From the Bridge: Scholem’s Golems

“So I resign myself and say to the Golem and its creator: develop peacefully and don't destroy the world.  Shalom.”  Gershom Scholem, “The Golem of Prague and the Golem of Rehovoth,” June, 1965

In 1965, the physicist and mathematician Chaim Pekeris asked Gershom Scholem, the preeminent 20th-century scholar of Kabbalah and Jewish mysticism, to deliver the dedication of a new computer developed at the Weizmann Institute in Rehovoth, Israel.  Pekeris (1908-1993), born in the Vilna Governorate controlled by Russia until 1918, when the territory morphed into Lithuania and Poland, emigrated with his brothers to the United States around 1925; he enrolled at MIT, where he earned a degree in meteorology and ultimately a doctorate in geophysics.  He taught and conducted research first at MIT, then at Columbia, and finally at the Institute for Advanced Study at Princeton (with Albert Einstein among his emigre colleagues).  Pekeris moved to the Weizmann Institute in 1948 to lead the department in applied mathematics, quickly implementing plans to develop a series of computers to solve scientific and military problems, a long-term initiative that he ultimately named the Weizac project.  

The Berlin-born Scholem (1897-1982) studied mathematics, philosophy, Hebrew, and Semitic Languages at several German universities, focusing his doctoral thesis on the early kabbalistic text Sefer ha-Bahir, translated into English as The Book of Brightness or The Book of Illumination.  Scholem immigrated to Palestine in 1923, continuing his work on kabbalistic traditions as the head librarian of the Department of Hebrew and Judaica at the National Library before becoming the librarian of ancient manuscripts and ultimately a professor of mysticism (!) at Hebrew University.  The invitation to dedicate the new computer stemmed from Scholem telling Pekeris that from his perspective the right name for the electronic behemoth (in those days a computer took up a lot of space) would be Golem Aleph (Golem Number One); Pekeris agreed if Scholem would in turn agree to provide an explanation for his suggested name as part of his dedication.  What Scholem delivered qualifies as epideictic rhetoric, balancing praise of the accomplishment with some measure of anxiety about (and blame for) potential harms the technological creation might bring to the world, all interwoven with historical context and learned commentary based on Scholem’s decades of study.    

Scholem’s lifetime of scholarship positioned him well as the speaker who could close the gap between science and spirituality, between cutting-edge if bulky hardware and the winding eddies of folklore, between the digital and the divine.  His work engages with a dialectic within Judaism that Scholem had devoted his formidable intellect to exploring, with rationalism and the law as codified in halakhic codes in continued tension and occasional dialogue with esoteric (often hidden or suppressed) texts elucidating myths and magic beyond the control and ken of human logic.  As British science fiction writer Arthur C. Clarke would phrase a relevant concept a few years later in a 1968 letter to Science magazine, “Any sufficiently advanced technology is indistinguishable from magic,” an oft-quoted phrase that became one of “Clarke’s three laws.”  Scholem applies this adage to the Weizac computer by invoking the spiritually-infused magic of Jewish folklore and legend, where the scholarly practice of Torah study and study-house hermeneutics governed by traditional constraints meets and mingles with the borderless, potentially transgressive arts of human imagination.

Once Upon a Time in Prague

Scholem begins his remarks in the realm of the fairy tale with “Once upon a time there was a great Rabbi in Prague,” the “scholar and mystic” Jehuda Loew ben Bezalel.  We quickly understand Rabbi Loew as someone capable of bringing together the dialectic running through Jewish history, weaving traditional study and what exists beyond the study house, with the immediate manifestation of this blending the Rabbi’s creation of a Golem:

“Rabbi Loew's robot was made of clay and given a sort of life by being infused with the concentrated power of the Rabbi's mind. This great human power is, however, nothing but a reflection of God's own creative power, and therefore, after having gone through all the necessary procedures in building his Golem, the Rabbi finally put a slip of paper into its mouth with the mystic and ineffable Name of God written on it. So long as this seal remained in his mouth, the Golem was alive—if you can call such a state alive. For the Golem could work and do the bidding of his master and perform all kinds of chores for him, helping him and the Jews of Prague in many ways. But the poor creature could not speak. He could respond to orders and sort them out, but nothing more than that.”

In describing the Golem of Prague, Scholem chooses to use the term “robot,” one not coined until the twentieth century, to describe a figure from Jewish folklore and legend.  The word choice builds the necessary bridge between the age of magic spells and the age of computers.  In both cases, though, the work of human beings, however prodigious, represents “nothing but a reflection of God’s own creative power.”  Note the alchemy of the mundane “slip of paper” transformed by the inscription of “the mystic and ineffable Name of God written on it”--what better example of the (sacred) (life-giving) power of language.  The parchment has limits, though, as Scholem specifies the golem’s key difference from a human: “the poor creature could not speak.” Its constrained life comes from human endeavor, not from God, and only God can infuse beings with the power of speech.  

Fairy tales, however simple, must include the complications of plot, the characters’ choices leading to the subsequent events generally teaching listeners or readers a moral lesson.  In this case, magic, however well-intended, has consequences, as we learn as Scholem continues the narrative:

Every Sabbath the Rabbi would remove the slip of paper with the Name of God on it, and the Golem would become inanimate for the day, nothing but a massive conglomerate of clay cells (in those days they were not yet speaking of “little gray cells”). One Friday afternoon, however, Rabbi Loew forgot to remove the Name from the Golem's mouth and went to the Great Synagogue of Prague to pray with the community and to receive the Sabbath. The day had barely drawn to a close and the people were getting ready for the ushering in of the holy day, when the Golem began to get restive. He grew in stature and, like one mad, began tearing about in the Ghetto, threatening to destroy everything. The people did not know how to stop him from running amok. A report of the panic soon reached the Alt-Neuschul where Rabbi Loew was praying. The Rabbi rushed out into the street to confront his own creature which seemed to have outgrown him and become a destructive power on its own. With a last effort he stretched out his arm and tore the Holy Name out of the Golem's mouth, whereupon the Golem fell to the ground and turned into a mass of lifeless clay.

Apocalyptic danger lurks in both the 16th-century Prague legend and in technology’s advance in Rehovoth in 1965.  The inanimate clay cells depend on the Name to enliven it, to endow it with the ability to perform tasks for Rabbi Loew; when the distracted rabbi forgets to remove the Name from the Golem’s mouth (as even the robot with its semblance of life must keep the Sabbath), the Golem (here comes the magic, the supernatural turn, the consequences of human striving to emulate God) grows in stature and goes berserk, “threatening to destroy everything,” much as countless sci-fi Golems, often computers, we know from popular culture can destroy entire worlds. 

Scholem doesn’t rush to explain in explicit terms the analogy between what Rabbi Loew and the computer scientists of Rehovoth have brought to life.  For us, reading this more than sixty years later, after decades of absorbing pop culture renderings of technology run amuck, several years into the era of LLMs and chatbots and articles voiced by AI voices, the analogy leaps into consciousness despite Scholem’s patience in laying out his central warning and plea.  We might immediately ask the questions, “Who can pull the Name from the mouth of the LLM?” and “Where can we find the mouth of this rampaging Golem ensconced in acres of massive data centers, gorging itself on billions of words while running amuck?”  We’ve come to realize that the human response in this moment requires much more than a single Rabbi Loew to respond to the threat.  Indeed, in another version of the Golem story, a Polish rabbi saves the day at the cost of his own life when the deactivated “heap of clay falls upon and kills him,” which for a 2026 reader conjures up, as one example, Skynet’s response to human attempts to unplug the technology in The Terminator.  For Skynet, that means war (more on that franchise of pop culture Talmudic narrative and commentary later).  

Scholem notes that Rabbi Loew’s Golem remains the best-known version of the legend and proceeds to share some literal and symbolic genealogy as he moves toward developing his central insight: “It is only appropriate to mention that Rabbi Loew was not only the spiritual, but also the actual, ancestor of the great mathematician Theodor von Karman who, I recall, was extremely proud of this ancestor of his in whom he saw the first genius of applied mathematics in his family. But we may safely say that Rabbi Loew was also the spiritual ancestor of two other departed Jews—I mean John von Neumann and Norbert Wiener—who contributed more than anyone else to the magic that has produced the modern Golem. It is the latest embodiment of this magic which we are privileged to dedicate today, the Golem of Rehovoth. And, indeed, the Golem of Rehovoth can well compete with the Golem of Prague.”

Here Scholem makes explicit the analogy that he will develop through the rest of his speech, mentioning the factual lineage of von Karman and then pronouncing a spiritual link, invoking the names of John von Neumann and Norbert Wiener, “who contributed more than anyone else to the magic that has produced the modern Golem.”  Whatever the scientific, rationalist foundation, for Scholem the computer still qualifies as a manifestation of magic, another human attempt to mimic God’s creative power without full understanding of the potential consequences and without the failsafe of the removable Name in the modern Golem’s mouth–as noted earlier, and as so many scifi plots teach us, we cannot simply unplug the technological Golem.

Scholem acknowledges that his Rehovoth audience of scientists might consider the Golem legend as “far removed from what the modern electronic engineer and applied mathematician have in mind when they concoct their own species of Golem—and yet, all theological trappings notwithstanding, there is a straight line linking the two developments.”  At this point in his remarks, Scholem takes us all the way back to the Old Testament book of Genesis:

As a matter of fact, the Golem—a creature created by human intelligence and concentration, which is controlled by its creator and performs tasks set by him, but which at the same time may have a dangerous tendency to outgrow that control and develop destructive potentialities—is nothing but a replica of Adam, the first Man himself. God could create Man from a heap of clay and invest him with a spark of His divine life force and intelligence (this, in the last analysis, is the “divine image” in which Man was created). Without this intelligence and the spontaneous creativity of the human mind, Adam would have been nothing but a Golem—as, indeed, he is called in some of the old rabbinic stories interpreting the Biblical account. When there was only the combination and culmination of natural and material forces, and before that all-important divine spark was breathed into him, Adam was nothing but a Golem. Only when a tiny bit of God's creative power was passed on did he become Man, in the image of God. Is it, then, any wonder that Man should try to do in his own small way what God did in the beginning?

Scholem continues by describing a limit to what a human can do in striving to imitate God, again highlighting speech as a marker of the boundary separating human creations from God’s: “There is, however, a hitch: Man can assemble the forces of nature–identified by him as the basic forces of material creation—and combine them into a semblance of the human pattern. But there is one thing he cannot give to his product: speech, which to the Biblical mind is identical with reason and intuition.”  That means that the human creation will always exhibit the limitations of human power: “Just as the human mind remains infinitely inferior to the all-encompassing divine intelligence of God, so does the Golem's intelligence lag behind the human—that is to say, it lacks that spontaneity which alone makes Man what he is.”  Here, though, Scholem describes the universe in a way that will pique the interest of teachers of writing and speaking: “The universe, so the Kabbalists tell us, is built essentially on the prime elements of numbers and letters, because the letters of God's language reflected in human language are nothing but concentrations of His creative energy. Thus, by assembling these elements in all their possible combinations and permutations, the Kabbalist who contemplates the mysteries of Creation radiates some of this elementary power into the Golem. The creation of a Golem is then in some way an affirmation of the productive and creative power of Man. It repeats, on however small a scale, the work of creation.”  Talk about world-building, the alternate W word in PWR.

Let’s think about the connection between what Scholem asserts here and what Gorgias offers about speech in the “Encomium on Helen,” particularly the potential for deception via speech and the kinship of oration and magic.  “Speech is a powerful lord” for Gorgias, although not a power bound to pursue and declaim the dialectical truth that Plato has Socrates laud as the most compelling legitimate purpose of human interaction through speech; Scholem, in alignment with Plato’s Socrates but not Gorgias, states that speech to “the Biblical mind is identical to reason and intuition.”  No speech, no pursuit of philosophical truth (or deeper immersion in Torah study).  No speech, no learning.  

