Employing the suggestive metaphor of neuroscientist Luiz Pessoa, David Brooks, in the New York Times, describes neural processing as “a flock of swirling starlings,” otherwise known as a murmuration. That metaphor alludes to the continuously changing patterns of activation across the plethora of complex, interconnected neuronal networks that populate both the cerebral cortex and the brain’s various subcortical structures. Such networks can contain anywhere from dozens to many tens of thousands of neurons. Those networks integrate inputs about:

• the current environment in which the person is situated and through which the person is navigating,
• the recent and current states of the person’s body and motor activities, and
• the host of the person’s mostly unconscious but potentially relevant cognitive and affective states (memories, emotions, desires, inferences, decisions, etc.).

These networks are also constantly forwarding the results of their integrative activities with all of that information on to other networks with which they are directly connected, both downstream and usually upstream—*as feedback—*as well. Such endlessly changing patterns of activation are just like, for the swirling flock, the collective consequences of the many movements of the individual starlings, who are primarily influenced by the motions of the six or seven other starlings closest to them.

It is worth noting that it is precisely the computational modeling four decades ago of these dynamics of neural networks that provided the foundational insights, tools, and techniques that have spawned the large language models that drive the new AI systems that have attracted so much attention over the past few years.

The Bucket Theory of Mind

Brooks contrasts this conception of the mind as swirling patterns of activation across scores of neural networks with the model that dominates so many of our educational efforts, namely, that students’ minds are empty vats that need to be filled with information.

Brooks’ observation is reminiscent of Karl Popper’s discussion of “the bucket theory of mind” in his book *Objective Knowledge. *Popper proposes that our commonsense conception of mind holds that at birth, the mind is like an empty bucket. On the basis of our sense experience, including what we hear and see people say, the bucket slowly gets filled with information. According to this commonsense account, our knowledge grows by accumulating ever more information in this fashion and reassembling it in new ways. One corollary of this picture of the mind is that education is primarily about getting as much good information into those (thoroughly similar) buckets as possible.

Leaky Buckets

For a host of reasons, Popper thinks that the bucket theory is “thoroughly mistaken.” The first reason he cites is the bucket theory’s assumption that learning how to get by in the world turns on gathering particular bits of sensory-based information in our buckets. Popper regards this entire story as philosophers’ contrivances. Instead, Popper suggests that the world relentlessly assails us with “chaotic messages” and that it is our emerging ability to ignore most of them and to detect and decode the important ones, especially the biologically important ones, which enables us to manage in the world. He holds that “innate dispositions” direct the detection and that trial and error learning informs the decoding. That we are so adept at both under most circumstances Popper attributes to our evolutionary heritage, which has given rise to “our incredible efficiency as biological systems.”

The bucket theory’s educational implications are no less questionable. It focuses on the content instead of the learners (as if they are all the same), and it emphasizes and tests the storage of informational bits, instead of exploring how they fit together, why they matter, and what might be done with them.

This is not to say that the command of facts does not matter. Inquirers need some appreciation of the basic assumptions that shape the discussions and controversies in any field, but they are better recollected and most promisingly possessed when they fit into richly integrated and articulated intellectual constructs. Otherwise, our mental buckets typically prove shockingly leaky. Many of those memorized names, dates, and facts are gone two weeks after the big test, and most, if not all, are gone a year later. Education cannot be primarily about their retention.

Brooks correctly notes that in a complicated world, a brain that is “able to improvise a vast number of networked ensembles that would dynamically affiliate and thus coordinate sensible responses” is far preferable to a vat-brain piled with items of information to be stored, conjured up, and reviewed.

References

Bechtel, W. and Abrahamsen, A. (2002). *Connectionism and the Mind: Parallel Processing, Dynamics, and Evolution in Networks *(Second Edition). Oxford: Wiley-Blackwell.

Popper, K. (1972). Objective Knowledge: An Evolutionary Approach. Oxford: Oxford University Press.

Rumelhart, D. and McClelland, J. L. (1986). Parallel Distributed Processing, Explorations in the Microstructure of Cognition, Volume 1: Foundations. Cambridge: The MIT Press.