Geometry and affect

I’m reprising some older blog posts from 2013 that consider vertigo, oblivion, melancholy, the motion of swings, and the emotional experience of urban spaces. Later on I drew this material together in my book Mood and Mobility (2016).

• 132. Swinging
• 133. Oblivion
• 134. The brain in the city
• 135. Well-being and geometry
• 136. Your inner child
• 137. The melancholy medium

With assistance from ChatGPT I’ve updated the sentiments of these posts under the rubric of geometry. I also attach the posts as audio cloned from my own voice (including the strangely hybrid aussie-UK accent and glitches).

Two by two

In 2013, I wrote about Galen, Vitruvius, and the classical inclination to interpret wellbeing through geometry. The ancients divided the world into fours — seasons, humours, temperaments, compass points — so that health and mood could be pictured as a matter of balance on a grid or within a circle. Geometry offered a way to make sense of states of mind. Even melancholy, sanguinity, and phlegmatic calm could be located in relation to one another, with the body framed as a system of intersecting axes.

I was also interested in other geometries and their associations with mood: the pendulum swing of ritual motion, the oscillations of brain waves measured by EEG, the arcs traced by animated pandas and dragons as they swept through the space of the cinema screen. These arcs, circles, grids and fourfold schemas placed emotion within a patterned order, as if affect could be understood when given a geometry.

Affective AI

In the age of AI, that geometrical impulse has intensified. Machine learning represents emotion as vectors in multidimensional space. Words, images, and voices are embedded as coordinates that can be measured, clustered, and interpolated.

When scholars of affect speak of “sentiment analysis,” they are really talking about geometries: distances between “joy” and “sorrow,” alignments of “frustration” with “engagement,” positions plotted within the emotional topographies of language models.

Cities are also overlaid with such geometries. We experimented with mobile EEG headsets in Edinburgh, mapping supposed emotional states as people walked through green space and busy streets.

It seems that the “affective AI” performs of today frame emotions via similar cartographies at scale, reading signals from cameras, sensors, and social media to generate mood-maps of whole populations. Urban geometry and emotional geometry collapse into one another, conjecturing new diagrams of wellbeing, anxiety, and attention.

Affective clone

Even the audio recording of my cloned voice that accompanies this post is part of this geometrisation of affect. Platforms such as ElevenLabs.com analyse pitch, cadence, and rhythm as numbers and curves before being reassembled into the semblance of (simulated) emotional engagement. Emotion becomes performance through calculation.

I’ll put ChatGPT’s conclusion in quotes as I’m not sure I could have produced such a cosy summary.

“Geometry and emotion are not such distant partners after all. From Galen’s fourfold humours to today’s multidimensional embeddings, we seem compelled to draw emotion into ordered shapes and patterned spaces. AI simply continues that tradition, turning the inchoate quality of feeling into points, vectors, and surfaces. Perhaps this helps explain both the allure and unease of our encounters with ‘intelligent’ machines: they invite us into geometries of emotion that feel at once strangely ancient and startlingly new.”

Note

• Featured image is by ChatGPT: Generate an image of Gallen’s ancient dusty lab bench including symbols of the humours..