, and modern academics, with different figures suggesting their theorized causes and generalizations.

Some of my favorite examples of sound symbolism are those found in Japanese mimetic words: the word jirojiro means to stare intently, kirakira to shine with a sparkle, dokidoki to be heart-poundingly nervous, fuwafuwa to be soft and fluffy, and subesube to be smooth like soft skin. These are some of my favorite words across any language due to how naturally they seem to match their definitions and how fun they are to use (more examples because I can’t resist: gorogoro may be thundering or represent a large object that begins to roll, potapota may be used for dripping water, and kurukuru may be used for something spinning, coiling, or constantly changing. There are over 1,000 words tagged as being mimetic to some extent on JapanDict!).

For fun I asked some of my friends with no prior knowledge of Japanese some questions about the above words, instructing them to pair them to their most-likely definitions, and their guesses were better than one would expect by random chance (although my sample size was certainly lacking for proper scientific rigor). The phonestheme page on Wikipedia tries to give us some English examples as well, such as noting that the English “gl-” occurs in a large number of words relating to light or vision, like “glitter”, “glisten”, “glow”, “gleam”, “glare”, “glint”, “glimmer”, “gloss”. It may also be worth thinking about why many of the rudest and most offensive words in English sound so sharp, often having very hard consonants in them, or why some categories of thinking/filler words (‘hmm’… ‘uhhh…’) sound so similar across different languages. There are some publications on styles of words that are found to be the most aesthetically elegant, including phrases such as ‘cellar door’, noted for sounding beautiful, but not having a beautiful meaning to go along with it.

Sound Symbolism in Machine Learning with CLIP

I would guess that many of the above aspects of sound symbolism are likely to be evident in the behavior some modern ML models as well. The reason for this is that many recent SOTA models often heavily utilize transformers, and when operating on text, use byte-pair encoding (original paper). The use of BPE allows the model to operate on textual input smaller than the size of a single word (CLIP has a BPE vocab size of 41,192), and thus build mappings of inputs and outputs between various subword units. Although these don’t correspond directly to phonemes (and of course, the model is given textual input rather than audio), it’s still likely that many interesting associations can be found here with a little exploration.

To try this out, we can use models such as CLIP+VQGAN or the more recent CLIP-guided diffusion, prompting them to generate an image of a ‘bouba’ or a ‘kiki’. One potential issue with this is that these words could have been directly learned in the training set, so we will also try some variants including making up our own. Below are the first four images of each object that resulted.

four images of “an image of a bouba | trending on artstation | unreal engine”
four images of “an image of a kiki | trending on artstation | unreal engine”

The above eight images were created with the prompt “an image of a bouba | trending on artstation | unreal engine”, and the equivalent prompt for a kiki. This method of prompting has become popular with CLIP-based image generation models, as you can add elements to your prompt such as “unreal engine” or “by Pablo Picasso” (and many, many others!) to steer the image style to a high-quality sample of your liking.

As we anticipated, the bouba-like images that we generated generally look very curved and elliptical, just like the phonemes that make up the word sound. I have to admit that the kiki images appear slightly less, well, kiki, than I had hoped, but nonetheless still look cool and seem to loosely resemble the word. A bit disappointed with this latter result, I decided to try the prompt with ‘the shape of a kikitakekikitakek’ instead, inserting a comically large amount of sharp phonemes all into a single made-up word, and the result couldn’t have been better:

the shape of a kikitakekikitakeki | trending on artstation | unreal engine

Having inserted all of the sharpest-sounding phonemes I could into a single made-up word and getting an image back that looks so amazingly sharp that it could slice me in half was probably the best output I could have hoped for (perhaps I got a lucky seed, but I just used 0 in this case). We can similarly change the prompt to add “The shape of” for our previous words, resulting in the shape of a bouba, maluma, kiki, and takete:

the shape of a bouba (top left), maluma (top right), kiki (bottom left), and takete (bottom right)

It’s cool to see that the phoneme-like associations within recent large models such as CLIP seem to align with our expectations, and it’s an interesting case study that helps us imagine all of the detail that is embedded within our own languages and reality – there’s a lot more to a word than just a single data point. There’s *a lot* of potential for additional exploration in this area and I’m definitely going to be having a lot of fun going through some similar categories of prompts over the next few days, hopefully finding something interesting enough to post again. If you find this topic interesting, some words you may want to search for along with their corresponding Wikipedia pages include: sound symbolism, phonestheme, phonaesthetics, synesthesia, ideathesia, and ideophone, although I’m not currently aware of other work that explores these with respect to machine learning/AI yet.

Thanks for reading! I appreciate feedback via any medium including email/Twitter, for more information and contact see my about page

This Anime Does Not Exist

This Anime Does Not Exist was launched on January 19th 2021, showcasing almost a million images (n=900,000 at launch, now n=1,800,000 images), every single one of which was generated by Aydao‘s Stylegan2 model. For an in-depth write-up on the progression of the field of anime art generation as well as techniques, datasets, past attempts, and the machine learning details of the model used, please read Gwern’s post. This post is a more concise and visual post discussing the website itself,, including more images, videos, and statistics.

