Coherence, Craft, and Creativity

Some thoughts on Hans Otto Storm's "Eolithism and design."

Evgeny Morozov’s “The AI We Deserve”, published in Boston Review, offers a genuinely compelling positive leftist vision for artificial intelligence not steeped either in Silicon Valley tropes or in their mindless inversion. He argues that, even though the shape of AI systems is inseparable from the machine dreams of bureaucratic rationality, we have the freedom of using them for playful experimentation detached from specific quantified goals, transform AI into what David Graeber called poetic technology in The Utopia of Rules — “that is, one where mechanical forms of organization, usually military in their ultimate inspiration, can be marshaled to the realization of impossible visions: to create cities out of nothing, scale the heavens, make the desert bloom.” The shape of AI systems is a constraint that deconstrains: We could use them to write 10x more boilerplate code in 10x less time, or we could stage the next information revolution.

Reading Morozov’s essay, I learned about Hans Otto Storm. A novelist and a radio engineer, Storm died in 1941, just days after the attack on Pearl Harbor, from an electrocution while installing a power transformer for the Army Signal Corps in San Francisco. He was friends with Thorstein Veblen, whose ideas on the role of “idle curiosity” in scientific discovery had influenced his thinking about the art and craft of engineering. Storm expounded some of his ideas in an essay titled “Eolithism and design,” which was published posthumously in 1953 in the Colorado Quarterly. In a way, this essay also gestures at something like poetic vs. bureaucratic technologies.

Storm wants to distinguish design, which he understands in the Weberian sense of rational action as matching available means to desired ends, from more organic, ecological modes of artifact-making, which he terms eolithic in reference to Stone Age eoliths, or “stones picked up and used by man, and even fashioned a little for his use.” The key difference between design and eolithism is this:

The important item of the definition from the point of view of method of craftsmanship, and the one which distinguishes the eolithic method fundamentally from that of design, is that the stones were picked up - picked up, that is to say, in a form already tolerably well adapted to the end in view and, more important, strongly suggestive of the end in view. We may imagine that person whom the anthropologists describe so formidably by the name of man strolling along in the stonefield, fed, contented, thinking preferably about nothing at all - for these are the conditions favorable to the art - when his eye lights by chance upon a stone just possibly suitable for a spearhead. That instant the project of the spear originates; the stone is picked up; the spear is, to use a modern term, in manufacture.

In other words, design is a rational activity; as Herbert Simon put it in The Sciences of the Artificial, “everyone designs who devises courses of action aimed at changing existing situations into preferred ones.” Here, the mention of preference is key, it is meant to indicate their temporal precedence to the process of design. A designer operates in what Jimmie Savage would call a small world, a setting where all possible states of affairs, contingencies, and consequences of various acts are codified and where a preference relation is given among the different consequences. It is a world where we can look before we leap, so an engineer can make use of simulation, prototyping, and numerical optimization to bring about the desired ends given the available means. By contrast, a large world is one where certain contingencies can only be encountered as a result of active exploration, rather than anticipated and accounted for in a theoretical model.

Storm, writing some thirty years before Simon, points out a similar distinction. Design operates in small worlds,

in those cases where the materials are uniform, and where there is considerable latitude for properly controlled experiment, so that the craftsman may have frequently impressed on his mind not only what happens according to the theory, but also what else may happen when the theory is ignored. Theory, experiment, and generalization become rather useless ornaments when applied to cases which do not in the nature of things repeat themselves, such as that of the biologist with twins, who had one twin baptized and saved the other as experimental control by which to estimate the effect.

By contrast, the eolithic mode is that of an encounter between (as Morozov nicely puts it) a Stone Age flâneur and an open world. There are no prefigured ends, values, or preferences; rather, they emerge spontaneously from this interaction, a particular affordance selected from among a multitude by the mind’s eye and a fortuitous coupling between an agent and an environment. There is either no theory guiding this interaction or, if there is one, it may be literally false. As Storm says,

I have known sailors who thought the tides came at the same hour every day, and mountaineers, outdoor people, who did not know that the fixed stars rose and set. Both of these specimens did from time to time order their movements by these wrongly apprehended natural phenomena, and both eventually muddled through - so that not only they but their theories with them survived into their ripe old age. … Their theories were not too wrong - after all, there were tides in the sea, and there were stars in the sky which did not move around very much. And the habit of being aware of these bulky natural verities brought in along with it an easy awareness of a multitude of other things individually quite below the threshold of classification, so that on the whole these people were workable, operating personalities who could get along in the world the way they found it, and whom one could depend on under the circumstances where one found them.