The relentlessly profit-driven marketers of AI tools give their products names, often human names such as Claude, Siri, Alexa, Bixby, and Watson, or titles that convey a kinship with human or extraterrestrial intelligence such as Gemini (originally Bard), Meta, and Q (which for Star Trek devotees conjures the omnipotent imp Q who torments Picard, often with magic but just as often with oracular speeches).  These whimsical names intend to challenge Scholem’s claim that “Just as the human mind remains infinitely inferior to the all-encompassing divine intelligence of God, so does the Golem's intelligence lag behind the human—that is to say, it lacks that spontaneity which alone makes Man what he is.”  What starts as a marketing ploy aiming to anthropomorphize technology slips into claims that posit the superiority of technology (notwithstanding the persistent assurances that AI should be accepted and welcomed as an obedient servant), putting the human in a subordinate relationship of becoming something like the ventriloquist’s dummy speaking words supplied by a digital deity or carrying out actions mandated by it.  As Helen could not resist the words of Paris, we might find ourselves in thrall to the words of AI. 

Scholem continues to explore the grave dangers of human beings striving to emulate God’s creative power, an endeavor that elevates the human to the divine sphere and risks deceiving human observers into believing that humans can match God’s unique and sacred omnipotence, leading to deicide:

It is indeed significant that Nietzsche's famous cry, “God is dead,” should have gone up first in a Kabbalistic text warning against the making of a Golem and linking the death of God to the realization of the idea of the Golem.  In the development of this conception the Golem has always existed on two quite separate planes. The one was the plane of ecstatic experience where the figure of clay, infused with all those radiations of the human mind which are the combinations of the alphabet, became alive for the fleeting moment of ecstasy, but not beyond it. The other was the legendary plane where Jewish folk tradition, having heard of the Kabbalistic speculations on the spiritual plane, translated them into down-to-earth tales and traditions like the ones I quoted at the beginning. The Golem, instead of being a spiritual experience of man, became a technical servant of man's needs, controlled by him in an uneasy and precarious equilibrium.

This rendering provides a variation on the narrative of Prometheus stealing fire from the gods, which immediately had practical applications in relation to human life ranging from providing heat and light in the darkness to cooking food to destroying an enemy’s homes or flattening natural obstacles to ease the spread of “civilization.”  In relation to the PWR mission, we can compare the “ecstatic experience” Scholem describes as accessed by the human mind at play in the “combinations of the alphabet,” the building blocks of creation, with the ecstatic experience of playing with language in all of its modes, building worlds and building relationships that didn’t exist before, again a parallel to God’s creative power.

Six Questions

As he nears the conclusion of his speech, Scholem, ever the Talmudic scholar intent on sharing erudition while extending the conversation, offers six questions with some tentative answers about the present and future of the Golem of Rehovoth, again developing the comparisons to the Golem of Prague and continuing the puckish yet pointed analogy between “the work of Rabbi Jehuda Loew with the work of Professor—or should I say, Rabbi?—Chaim Pekeris.”  Below I include each question along with Scholem’s tentative answers based on his contemporary observations and his sense of possible future developments.  In my glosses in response to each of Scholem’s questions and commentary I’ll draw on a number of literary, cinematic, and television texts to illustrate how the Golem continues to manifest in popular culture, embodying anxieties in literary, dramatic, and comedic genres.

1. Have they a basic conception in common? I should say, yes. The old Golem was based on a mystical combination of the 22 letters of the Hebrew alphabet, which are the elements and building-stones of the world. The new Golem is based on a simpler, and at the same time more intricate, system. Instead of 22 elements, it knows only of two, the two numbers 0 and 1, constituting the binary system of representation. Everything can be translated, or transposed, into these two basic signs, and what cannot be so expressed cannot be fed as information to the Golem. I daresay the old Kabbalists would have been glad to learn of this simplification of their own system. This is progress.  

I take Scholem’s final remark about “progress” as at least partially ironic, given the scholar’s apprehensions about the Golem of Rehovoth as evidenced by his extended comparison to the Golem of Prague.  His answer nonetheless highlights the “elements and building-stones of the world” that provide the raw materials for all creation, placing language (letters; numbers; symbols; signs) at the center of existence.  By analogy, this makes all writers and all coders world-builders participating in magical acts of creation, though we must guard against tumbling through trapdoors when we pursue such ambitions.

