I’ll start with a proposed equation: “Artificial” + “Intelligence” + “Jewish” = Golem.
Maybe, maybe not: In Yiddish, goylem can be used, disparagingly, to indicate a lack of smarts, a real stupidity, if you will, which is precisely the opposite. Still, there’s a nagging, fundamental instinct that the golem is something more than just a brute creature, something that worries at the most fundamental boundaries of intelligence, humanity, and, yes, technology. This, in many ways, is the definition of monstrosity; and, like all monsters, whether the golem is a thinking creature or not, he is certainly something we can think with, something that can help us define and explore those boundaries. A brief journey through the history of the golem might help us see how that’s so.
The ur-golem starts in poetry.
Psalm 139:16 reads, in part, Golmi ra’u eynekha ve’al sifrekha kulam yikatevu. The Hebrew’s tricky. The good King James has it as “Thine eyes did see my substance, yet being unperfect; and in thy book all my members were written.” The JPS 1985 translation has “Your eyes saw my unformed limbs; they were all recorded in your book.” Glm, the root, then, with its sense of imperfection, not-quite- or pre-human; but this sense of “unperfect” or “unformed” suggests a kind of necessary crafting, or, we might say, techne, in order to get us where we need to be. In the Psalmist’s telling, that crafting seems to belong to Divine authority; but what kind of humans would we be if we didn’t try to arrogate to ourselves the technical capabilities of God and nature?
Indeed, every golem story, in one way or another, is a story of humans seeking to imitate that Divine creative urge. When the golem next appears in Jewish literature, we find the rabbis doing just that — and, in so doing, encountering certain limits. This is from the Babylonian Talmud, Sanhedrin 65b:
Raba said: If the righteous desired it, they could [by living a life of absolute purity] be creators, for it is written (Is. 59:2) “But your iniquities have made a separation [between you and your God].” Raba created a man and sent him to R. Zera. R. Zera spoke to him but received no answer. Thereupon he said to him: “Thou art a creature of the magicians. Return to thy dust.” R. Hanina and R. Oshaia spent every Sabbath eve in studying the “Book of Creation” [Sefer Yetsira] by means of which they created a third-grown calf and ate it.
We have here a Talmudic Turing test, basically: a fairly straightforward definition of perfect, divinely created humanity versus imperfect human technology. Following a line of argument most famously associated with Aristotle, the Talmud locates speech as the essential defining nature of personhood: If you don’t have it, you’re something else.
But it’s also — as everything is in the rabbinic world — a moral argument. Raba here suggests that this sort of creation is a mark of moral virtue, of human proximity to the Divine. It’s sort of the rabbinic equivalent of tech-bro power: If you can do something, you should; and, what’s more, if it gets done, it’s a sign of something complimentary about you. Rabbi Zera, by contrast, focuses not on the technical achievement, but on the gap between the accomplishment and the ambition: Raba may have been able to convince himself he’s created a human. He may even have been able to convince the narrator that it’s a human, who refers to it as such. But, impressive as such an act is, Rabbi Zera reminds it — and us — that it fails in the most basic acts of humanity, and thus does not, in fact, belong. It is, perhaps, monstrous.
The programmed golem. Despite the cautionary aspects in the Talmud’s account, there’s a pretty clear sense of technology’s benefit there, too: There’s that calf the rabbis create every Friday night for Shabbat dinner, after all. And so, that “Book of Creation” mentioned in the last sentence became the germ, through works and commentaries attributed to personages as elevated as Abraham, Rabbi Akiva, and Saadia Gaon, of a kind of programmer’s manual: High-performing golems are then the result of proper programming and instruction. As with today’s technological golems, the formulas are the result of the proper inputting of letters (which, in Hebrew, are also numbers). Here, for example, is a 13th-century commentary on the Sefer Yetsirah ascribed to Saadia:
They make a circle around the creatures and walk around the circle and recite the 221 alphabets, as they are noted, and some say that the Creator put power into the letters, so that a man makes a creature from virgin earth and kneads it and buries it in the ground, draws a circle and a sphere around the creature, and each time he goes around it recites one of the alphabets. This he should do 442 times. If he walks forward, the creature rises up alive, by virtue of the power inherent in the recitation of the letters. But if he wishes to destroy what he has made, he goes round backward, reciting the same alphabets from end to beginning. Then the creature sinks into the ground itself and dies.
