Hacker News new | past | comments | ask | show | jobs | submit login
Why spiky shapes seem angry and round sounds are calming (bps.org.uk)
74 points by hhs 54 days ago | hide | past | web | favorite | 21 comments

'we automatically associate variations in one particular property of images or sounds with variations in levels of emotional arousal.'

The future is going to be great, the implications of these findings seem obvious: auto-tune for my voice is coming so that people think I am thrilled that they called.

To monetize it, they can sell it to contact centres so that sales reps sound super engaged and interesting, and customer support reps seem genuinely concerned.

> they can sell it to contact centres so that sales reps sound super engaged and interesting, and customer support reps seem genuinely concerned.

This sounds so dystopian!

It might even be baked into some phones by default, kind of like image softening to remove skin blemishes is the default post-processing for most pictures in most phones today.

There’s help available, if not for you, for your children: https://books.google.com/books?id=R6E0AAAAMAAJ&pg=PA74&lpg=P...

That should be a fairly straightforward application of the same techniques that transfer style between images.

Seems similar to the Bouba/Kiki effect [1] where we automatically associate spiky and smooth shapes with certain names.

I did this test with my toddler and he chose “correctly.”

[1] https://en.m.wikipedia.org/wiki/Bouba/kiki_effect

They mention this in the article:

This process could also account for the classic “cross-modal” finding, first reported in 1929 and replicated around the world many times since, that people tend to pair a blobby shape with a word-sound like “bouba” and a spiky shape with a word like “kiki”.

I think a possible alternate answer to why these associations exist would be how the brain recognizes and utilizes patterns. The human mind appears inclined to find patterns in stimuli and form conclusions from those patterns. The inordinate amount of brain real estate that is lit up during facial recognition is an example, as well as our tendency to hallucinate fake stimuli in sensory deprivation scenarios after a time. Shrill, staccato noises and their counterpart pointed, sharp images represent sudden changes in stimulus. Unless those changes are regular, it tends to disrupt the brain's ability to recognize the changes as a pattern e.g. a pop song, with often tread chord progressions and rhythms tends to get stuck in your head much more easily than say a Rush song, with odd time signatures and discordant melodies.

If you draw out these patterns, or lack thereof, as a graph chronologically, be it changes in pitch, rhythm, timbre of a sound/music, or the characteristics of drawn figures, such as change of the direction of a line over distance, dramatic and frequent changes would make it harder to interpolate what would happen if we kept the graph going. I postulate that the brain associates regular, "round" patterns with happy or content emotions because it too tries to interpolate the future state of these stimuli, and uncertainty about that future gets associated with "negative" emotions that are rooted in the amygdala and are tied to more survival-centric emotions such as anger and fear i.e. fight or flight.

This might be total bunk, but that's the 2 cents from a non-psychologist :)

Thanks for the comment. Thoughts not dissimilar from yours are what drove us to undertake these studies. :)

I definitely suspect that predictive processing in the brain is part of the story, and we discuss that a bit in the paper. Interestingly, local entropy is highly correlated with our "spikiness" measure, the spectral centroid. However, where your comment focuses on macro-level regularity, like Bach versus Rush, our study focused on micro-level texture, like the hum of a box fan versus squealing brake pads. Although macro-level regularity definitely has an impact on emotion perception as well! This is something my other research has touched on, e.g.: https://www.pnas.org/content/110/1/70

We also found that "spikiness" predicted emotional arousal for both positive and negative emotions. E.g., "angry" and "excited" are both spiky across the senses, while "sad" and "peaceful" are both smooth. So the results can't be driven solely by uncertainty becoming associated with negative emotion.

Thanks for your thoughts! And here's a link to a preprint of the paper: https://psyarxiv.com/wucs4

This is a great observation, and lines up well with something I've been thinking about recently. Some HN posters may be familiar with ASMR; if not, https://en.wikipedia.org/wiki/Autonomous_sensory_meridian_re... is a reasonable jumping off point. I'll assume basic familiarity with the idea of "triggers" - primarily auditory, these are actions / sounds that cause the tingly ASMR effect. Some triggers seem to work on really large groups of people (e.g. mouth sounds, repeated words).

From an audio standpoint, triggers often seem to be highly correlated with audio transients ( https://en.wikipedia.org/wiki/Transient_(acoustics) ) - quick little sounds with a lot of detail. From an information-science standpoint, transient sounds convey more information in a short time period than does, say, a uniform sine wave, or music, or even talking.

There's a lot happening there, and it's not very predictable; but at the same time, the entire setting of the content is generally predicated upon relaxation. It's possible that the triggering sounds are stimulating that kind of negative / fight-or-flight / sudden-action-required sort of interrupt you're describing. The difference is that because of the context, the listener can immediately return back to feeling safe and comfortable again after the transient is past.

I have personally noticed that the most triggering kind of content for myself is content where I can _almost_ predict what's going to happen, but when it surprises me slightly. Repeated trigger words work really well for this - by the 3rd repetition my brain has an excellent profile of how the ASMR practitioner is going to pronounce a word, but there are always slight differences, unexpected timing, etc. that go into it. I'm not yet sure whether the trigger happens more on a prediction hit or miss; leaning towards a hit at this point.

Spiky shapes are like a dog showing its teeth or the fur standing up along its back. It's how animals and even plants get you to keep your distance.

I figure this has to be it. In nature spikes are almost always dangerous, whereas blobs are almost always relatively safe objects (typically soft, et cet).

Markings on snakes are one of the more prominent examples. I don't have a source, but I've read that human eyes are very well tuned to spot them- an evolutionary advantage from when we were living in forests?

They say "Here's why", but they never explain why. They just found a correlation in what people think. No reason as to why we think that way...

This could even potentially be just from societal conditioning.

In the paper we do take a shot at explaining why. Specifically, we draw from the literature on cognitive ethology to argue that the ability to quickly and reliably read out emotion content from low-level stimulus features provides better prediction of others' behavior, and therefore an evolutionary selection advantage.

We don't rule out learning processes (e.g., social conditioning), but we do examine evidence for some level of innateness. For example, cross-modal Bouba–Kiki-like effects are present in pre-linguistic infants, and arousal signals can be easily understood across species. It may end up that what is innate is a predisposition to track and learn cross-modal correspondences that are widespread in the environment.

Link to paper pre-print: https://psyarxiv.com/wucs4

That is an extremely common complaint about this kind of study/argument/proposal, and honestly no less valid because of it.

So cool to see this hitting the front page of Hacker News! I've known Beau since we were in high school.

He occasionally posts about his research on twitter - https://twitter.com/beausievers

I am first author on this paper; glad to see it getting discussed on Hacker News.

Here's a free, non-paywalled link to a pre-print version of the paper: https://psyarxiv.com/wucs4

And here's a paywalled link to the published version: https://royalsocietypublishing.org/doi/abs/10.1098/rspb.2019...

Imagine something spiky hitting your face. It makes you angry. Then you see another spiky. You have empathy with your past self, and project it onto the thing. The new spiky is angry.

surprised nobody else here mentioend this , but perhaps this is just a very common form of synesthesia?

It would be interesting to see how different materials factor into these perceptions. They mention the straight lines and sharp edges of brutalist architecture, but if constructed out of wood instead of concrete, for example, I would expect the forms to be perceived as softer and more organic.

Spiky things hurt and round things don’t — would be interesting to simulate pleasure / pain of shapes against a virtual hand using a simulation then train a deep net to learn the pain response. Bet you would have similar findings.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact