Alchemy, Mythology, and Artificial Intelligence

Podcast: The Ancient Roots of Transhumanist Thinking with Lenn E. Goodman

But was the word God’s correlative or was it implicitly divine? Emanation was long symbolized as the flow of water from a spring or light from the sun. But Proclus compared it to the flow of theorems from axioms. Theistic voluntarists made emanation an act of grace, God’s word eternal, coextensive with His wisdom. But the world, for them, was contingent and of finite duration.

Matter, Neoplatonically, was pure receptivity, with no being of its own. For being, to Aristotelians, meant determinacy: What is real should be a this and a such. Matter makes a thing a this—that is, a particular—but only Form could make a such. Yet matter, as defenders of miracles would insist, might receive any form. Hadn’t Aristotle himself explained the perpetuity of the cosmos by appeal to matter’s eternal recycling, the elements ever transformed by one another? For homo faber, the appeal was irresistible: to share in divine creativity by shifting forms in matter. For if essences are gifts of the star souls that steer celestial bodies, imparting terrestrial natures (and inspiring prophetic insights), the alchemist had but to harness and direct those energies.

That project long antedates its Neoplatonic rationale. It harks back, as Paul Kraus showed, to ancient procedures for consecrating idols.¹ The artisans who modeled human or animal forms as objects of veneration well knew the natures of their materials, limits trenchantly spelled out by the prophets of Israel:

Idol makers—all dross! Worthless their pride! Blind witnesses, ignorant, too ashamed to confess it. Who could make a god? Cast a helpless icon! All its devotees, embarrassed; the craftsmen, struck dumb, stand together facing their shame. The smith wields his iron, hammers at the forge with brawny arm, shaping his work, weak with hunger, fainting for a drink. The carver draws taut his line, marks a shape in chalk, carves it with his chisel, scribes it with his compass, makes a human figure, a handsome man, to dwell within a shrine. He cut the cedars, cypresses, or oaks, among the trees he grew in the woods, the firs he planted, watered by the rain, to use for firewood. Some he takes for heat, builds a fire, bakes bread. And of some he makes a god to worship, a graven image that he kneels to. Half he burns and eats the meat he’s roasted. Warm and sated, he says, “Ah, the heat feels good! What a fire!” The rest he makes a god—handcarved!—kneels and bows and prays to it: “Save me! You are my god!” (Isaiah 44:9–17).

How could anyone, Isaiah reasons, be so naive? Only social expectations sustain the pretense. Maimonides will agree: “No one ever imagined or ever will, that the form he casts in metal, or works in wood or stone, created and rules heaven and earth. Idols are served only as symbols of some thing that mediates between us and God.”² The rationale is well-known from Porphyry. But an alternative reflects centuries of pagan piety. Akkadian and Sumerian texts from the first millennium BCE neo-Babylonian and neo-Assyrian libraries describe the activation of a divine image (śalmu, cf. Hebrew tzelem) by ceremonies of mouth-opening and mouth-cleansing (pīt pî and mīs pî). The references run back to the third millennium BCE.³ The object was to invoke the god to be “born” into the image. In some versions the god invests the carving as soon as it is finished. In others it is not divine until the ceremony is complete. One text avows: “This statue, unless its mouth is opened, cannot smell incense, eat food, or drink water.”⁴

The ritual was fraught with meaning for worshipers, but also for the god: mere wood became “godflesh.” Later Christian controversies about transubstantiation preserve the thought. Lanfranc, Anselm’s mentor, held that Christ’s body was “really and substantially” in the eucharist. God’s mercy masked the bloody sight, but sometimes it broke through and the flesh was seen. Isaiah’s irony finds a counterpart in the Hellenistic story of a sculptor facing two offers for the same statue: one to be dedicated in a temple and the other as a memorial. That night the statue told the sculptor in a dream that its fate was in his hands: to make him a god, or a dead man.