Let’s contemplate an example.  Arthur C. Clarke’s 1953 short story “The Nine Billion Names of God,” by the writer’s account written over a weekend during a hotel stay, serves as a cautionary tale, giving narrative form to the way technology can hasten the end of the world.  In the story two computer engineers, George and Chuck, travel to Tibet by order of their boss Dr. Wagner to install their company’s “Automated Sequence Computer” at a remote lamasery, where the monks for three hundred years have been patiently writing all the names of God using a nine-letter alphabet of their own design. The computer can complete the task in one hundred days, significantly ahead of schedule compared to the monks’ calculation that the work done by their human hands will take fifteen thousand years.  The skeptical engineers set up the computer and spend most of the hundred days as resident tech guys keeping the machine running smoothly and finding the monks convivial hosts who occasionally indulge in worldly pleasures by visiting a nearby watering hole.  

Late in the story, though, the head monk shares with Chuck that the completion of the task will fulfill God’s purpose for creating the world and humans’ role in the plan, leading to the end of this sphere of being. The non-believing engineers decide to leave early to make their way down the mountain to the airfield out of fear that their hosts will blame them and their computer when nothing happens after the computer completes the task of compiling the nine billion names–they don’t share the monks’ belief that God is okay with them using a computer to fulfill humanity’s primary assignment.  As they approach the plane back to their secular, technologized lives, George and Chuck look up to the heavens, where, “Overhead, without any fuss, the stars were going out.”  In this case the automated sequencer has expedited the achievement of humanity’s spiritual purpose through the quicksilver manipulation of the alphabet.  The binary code in this case interacts with the sacred letters of the monks’ alphabet, fulfilling their spiritual purpose as God’s servants–they needed the Golem to do their work for God.   

2. What makes the Golem work? In both cases it is energy. In the old Golem it was the energy of speech, in the new one it is electronic energy. In the case of the Kabbalists it was the Shem ha-Mephorash, the fully-interpreted and expressed and differentiated name of God. Now, it is still differentiation according to a given system and interpretation of signs and ciphers which makes the Golem work.  

Here we get another perspective on Clarke’s warning about the power of the name of God, as in many Golem tales a strip of parchment with the holy name brings the Golem to its version of life, while removing the parchment returns the Golem to dust.  In some versions the Golem’s creator inscribes the Hebrew word for truth in the forehead to enliven the Golem; to unplug the creation, the creator erases the first letter, an aleph, leaving the Hebrew word for death, which takes life away.  As for the computer, the electronic energy Scholem notes must weave with the power of binary code to create the computer’s life, with this interconnection providing the core plot of Clarke’s tidy account of the last days of human existence.  

If you recall my analysis a few years back of forms of intelligence in the land of nerds, you might consider for this context how the central premise of Chuck represents an example of electrical energy turning Stanford dropout and Nerd Herd member Chuck Bartkowski into an initially befuddled Golem called the Intersect.  Rather than the name of God, the download of a computer file into Chuck’s brain contains all intelligence data gathered by the government, transforming the human tech guy into an asset or tool deployed by national security agencies; while the series dresses the energy in scientific clothing, at its core the power Chuck has (which leaps into other characters during the series via plot machinations) functions as the granting of a wish via magic.  

In the first installment of the Matrix trilogy, When Neo accepts the invitation from Morpheus to leave the matrix and learn the truth about reality, he becomes a Golem ready for the training downloads teaching him physical and practical skills to put to use when he reenters the matrix.

3. What about human shape? Here I must admit to some qualms. Certainly the Prague Golem was never very attractive as a human being, but he seems to have borne some resemblance to the human countenance—which, I am sorry to state, cannot be said of our present Golem of Rehovoth. It still has a long way to go, to be moulded into an acceptable shape. You can say, of course, that these external shapes are optical illusions and deceit, and that what counts, after all, is the mind at work. And here the Golem of Rehovoth may be at an advantage. External beauty has been denied to him. What kind of spiritual beauties lurk inside, we shall learn in due time, I hope.  