Storm’s concepts of eolithism and design could be mapped onto what James C. Scott called métis (local, practical knowledge acquired through experience) as distinguished from techne (rational knowledge put to work in the realm of uniform material and plans). This mapping is not entirely one-to-one because, unlike the eolithic craft, metis still presupposes goals that are stated beforehand. What is more interesting, though, is what eolithism and design have in common — the character of operational coherence.

In the context of scientific and engineering inquiry, this notion lies at the core of Hasok Chang’s latest book, Realism for Realistic People: A New Pragmatist Philosophy of Science. According to Chang,

operational coherence consists in aim-oriented coordination. A coherent activity is one that is well designed for the achievement of its aim, even though it cannot be expected to be successful in each and every instance. Operational coherence is based on pragmatic understanding; it consists in doing what makes sense to do in specific situations of purposive action

(emphasis in the original). Note that, even though Chang insists on talking about “design,” “aims,” and “purposes,” he does not stipulate their origin. Aims and purposes could be stated and codified ahead of time or they could be emerging spontaneously in the course of one’s interaction with the world. That makes both design and eolithism coherent. The stipulation of no guaranteed success is quite important here as well — think of Storm’s example of sailors or mountaineers achieving their aims despite holding theories that are literally false. Elsewhere in the book, Chang wants to emphasize that coherence is a much broader concept than logical consistency; it is context-dependent, open-ended, and provisional. Coherence can be gained and it can just as easily be lost.

Although Chang makes no mention of it, the notion of coherence is central to process ontologies that emphasize events rather than things, becoming rather than being. For example, Schopenhauer writes the following in the introduction to the first edition of The World as Will and Representation:

A system of thought must always have an architectonic connection or coherence, that is, a connection in which one part always supports the other, though the latter does not support the former, in which ultimately the foundation supports all the rest without being supported by it, and the apex is supported without supporting. On the other hand, a single thought, however comprehensive it may be, must preserve the most perfect unity. If it admits of being broken up into parts to facilitate its communication, the connection of these parts must yet be organic, i.e., it must be a connection in which every part supports the whole just as much as it is supported by it, a connection in which there is no first and no last, in which the whole thought gains distinctness through every part, and even the smallest part cannot be completely understood unless the whole has already been grasped.

Alfred North Whitehead’s Process and Reality defines it as follows:

’Coherence,’ as here employed, means that the fundamental ideas, in terms of which the scheme is developed, presuppose each other so that in isolation they are meaningless. This requirement does not mean that they are definable in terms of each other; it means that what is indefinable in one such notion cannot be abstracted from its relevance to the other notions. It is the ideal of speculative philosophy that its fundamental notions shall not seem capable of abstraction from each other. In other words, it is presupposed that no entity can be conceived in complete abstraction from the system of the universe, and that it is the business of speculative philosophy to exhibit this truth. This character is its coherence.

In the context of Storm’s distinction between design and eolithism, coherence can either be imposed and maintained in a top-down fashion or it can arise in a bottom-up way through a reciprocal interaction between a subject and an environment. Both bureaucratic and poetic technologies must be operationally coherent; we can apprehend technology as both the locus of instrumental rationality and as a crucial enabler of open-ended exploration driven by idle curiosity and play. We can either passively wallow in Heideggerian resignation to technology as enframing of the world as standing-reserve, or we can follow Whitehead and reconfigure our attitude toward engineering as “creative advance into novelty.” As he wrote in The Concept of Nature,

For natural philosophy everything perceived is in nature. We may not pick up and choose. For us the red glow of the sunset should be as much part of nature as are the molecules and electric waves by which men of science would explain the phenomenon.

This should be the motto of every poetic technologist.