Following are some of my reflections on the first week of JF Martel’s Art and Contemplation course, combined with some thoughts that have been percolating for some time. I am going to italicize many words for emphasis — fair warning for those of you who shriek at that sort of thing.
Martel makes a distinction between art and artifice that is somewhat peculiar to his context, wherein both are aesthetic artifacts (in the sense of being artificial, made by humans) intended to have some kind of affect (or affective effect) on the viewer, but whereas artifice does so to effect a kind of magical glamor on the viewer (usually to control them), art points to Mystery or the Real in ways that create rifts in ordinary experience — ordinary experience being often dominated by the enchanting and entrancing spell of artifice. (Here I am remembering a scene from the film Johnny Mnemonic, when the Lo-Teks hack the television signal so their leader J-Bone can briefly interrupt the news stream with his message, “Snatch back your brain, zombie!”) N.B., there are, broadly speaking, two, contradistinct kinds of magic being indicated here: the stuff of rifts, liminal spaces, initiations, rites of passage, empowerment, agency, &c.; and the stuff of glamor and enchantment to hypnotize and control. There is much to unpack in all this but I am only going to get at a few bits pertaining to some thoughts I have about artificial intelligence and real magic.
Among my collection of relatively obscure books about cybernetics and related subjects is Richard R. Landers’ Man’s Place in the Dybosphere, published by Prentice-Hall in 1966 and given a second printing in 1967. Landers (1919–2017) coined the word ‘dybosphere’ from the Hebrew word dybbuk, a malicious spirit that clings to a human host (דיבוק is derived from דָּבַק, “adhere” or “cling”); to denote “a world of mechanized men and humanized machines,” a super-mechanical world similar to something Philip K. Dick once observed (“The Android and the Human”):
[O]ur environment, and I mean our man-made world of machines, artificial constructs, computers, electronic systems, interlinking homeostatic components — all of this is in fact beginning more and more to possess what the earnest psychologists fear the primitive sees in his environment: animation. In a very real sense our environment is becoming alive, or at least quasi-alive, and in ways specifically and fundamentally analogous to ourselves.
Today, we are surrounded by so-called “smart” devices possessed of embedded computers that coordinate interactions with sensors and actuators or effectors in feedback circles, granting these artifacts a kind of intelligence or lifelikeness. Such artifacts are intelligent in the sense proposed by W. Ross Ashby (An Introduction to Cybernetics): they are (more or less) capable of appropriate selection, and the more appropriate selections they can make the more intelligent they appear to be and in some sense genuinely are — a sense quite like how a living organism appears to be intelligently designed because natural selection has awarded it such traits as are well adapted (i.e., appropriate) to its niche, thus allowing it to survive and thrive in ways that appear to be downright clever even if they do not involve any cognition or deliberation per se on the part of the organism. But while many “smart” artifacts are cleverly designed, they are also notoriously unintelligent in sometimes humorous or even dangerous ways when confronted with perturbations of greater variety than they were made to compensate.
Back to Martel: he lists five qualities that art (vis-à-vis artifice) possesses; it is irreplaceable, integral, important, interesting, and intelligent. These are, not coincidentally, also qualities of persons, and connected to Martin Buber’s distinction between I-It and I-Thou relations (I and Thou). Martel suggests approaching art as a Thou rather than an It — as a person rather than a thing — which we do in magic to animate a thing (an it) such as a stone or statue. This evokes the two sides of idolatry: on one hand, the fetish is a genuine material or physical abode, manifestation, or expression of the spirit or spiritual; on the other, the fetish diverts our gaze away from the spirit, keeping us fixated on the symbol such that we ignore its referent, like the finger pointing at the moon (Śūraṅgama Sūtra).