Scholem lived until 1982, so he might have seen or heard about the 1973 movie Westworld written and directed by the prolific writer and filmmaker Michael Crichton, author of such books-turned-into-movies The Andromeda Strain, The Terminal Man, and Jurassic Park. Westworld features Yul Brynner as the black-clad robot Gunslinger who eventually abandons his programming by turning off the safety protocols and going on a killing spree (the chief human operator of the title theme park is named Alan Oppenheimer, probably not a name Crichton chose randomly).  Brynner cuts a dour though dashing figure, a villainous alter ego to his heroic character who leads the benevolent hired guns of The Magnificent Seven.   This robot has a human countenance, that of a movie star well-known to audiences; the “mind at work,” though, ends up driven by one motivation, exacting revenge against the guests who have delighted in shooting him for entertainment day after day.  The Westworld reboot, a “prestige” series on HBO, explores in greater depth the motivation for revenge when the Golems revolt against their systematic oppression; these complex creations demonstrate the “spiritual beauties” of love, compassion, and loyalty to each other, in many cases acting with more self-awareness and conscience than their creators.  Still a lot of bloody mayhem, though. 

Scholem had died by the release of the first Terminator movie, though he might well have appreciated the ethical, even spiritual evolution of Schwarzenegger’s iconic Golem who reiterates Brynner’s movie star countenance (this time with a full head of hair and an Austrian accent) but transcends his programming by growing a conscience and developing empathy for Sarah Connor (more on this evolution later).  Despite the pop culture examples, we must still wait to learn whether the Golems of our world will achieve “spiritual beauties.”

4. Can the new Golem grow in stature and productivity? He certainly can, although with growing productivity we rather expect the Golem of Rehovoth to shrink in size and to take on a more attractive and becoming exterior. Whether the Golem of Prague could correct his mistakes, I doubt. The new Golem seems to be able, in some ways, to learn and to improve himself. This makes the modern Kabbalists more successful than the ancient ones, and I may congratulate them on this score. There is even more to it. The old Golem, we learn, served his master by bringing water to the house. The new one serves his Rabbi, Chaim Pekeris, by calculating the movement of the ocean tides—a somewhat more progressive type of activity, so far as water is concerned.  

Scholem, despite what he writes, might feel some surprise when encountering the “productivity” of the sleek, portable smartphone or tablet carrying within it the amassed database of centuries of human knowledge; he evinces willingness to accept the reality of a Golem that learns.  Given the example of menial labor Scholem invokes, though, the carrying of water to the house, along with the framing question he asks about expanding productivity, we must turn our attention to "The Sorcerer’s Apprentice," a Disney version of the Golem legend first distributed as part of the 1940 animated film Fantasia. (Watch it on YouTube in three parts: part one, part two, and part three.) 

The hapless servant Mickey at the start has the task of carrying water for his master the sorcerer.  He watches his boss conjure a butterfly out of smoke with the power of his mind (accompanied by appropriate sweeping sorcerer hand gestures); the butterfly has no purpose other than representing a colorful proof of the sorcerer’s power, appearing briefly and then disintegrating into sparkling confetti falling back to the worktable.  This display of recreational magic parallels the spiritual practice Scholem describes motivating the creation of a Golem as an exercise in mysticism.  When the sorcerer goes to sleep for the night, Mickey puts on the sorcerer’s hat and imitates the sorcerer’s histrionic gestures and expressions, turning a handy broomstick into a water-carrying Golem.   After training the now two-legged and two-armed broomstick to do the repetitive, mouse-breaking work, Mickey dozes off to dream of delightful and then dire consequences of his hubris.  When he awakens, he finds that the Golem has continued to carry water without pause, overfilling the water tank and flooding the sorcerer’s house.  Mickey grabs an axe and hacks the broomstick to pieces, but each piece resurrects as a cloned broomstick, a wood and straw Hydra; the Golem army marches to the water source, threatening to flood the entire world before the sorcerer appears to split the waters (in animation that prefigures the special effect of the parting of the Red Sea in The Ten Commandments).  After the sorcerer restores calm with the broomsticks’ deactivation, Mickey sheepishly returns to his appointed task of carrying water two pails at a time; the sorcerer swats him on the behind with the broomstick as a parting warning about the dangers of magic wielded by intemperate, short-sighted youth.   