The Creativity Slider and Creativity Montage

Previous versions of ‘This X Does Not Exist’ (including anime faces, furry faces, and much more) featured a similar layout: an ‘infinite’ sliding grid of randomly-selected samples using the javascript library TXDNE.js, written by Obormot. Something more unique about Aydao’s model was that samples remained relatively stable, sometimes significantly increasing in beauty and detail as the ‘creativity’ value (see: psi) was increased all the way from 0.3 to 2.0. This led to the creation of the ‘creativity slider’ as well as linking every image to a page with a tiled layout showing every version of the image:

“The more entropy you give me, the more it makes me want to smile!”, 18 samples of image # X from creativity 0.3 to 2.0
One of the best and most common effects of higher creativity is better colorization, including shading, reflections, vibrancy, color diversity, and more
Another tile of seed 7087, showcasing stability with an increasingly interesting artistic style and details

Although the above two are among my favorite examples of creativity montages, many users found some other interesting things that increased creativity could do:

Sometimes it appeared as if ‘creativity’ was simply an alias for the increase in volume of a certain bodily area coupled with a proportional reduction of garment-covering surface area (seed 30700)
Sometimes increased creativity also leads to mitosis (seed 28499)
Sometimes you get… a lot of mitosis?

Selected artwork

The best way to demonstrate the stunning potential of this model is to show some of the samples that users have enjoyed the most:

The original thumbnail image for This Anime Does Not Exist: “Notice me, Onee-chan!”
Other earlier candidates considered for the website’s thumbnail (seeds: 1013, 4606, 8674, 8859)
Four popular images selected by Twitter users
Images from seeds 16736, 23655, 62731, and 93887
Images from seeds 84995, 50428, 10157, and 52476

Weirdest images

Incidentally, it also seems that users love sharing images that are not the most beautiful, but the weirdest as well. The below four were among the most popular images on the website during the first few hours, when there was only n=10,000 of each image:

A montage of the interesting and popular results from images 3313, 7820, 4437, and 3103. Sample 3103 would be a particularly good album cover for a death metal band.

Collapsed images

Sometimes the result collapses enough to lead to an image that, although pretty, does not at all resemble the ideal target:

interesting but unintended results from samples 0544, 31975, 3997, 0557


In many cases the model will produce writing which looks distinctly Japanese, however upon closer introspection, is not legible, with each character closely resembling distinct counterparts in Japnese scripts, however diverging just enough to produce confusion with a lingering feeling of otherworldlyness.

Although some characters are easily recognizable, many are not, and are nonetheless *usually* combined in incoherent manners

Videos and gifs

As it’s possible to produce any number of images from the model, we can also use these images to produce videos and animated gifs. The primary style of this is referred to as an interpolation video which is produced by iterating through the latent variables frame by frame, transitioning in between different samples seamlessly:

Interpolation video produced by arfa

Additionally, I decided to make a few videos that used a constant image seed, but modified only the creativity value instead (instructions on how I did this are here):

I chose this particularly wholesome sample to demonstrate a gif with creativity 0.3-2.0 and a frame difference of 0.01

Statistics and users

After the first day, the website had served over 100 million requests, using over 40TB of bandwidth:

First-day traffic statistics from CloudFlare’s Traffic Analytics

At launch, the largest contributor to traffic by country was the United States, followed by Russia, but over the next two days this quickly shifted to Japan.

A map showing which of Cloudflare’s data centers served the most DNS query results for the domain

As of writing this post (Jan 21st, 2021), the largest sources of traffic appear to be from Twitter, Hacker News, Gigazine, Hatena, and Himasoku.


Compare to similarly-trafficed websites, was relatively cheap, thanks to not requiring server-side code (serving only static content):

  • Domain ‘’: for two years from NameCheap: $110

  • Cloudflare premium, although not required, improves image load time significantly via caching with their global CDN: $20

  • Image generation: 1,800,000 images, 10,000 generated per hour with a batch size of 16 using a g4dn.xlarge instance, which has a single V100 GPU with 16GB of VRAM at $0.526 per hour on-demand: $95

  • “Accidentally” hosting the website from AWS for the first day, resulting in over 10TB of non-cached bandwidth charges: >$1,000 (via credits)
An image of what my desktop looks like while generating an additional million anime images
What the access log for nginx looks like while serving >1,000 anime images per second
Yet another example of why to not use AWS for high-bandwidth content: AWS has some of the most expensive bandwidth at $0.09 per GB (luckily this was paid for entirely with credits and the migration off of AWS was complete in less than a day to a more sustainable provider)

Conclusions and the future

All in all, this was a fun project to put together and host, and I’m glad that hundreds of thousands of people have visited it, discussed it, and enjoyed this novel style of artwork.

If you want to read in-depth about the ML behind this model and everything related to it, please read Gwern’s post.

Thank you to Aydao for relentlessly improving Stylegan2 and training on danbooru2019 until these results were achieved as well as releasing the model for anyone to use, Obormot for the base javascript library used, TXDNE.js, Gwern for producing high-quality writeups, releasing the danbooru datasets, and and other members of Tensorfork including Gwern, arfa, shawwn, Skyli0n, Cory li, and more.

The field of AI artwork, content generation, and anything related is moving very quickly and I expect these results to be surpassed before the year is over, possibly even within a few months.

If you have a technical background and are looking for an area to specialize in, I cannot emphasize the extent that I’d strongly suggest machine learning/artificial intelligence: it will have the largest impact, it will affect the most fields, it will help the most people, it will pay the most, and it will cause you to be surrounded by the best and smartest people you could hope for.

Thanks for reading and I hope you enjoyed the website! For more about myself, feel free to read my about page or see my Twitter.

Bonus: additional image compilations from Gwern