Here I want to mention Heinz von Foerster’s notion of Trivial vs. Non-Trivial Machines. A TM embodies a simple function that transforms one thing or state into another — like a toaster transforms bread into toasted bread — and it is (theoretically, in principle) invariant: every time you put in bread and push down the switch, the machine outputs toast. A NTM, by comparison, changes its function, or what state is transforms to, depending on its previous state. As Frank Chimero said (“The Shape of Design”), if you ask me if I want pancakes for dinner tonight then I will enthusiastically reply yes, but if you ask me the same question tomorrow night, I would not be so enthusiastic. People do not do the same thing over and again like machines — except when they do, and we remark they are behaving like a machine; and what they do usually depends on what they have done previously. We can make a machine with memory and feedback, but does that really make the machine non-trivial? And we can treat a person like a machine, or in other words we can trivialize people — usually by denying them one or more of the five qualities that Martel ascribed to art! (Here I am remembering how humans pejoratively called the Cylons toasters in the Battlestar Galactica 2004 series; the word connoting, “you are nothing but a [trivial] machine.” This is of course a well-known tactic in warfare: reducing your enemy to something less than human, devoid of personhood.)
Clearly, artificial intelligence of the sort that people fancy when they think of Droids, Transformers, Replicants, &c. — “strong” AI — would need to be responsive to various stimuli, meaning it would need to be capable of varying its responses. i.e., it would need to be dynamic. Whether or not it would be trivial depends on where we draw the line of non/triviality, but it should be obvious that the more complex and lifelike AI becomes, the more it calls us to relate to it as a Thou rather than an It. I suppose the reason many folks wince at those videos of people hitting and pushing around the Boston Dynamics robots is because they are witnessing something that in many ways looks and acts like a Thou, being treated like an It (I would much prefer to watch them dance).
Meanwhile (it is always meantime somewhere — somewhen?), my way of making machines non-trivial is to treat them animistically: with magic.
The magical art that I call “robomancy” is chiefly concerned with the creation of a creature, viz., the Robot of Art (a nod to the Triangle of Art in magical evocation), and regardless of the robot’s mechanical complexity, the “magic” begins with approaching the Robot as a Thou rather than an It. It is an animistic, daemonic process of non-trivialization — kind of ironically, since demonizing people can be synonymous with trivializing them. What differentiates the Robot of Art from other occult artifacts is precisely that it is dynamic in the physical space (compare kinetic art), and the changes it undergoes are symbolic in the same way that ritual actions and magical icons and idols and so on are symbolic — but icon and idol are typically static in the physical space (at least in an ordinary sense) although they may be magically or spiritually dynamic. The Robot of Art, like all art, transcends utility (this is another quality of art that Martel talked about), and especially the utility we typically associate with robots. Of course, sorcery and so-called “low” magic have a strong utilitarian aspect to them, which is also part of robomancy; but if sorcery is the trivial machine of magic, inasmuch as it is magic it remains a non-trivial trivial machine in the sense that Jesper Sørensen said, “Magic is about changing the state or essence of persons, objects, acts and events through certain special and non-trivial kinds of actions with opaque causal mediation” (A Cognitive Theory of Magic, my emphasis).
The dybosphere of “smart” devices surrounding us today are usually more artifice than art. They may fascinate us with their flashing and beeping charms, but they merely pretend at being intelligent or alive, quite mockingly sometimes. By contrast, the Robot of Art plays at being numinous, along aesthetic (or perhaps aesthetico-functional) dimensions; and it is actually “cyber-animistic” in the sense that Dukes suggested and I suspect Martel is also getting at. As Dukes said in Words Made Flesh:
Instead of deducing that we are 'nothing but' machines, let us increase the mechanical world to embrace mystery. The idea of a computer which suffers from a nervous breakdown, or a robot which becomes a born-again christian, does not threaten us — it amuses us. It does so because it suggests machines growing into areas of unpredictability and craziness that were once beyond their reach — machines becoming 'unmechanical'. But the idea of a computer program which could precisely predict your next nervous breakdown would be a threat, because it would reduce that area of craziness to mechanical predictability. If I create a machine that can full-bloodedly suffer the agonies of falling in love, then I have done something wonderful: I have revealed 'soul' in the motions of matter. If consciousness is totally the byproduct of chemical reactions, then the rudiments of consciousness lie all about us: I can once more say that a flower turns toward the sun because it 'loves' the sun. The universe is now a living being. If, however, someone else chooses to take the same evidence to prove instead that 'love is nothing but a chemical reaction', then I feel sorry for that person in his smaller world. May this book offer healing.
Or, mechanism without reductionism.