Not surprisingly, the role of priests in sanctifying images was shrouded in secrecy. In some ceremonies, the artisan ritually denied the product was of his own making. In one Babylonian liturgy, the artisan was to say: “I did not make him; Ningal, god of the smith, made him.”⁵ Underscoring the point, artisans’ hands might be symbolically severed. In theory, the work was a collaboration of divine and human agency, but the stress was on the divine contribution.⁶

The ancient rituals are secularized somewhat in the Egyptian practice of inspiriting various objects, including mummies to be revived in the hereafter.⁷ Philo of Byblos (first century CE) cites a Phoenician genealogy of “Bethyl,” a son of Ouranos, reflecting the belief that the god of the heavens created betyls, ensouled stones (baitulia lithous empsychous).⁸ Archaeologists have long studied the so-called betyls enshrined by the Nabataeans in Petra. Some were memorials and called nefesh. But betyls proper were classed as gods and heaven-born.⁹ The black stone within the Kaaba in Mecca, perhaps of meteoric origin, is venerated today. The name βαιτυλια (baetylia) may echo the Hebrew לא תב (Beth El); Nabataean stones, aniconic or carved with eyes and other facial features, recall the thought of stones as dwelling places of divinities (cf. Genesis 28:22).

The Philosophy of Oracles, ascribed to Porphyry (ca. 234–305), piously outlines procedures for investing graven images with divine or daemonic spirits.¹⁰ In a later (or more certain) work, perhaps reflecting the teachings of Plotinus (204/5–270) or Longinus,¹¹ Porphyry responds to monotheist critics: Pagans do not mistake statues for gods. The figures are symbolic.¹² Maimonides echoes this view in giving pagan piety the benefit of the doubt. But Porphyry’s student Iamblichus (d. ca. 330) rejects the Plotinian account for what may be a more traditional view.

Porphyry had raised a spate of objections to idolatry in the form of a letter purportedly to Anebon, an Egyptian priest, professedly seeking to allay his doubts, but more likely to expose the weaknesses of naive devotion. Iamblichus answers on the priest’s behalf that humans can fashion efficacious images, evidently still a traditional belief in Iamblichus’s time. Porphyry’s challenge: “Why should anyone exchange realities for images?” Wasn’t that the intent of the “Egyptians” in invoking divine spirits to dwell in their images?¹³

Here is the core idea that Jābir b. Ĥayyān, the notional doyen of Islamic alchemy,¹⁴ fathers upon Porphyry, perhaps influenced by the earlier, perhaps pseudonymous, work. Iamblichus, in De Mysteriis, argues that mere mingling of elements cannot generate daemonic forms: Daemons are not physical. Yet humans can share in divine creative work: Eidolopoeia, or idol making, does seek to engage with spirits emanating from the divine world. Yet gods, surely, are not the only supernatural beings.¹⁵

In Hermetic writings, the idea that Porphyry scouts and Iamblichus rationalizes is parsed thus: Earthly spirits (daemons) will dwell in images made by human hands, but gods, being celestial, invest bodies fallen from the heavens—and must not be touched. Augustine, despite his vehement disapproval of pagan piety, cites the teaching of Hermes, “the Egyptian, called Trismegistus,” that “some gods are made by the supreme God, and some by men.” Visible, tangible images, on the pagan account, he explains, are but the bodies of the gods. The spirits invoked to dwell in them have powers to inflict harm or fulfill the desires of those who honor and pay them divine service. The art that unites such spirits with bodies, making them animated by the spirits that invest them, is the making of gods, a power that human beings have been given.

Just as the highest God is the maker of the gods in heaven, Augustine voices the view he finds in “Hermes,” so is man the maker of the gods content to dwell near men in temples. So human beings can imitate divine creativity: “Just as the Lord and Father made eternal gods like Himself, so mankind has fashioned gods in its own likeness.” Does this mean statues? Hermes replies: “Yes, statues, hard as it may seem for you to believe, sentient and ensouled, that do great and marvelous things—foresee and foretell the future—by lot, by prophet, by dreams, and in many other ways; they make men ill or well and grant joy or sorrow, as men deserve.”¹⁶

II

The alchemical question was a practical one: If Ouranos brought dead stones to life as gods and Israel’s God had animated clods of earth to create Adam, could humans not emulate such acts by fashioning a man—or even a god?

Jābir lays out the rationale: Metals differ only in their accidents, depending ultimately on the blend of mercury and sulfur of diverse kinds. All elements get their natures from the balance of the four cardinal qualities: hot, cold, wet, and dry. Lead and iron, copper and gold, and mercury and silver all reflect that balance and are in principle interchangeable.¹⁷ In his book On Actualizing Potentials, Jābir lists seven arts, culminating in the Art of Generation, based on balancing the four cardinal natures. He considers the sciences of talismans and theurgy, or istikhdām al-rūĥāniyyāt—putting into service spirits, or supernal beings (¢ulwiyyāt)—to be the science of properties, and, of course, medicine and alchemy.¹⁸

Generation means producing things in all three natural realms: mineral, vegetable, and animal. The crowning production, an artificial man, depends on the proper confluence of natures. Just as nature obeys the law of proportions in producing human beings, to generate an artificial man one must carefully observe the rules of equilibrium. Critics may challenge us, Jābir writes, denying this possibility. But they have no proof. Grant us, “masters of natures” (aśĥāb al-ţabā’i¢), that we can make minerals, vegetables, and nonhuman animals, or we’ll demonstrate that we can. So how can you deny our power to add just one more species when we can readily transform even things of one genus into another!¹⁹

Beyond the Platonizing idea that art imitates nature, Jābir boldly sought to put into practice Plato’s challenge that man must imitate God—not by seeking self-perfection, as ethical monotheism has it, but by emulating God’s creative powers,²⁰ putting oneself in the place of the divine creative spirit, isolating and recombining the natures of things, to make anything one desires. Divine creation projects all the natural forms at one fell swoop, but an alchemist must proceed gradually, step by step.

Jābir distinguishes the arts that can produce lower animals like wasps and bees, snakes, and the animals that arise in nature by spontaneous generation, from the art needed to produce quicker, “warmer” animals, endowed from the start with sound instincts. To produce an animal that is not only living but intelligent, insightful, imaginative, even capable of legislation (aśĥāb al-nawāmīs) is a higher challenge. That means, Kraus proposes, the making of prophets,²¹ reviving the ancient idea that the god-maker might induce prophecy.

Reliable witnesses have told Jābir, he avows, of partially formed animals found in clay; he has seen such things—fossils, perhaps. We humans were made just so, as scripture attests; and we cannot doubt we will be resurrected by such means. Ibn Ţufayl employs the notion in characterizing the scientific version of a man’s origins.²² Jābir details the techniques needed to produce various animals using figures, cosmic models, and organic materials, quoting “the great Porphyry” on the methods needed to control the color of the creature produced.

An allusion follows to the production of a perpetuum mobile, regrettably in a textually corrupt passage: The mold, Jābir writes, is critical. It may have a human or animal form, and one should build the apparatus of crystal or stone (crystal is best) in the desired form (body of a girl, face of a man, intelligence of an adult, etc.); build a sphere half again the size of the mold; add smooth clay; cover and let dry; harden and polish as shiny as a mirror; add sperm, not chilled, or a bit of flesh from the animal desired plus various drugs to provide the right balance; put all the organs in place; and spin. The movement, we learn, will become spontaneous and perpetual, as in a mill or water wheel. An example is the ever-moving statues of Egypt placed on an iron column with a polished concave mirror below (cf. our solar tops). The movement will match that of the spheres and will continue to the end of time. The technique is attributed to Porphyry and to one Zosima (Kraus, 110–13). The use of sperm is critical (116)—as Paracelsus also held.

If one makes intelligent, autonomous, or lawgiving beings, Jābir asks, where do they get their understanding and discourse of reason? Some who claim great expertise say from the brain, seat of the imagination, memory, and thought.²³ Others say from the apparatus and “instruction” of the operator. Some call it spontaneous, like the insights of Socrates and other great philosophers. Still others deny these claims, insisting that souls must enter a body ignorant but capable of learning. But even they seem to concede that souls must be infused—as the divine spirit was into those idols and as Adam’s soul was.

Many a medieval and renaissance man was entranced by the Faustian ideal that reached an apex in Paracelsus’s dream of the homunculus. 

III

Youngsters in Sunday school sometimes ask: Can God make a stone He cannot lift? College sophomores, and sometimes their elders too, may ask: If God made everything, who made God? In the ancient texts that lack Plato’s idea of an absolute, or the monotheistic idea of a creator, it hardly seems impossible to ask: Can God make a god?

There’s a hope and fear distinctive to the world of homo faber that we humans might make a creature more powerful than we are. The creatures of this mythology bear gifts and threats, reflecting the ambivalence that engenders them. The Frankensteins and golems, homunculi and mutant succubi, like the extraterrestrials of another branch of scientistic myth, are both saviors and terminators, judges, innocents, and avengers, their visits not always fleeting.

Today we know what an element is, and we can, at times, transform one element into another or even generate new elements. A consulting company once did make a silk purse out of a sow’s ear. They have not quite reversed the process. But we do have cloning now. We have no perpetuum mobile, but we have learned to translate matter into energy with the controlled, and at times uncontrollable, power of nuclear fission, harnessed in productive work or unleashed in devastating destruction. Having not quite mastered fission, we pursue fusion, with a potential, as Sir Arthur Eddington foresaw, of virtually unlimited energy.

An ordinary computer not very different from the one that holds the present text can perform billions of logical operations per second; some machines can fly much farther and faster. Human brains are vastly more complex. But an ordinary laptop or handheld computer can house libraries and reach beyond its shell to unlimited information, sound and unsound. Computer history has accelerated since the Babylonians and the Chinese first used the abacus to aid their calculations.

It was in 1642 that Pascal, at the age of nineteen, invented the first gear-driven computer to help his father keep government accounts. Thirty years later, Leibniz (1646–1716) devised a computer that could add, subtract, multiply, divide, and extract roots. He also conceived a universal binary code meant to be able to express all concepts and fuel a truth machine. Less than three hundred years later, Turing conceived the possibility that a machine might pass for human. But notice the price to be paid: defining humanity in purely functional terms, as if we humans were input-output mechanisms.

Pascal and Leibniz both worried and wondered about the boundaries they touched, and today futurists and philosophers, not just science fiction writers, anticipate a day, perhaps not far off, when computers will become more than the oracles credited or blamed for the answers they provide to questions that mere human minds may seem too slow or dull to answer or even adequately frame. Science fantasists envision autonomous supercomputers that will govern, direct, and use their human subjects —a forecast feared by some and eagerly awaited by others.

Čapek, who coined the term robot in his 1921 play R.U.R., or Rossum’s Universal Robots, envisions the creation of an artificial species that escapes the limitations of mere mechanism, destroys the human race, and ultimately discovers love and self-sacrifice, letting the world, as it were, begin again. Čapek rewrote his play when he realized that he had recapitulated the legend of the golem, the portentous homunculus with the power to serve, perhaps to save, but also to destroy humanity. I doubt if Čapek knew that the golem itself, a creature of rabbinic legend, and of homilies still in use today in film and in fiction, was a transfigured idol.

Where ancient priests used ritual to draw down a god to invest dead matter with powers for good or ill, modern dreamers who see the soul as a mere computer program anticipate transposing one’s identity into a self-repairing machine and winning immortality not with the odd bionic limb or artificial organ but a full suite of nonorganic parts—eternal life not in heaven or some cryonic vat but as computer software.

Robert Geraci, studying the work of contemporary AI investigators like Marvin Minsky, Ray Kurzweil, and Hugo de Garis and robotics specialists like Hans Moravec and Kevin Warwick, explains that these science fantasists, having a profound distaste for our human embodiment and its concomitants of aging, illness, and death—and impatient with the limits embodiment imposes on their scientific work—expect AI and robotics to “create a paradise for humanity in the short term” but ultimately to render humanity obsolete: human beings, these transhumanists prophesy, “will need to upload their minds into machine bodies in order to remain a viable life-form.... Mere human beings will not fit in” to this newer brave new world. The new AI, as its advocates believe, “will make robot life better than human life.”²⁴

I’m dubious about such claims on both technical and moral grounds. Even conceptually, the idea of machine intelligence runs aground as a result of its rather primitive notion of intelligence. Phenomenologists like Merleau-Ponty and Hubert Dreyfus have rightly stressed the importance of our embodiment in situating human consciousness; so have analysts like John Searle. The medium, they argue, pace reductionists like Daniel Dennett, is not irrelevant to the message: human thought is situated. Perspective matters in grounding the values that orient real thinking. Continuity, to add a Kantian, Jamesian, and Bergsonian point, is critical: we can’t know if we don’t care; and we won’t care if we don’t in some way know who we are. Values are not just a decorative flourish added to purely cognitive data. As Spinoza saw clearly, a valuing, committed, emotive stance is constitutive in the very act of thinking.

Fantasists may imagine that machines too, given their heads, may establish values for themselves. Perhaps that’s hinted in Jābir’s idea of a legislating homunculus. But the notion is highly problematic. The obvious riposte is that a machine’s values, if it had any, would reflect what was programmed into it—serving the designer’s ends or some further function such as self-perpetuation, or perhaps propagating more machines. That is not autonomy. Autonomy in the human case involves the projection (and even critique) of values of one’s own.

We can get a clear idea of the incoherence of the notion of machine intelligence if we consider just how primitive it is to think of knowledge as an array of facts. The obvious question that arises about vetting data is a timeless issue gone critical in the new information age.