Woe betide the creator who would recite the alphabet merely 441 times. Best-case scenario, presumably, the program wouldn’t work, and the creature would not rise (or would die). Worst case? Well, here’s a 17th-century account from a responsum written by Rabbi Tzvi Ashkenazi, better known as the Chacham Tzvi:
It has also been asserted concerning my grandfather, the Gaon, our master and teacher, Rabbi Elijah, chief rabbi of the holy community of Chelm [that he created a golem]. . . . When the Rabbi [Elijah of Chelm] saw that this creation of his hands grew larger and stronger because of the Name — which, written on parchment, was fastened to its forehead — he became afraid that the golem might cause havoc and destruction. Rabbi Elijah summoned enough courage and tore the parchment with the Name from his forehead. Then it collapsed like a clod of earth; but in falling, it damaged its master and scratched his face.
In the Chacham Tzvi’s telling, it’s not that the programming is performed incorrectly, exactly; it just fails to account for the consequences of, let’s say, too much computing power. Putting the Divine name on a creature’s forehead might be like trying to wire a modern motherboard into a Mac Classic: You might get it to work for a while, but don’t be surprised if it blows up. But it also suggests — predating Mary Shelley’s novel by more than a century — that technology inherently holds within it the germ of its own catastrophic failure: You can’t make a golem without using the Divine name, and you can’t use the Divine name without loss of control, because it transcends your control to begin with. In some ways, the kabbalistic model — of controlling the universe through its essential building blocks — is a fantasy. It will end in destruction, whether of the world around you or, failing that, of you, yourself (note the apparently extraneous end detail about the grandfather’s scratched face, which is in fact not extraneous at all).
It’s significant that this story comes in a rabbinical responsum, an answer to a legal question, about whether a golem can be a member of a prayer quorum, a minyan, and that the Chacham Tzvi answers firmly in the negative: It’s his belief that there’s an aspect of humanity that the golem cannot replicate. If a prayer quorum, as is famously believed, brings God’s presence to itself, that human capacity is not shared by what is, in the end, a tool.
The human golem. These earlier golems were creatures of a world in which humanity perceived itself in contradistinction to Divinity, so their attempts to emulate or echo that Divinity were doomed to condemnation — despite, as these stories also make clear, being capable of achieving some real success. But modernity, where that perception wavers along with belief, is a different story; and the lessons of the golem are trickier to draw. The great Yiddish writer I.L. Peretz, who wrote a short story called “The Golem” in 1894, ends his tale with the portrait of the figure lying
concealed in the uppermost part of the synagogue of Prague, covered with cobwebs that have been spun from wall to wall to encase the whole arcade so that it should be hidden from all human eyes. . . . The golem, you see, has not been forgotten. It is here! But the name that could bring the golem to life in times of need, that name has vanished into thin air. And no one is allowed to touch the cobwebs that thicken.
Do something — if you can!
For Peretz, the golem is more than just a force capable of supernatural protection of the Jewish from Gentile violence (though it is that), and more than just a creature that can go on the rampage (though it’s that, too). It’s a repository, potentially, of Jewish imagination and animation — all those letters and formulas and legal questions and Psalmic poetry taking something inert and bringing it to life. That’s what Peretz wants us to do: animate the golem through our imagination. And to do so in the form of stories like the one Peretz is telling: works that bring together ancient ideas and modern sensibilities to illuminate contemporary concerns — in Peretz’s case, the perils of a fragmenting, dissipating sense of national identity in the face of modernity.
A lot to place on a golem, perhaps; and not even a golem, but a story about a golem. But if there’s anything our contemporary anxieties about AI chatbots and large language models suggest, it’s that words have power.
Power enough, it should be said, that they’re the key to solving the problem of the golem, not just creating it. The stories tend to agree on what to do if faced with an out-of-control golem: Simply (if simply it is) remove the first aleph from the word emet engraved on its forehead, rendering the word as met, from “truth” to “death.” The golem, faced with a rewriting of its essence, has no choice but to obey its inscribed code, and it’s rendered inert, as much a brick as that old phone.
Which is a significant difference from human beings, needless to say: You can tell an enemy to drop dead, after all, but it’s a highly ineffective combat strategy. And so, in the end, it’s not a lack of language that differentiates us from our modern-day golems, but our ability to sidestep, to dance around, to liberate ourselves from its commands: which is not only, as it turns out, what makes us human, but what marks the difference between the controller and the controlled. In the end, it’s our flexibility of interpretation, of definition and redefinition, of story-making and boundary-setting and limit-determining, that’s the best (and only?) means of dealing with our new creations, their enormities and our anxieties.
