Ed Yong’s An Immense World: How animal senses reveal the hidden realms around us describes the many ways creatures perceive their surroundings through their very different sensory apparatus. It made me wonder about the subjective perceptual universe of ogregores.
According to the publisher’s blurb, in the book we “encounter beetles that are drawn to fires, turtles that can track the Earth’s magnetic fields, fish that fill rivers with electrical messages, and even humans who wield sonar like bats.”
In a New Scientist review, Anna Demming notes that Yong uses the word umwelt to describe this world. He borrowed the term from Baltic-German zoologist Jakob von Uexküll, who used it to describe not just an animal’s environment, but also its experience of that environment. In a New Scientist story about biosemiotics, Von Uexküll is quoted thus: “If we stand before a meadow covered with flowers, full of buzzing bees, fluttering butterflies, darting dragonflies, grasshoppers jumping over blades of grass, mice scurrying, and snails crawling about, we would be inclined to ask ourselves the unintended question: Does the meadow present the same view to the eyes of so many various animals as it does to ours?”
I’ll focus on for-profit companies to explore what ogregores might perceive. (Ogregores, aka orgregores, are groups of people which have, or are seen to have, individuality and agency over and above the means and motives of their human constituents.) All companies collect data on costs and sales. Large ones generate and collect information from many sources in many places. They collect information on both stocks and flows. As data flows up the org chart it may be summarized and compressed, for example to be presented in business dashboards. Data values fluctuate, ebbing and flowing with economic cycles and the seasons (e.g., the volume of “grill” vs “chill” fast food sales changing from winter to summer). Businesses have a first-line view of data like orders, inventories, and hours worked that are aggregated to create economic indicators.
Companies not only observe their internal processes and customer interactions, but also their competitors. I guess such information is less automated and less private since it relates to behavior that is largely public. However, there are emerging scenarios such as satellite earth observation (NSR research report; NYTimes) where data gatherers are selling intelligence not just to government and military clients but also private sector customers like financial and insurance companies.
Digital technology increased the volume and increased the diversity of data, with even traditional companies using it to track user attitudes, such as sentiment analysis of customer interactions and social media posts. Digital tech businesses such as social media and behavioral advertising more broadly, can observe real-time nuances in user engagement (what topics are trending, what’s resonating with specific user types) and ad sales (what keywords are selling best, which parts of the market are busy or slow).
Companies obviously have rich representations of their surroundings. I’m agnostic about whether they have a subjective experience of these representations. (This is a distinction between defining perception as requiring subjective experience versus merely requiring representation.) I can’t adduce good evidence that they do, but I would like it to be true. For the purposes of this post it’s sufficient that ogregores have representations of the world that are different to the ones individual humans do.
Most if not all the conventional business data described above is perceptible by humans: it’s collected by them, and summaries of aggregates are presented to managers. I suspect that something qualitative changes for very large volumes of such data, but it’s not as if we’re talking about sensory modalities (like electric or magnetic fields) that are inaccessible to humans.
Digital tech changes things. On a mundane level, I doubt humans have effective sensory access to the real-time dynamics of adword auctions and algorithmic stock trading. Too much is happening too fast for it to be perceptible. The same is true for automated cyber intrusion detection systems (IDS) that monitor a network or systems for malicious activity or policy violations.
The parameters of neural networks are even less accessible. The algorithms that drive social media experiences and retail offerings represent the world in terms of “weights,” equation coefficients that determine a network’s output. As neural nets are trained and evolve, the weights change. The number of weights can be very large; the GPT-3 language model debuted with 175 billion parameters (Venturebeat; StackExchange; a quick search didn’t yield any information on the number of weights used by social media algorithms). It is very hard, if not impossible, for humans to make sense of the weights.
Ed Yong’s book implies that different animals invest in different kinds of sensory apparatus. Different companies seem to do so as well. A brick-and-mortar retailer doesn’t track customer engagement in the same way as a social media platform, and a physical production line tracks different indicators to a software development chain. Berkshire Hathaway has a very different time horizon from an algorithmic trader. I imagine that these organizations have different processing and representation systems (both human and technical) which will affect their umwelt.
Kalevi Kull at the University of Tartu in Estonia, quoted in the New Scientist story about biosemiotics, takes a line similar to Von Uexküll which points a way forward for ogregores.
“Biology has studied how organisms and living communities are built. But it is no less important to understand what such living systems know, in a broad sense; that is, what they remember (what agent-object sign relations are biologically preserved), what they recognise (what distinctions they are capable and not capable of), what signs they explore (how they communicate, make meanings and use signs) and so on. These questions are all about how different living systems perceive the world, how they model the world, what experience motivates what actions, based on those perceptions.”
What do ogregores know, remember, and recognize?
No comments:
Post a Comment