Beyond that, consider the fabric of our knowledge: most of what we know scientifically, quite apart from the seeming imponderables and intangibles of the humanities and the arts, takes shape in theories. The interpretation and even recognition of data is intimately bound up in the theories that offer nesting places for observations and aperçus. Anyone can see the swirling of water running down the drain. But without a theory of the earth’s rotation, the effect is unlikely to be seen as evidence for that rotation.

Could a machine be said to know if it lacks the capability of framing theories? And could it be said to learn if it lacks a way of weighting the values on which theory choice depends? How do we relate values like accuracy and breadth, coherence and consistency, objectivity and economy, consilience and heuristic fertility, not to mention the fugitive amalgam known as elegance? There’s no algorithm for that, and the effort to build one will inevitably privilege certain styles of inquiry, if not specific bodies of theory. A machine cannot know if a machine cannot discover, and there’s no algorithm for insight—or creativity.

The vastness of the promised powers of machines—the limitless capacity for work, for computation and memory storage, for production and destruction contemplated in transhuman scenarios—reflects a mythic vision. The ambivalence of the imagined powers confirms that suspicion: hope and fear drive the projections, much as they did the promises of ancient alchemists. But the kernel of truth latent in the new myths lies in the same bivalence or ambivalence: limitlessness is both the dream and the lie. The persistence of moral choice is the truth lurking in the myths, with all their smoke and thunder, laser light and display. Computers, no matter how fast, will still need thinkers to set their parameters and assign their purposes.

Hence the warning implicit in Goethe’s (1749–1832) ballad, “The Sorcerer’s Apprentice” (1797): be careful what you create. That’s the repeated message in medieval and early modern tales of golems: written on a golem’s forehead, plain for all to see, was the Hebrew word for Truth, Emet (תמא), a word composed of the first, middle, and last letters of the Hebrew alphabet and a name for God, the ultimate Truth and source of justice, which is truth.

The Talmud relates that Rava, a rabbinic sage, created a calf and later a golem using the alphabets preserved in the magical Sefer Yetzirah, or The Book of Creation. (Simon Magus would make gods, but Rava only emulated God’s work.) In one legend, Rava sent his creature to a colleague who addressed it, but it did not answer. Was that because we humans fall short of God’s power to create a man with the discourse of reason? Or was it that it would not speak, given humanity’s many failings?

In a commentary on the Sefer Yetzirah, the prophets Jeremiah and Ben Sira are said to have made a man using that book’s formulae—but the man erased the aleph on his head and died. The erasure is portentous, since aleph is the number 1, and thus stands for God. Its deletion leaves just the one-syllable word met (תמ): dead. Loss of the divine spirit is death.

In another version of the story of Jeremiah and Ben Sira’s golem, the work took three years; they made a man that did speak, but only to warn them to erase the aleph “lest the world revert to idolatry.” So they reversed their spells, erased the letter, and let it die. In a third version, the golem has a full sentence on its forehead: The Lord God is Truth. But it uses a knife to scratch out the aleph, leaving the words: The Lord God is dead. Jeremiah weeps at the blasphemy and warns in a parable against turning from admiring the architect to admiring the imitator. Here the moral becomes explicit: hubris brings the death of God.²⁵ In a pious song still sung with brio, pairs of attributes are catalogued alphabetically for ascription, eminently to God. Gershom Scholem singles out the fourth verse: discourse and discernment (ha-de‘ah ve-ha-dibbur) are His whose life is eternal—His  alone to give or withhold.

Early golem stories may have been told simply to show how close to God some ancient sage had come. But the later stories are cautionary tales. When the golem ceases to be a servant and becomes a monster, as these fables warn, much as many a robotic fable warns, they must be destroyed or they will destroy their makers. The doubts are not about the powers of technology but about its ambiguity.

Just as hope and fear are deep sources of fascination with the idea of an artificial man, or mind, or god, so are there two ways of responding to the image that such projects and projections conjure up. Both are humanistic in intent. Miskawayh (930–1030), a Muslim humanist historian, moralist, and statesman, typifies the optimistic view, reassuring doubters that no one could practice alchemy without passing through the alembic of philosophy and thus becoming immunized against abuses: homo faber will serve human interests, the goals toward which philosophy will turn those able to wield the power to create powers greater than our own. In that hopeful vision, every alchemist is an Einstein.