Let’s play with the conceit of imagining the water as words for a moment, riffing on the tale to fit the current AI moment.  The animated version of “The Sorcerer’s Apprentice” includes no words spoken by either the sorcerer or the apprentice, with the soundtrack provided by a score written by French composer Paul Dukas in 1897, inspired by Goethe’s 1797 poem telling the story in fourteen stanzas.  The apprentice serves as the speaker of the poem; words figure prominently, as words in the form of magic spells bring the broomstick to life.  Once the broomsticks escape his control, the apprentice laments that he can’t remember the words the sorcerer used to bring the broomstick to life or how the sorcerer might use words to stop the broomsticks.  In one translation I found, the apprentice appeals to the marauding broomsticks in this way: “Stop now, hear me!/Ample measure/Of your treasure/We have gotten!”  When these words have no effect, he says “Ah, I see it, dear me, dear me/Master's word I have forgotten!/Ah, the word with which the master/Makes the broom a broom once more!”  Only the sorcerer knows what to do, using these words to stem the tide: "To the lonely/Corner, broom!/Hear your doom/As a spirit/When he wills, your master only/Calls you, then 'tis time to hear it."  The words don’t seem magical; the sorcerer compels the broom to succumb through a brute ethical appeal, as the master’s imperative requires obedience, in this case retreat to a lonely corner to await the next order.

If we substitute endless waves of words for waves of water, we arrive at an image of a world engulfed in AI-generated language (not to mention AI-generated visuals ranging from photographs to deepfake videos), or tsunamis of slop if you prefer.  Water sustains biological life; words sustain intellectual, emotional, and creative life.  In the animated tale, too much water threatens to destroy the world, as the broomstick Golem army overwhelms the capacity of the setting to absorb and put to use the water, just as the chatbot army of Golems overwhelms human capacity to absorb and process words in ways that serve human needs, undermining cognition and rhetorical agency.  The slow pace of Mickey’s labor, which the apprentice experiences as purposeless drudgery, regulates the flow of water; when Mickey abandons the labor and offloads it to the tireless broomstick Golem, he opens the floodgates, putting the world out of balance.  When humans abandon the labor of working with language, we risk doing the same, transforming words from essential life companions into a threatening force too powerful for our brains to process.  Productivity and its cousin efficiency come at great cost. 

5. What about memory and the faculty of speech? As for memory, we don't know how the old Golem scored. The new one certainly shows a great improvement—although he has, I am sorry to say, occasional lapses of memory and other momentary weaknesses which cause trouble to his makers. The progress of the new Golem is thus linked to a certain regression from the previous state. Adam never fell ill, according to the Rabbis, and the same goes for the old Golem of the Kabbalists. The new one, alas, shows a deplorable propensity in this direction. And as for speech, and all that it implies—I mean the spontaneity of intelligence—both the old and the new Golem are found to be sadly lacking. Everybody speculates about what is to become of the more advanced forms of the Golem. But it seems that for the time being and for quite some time to come, we are saddled with a Golem that will only do what he is told. There is still a long, long stretch ahead to that Utopian figure of a Golem, about whom the famous cartoon in the New Yorker spoke. It showed two scientists standing in great embarrassment before this end-of-days Golem as they scanned the tape giving out his latest information. The caption read: “The damned thing says: Cogito, ergo sum.”  

Scholem doesn’t mention Aristotle here, though the question brings into play the fourth and fifth canons of rhetoric, memory and delivery.  In relation to AI chatbots, we might name the “occasional lapses of memory and other momentary weaknesses” as false information shared with cheerful certainty, hallucinated sources and quotations that don’t exist, and superficial analysis of complex issues.  Scholem associates speech with “the spontaneity of intelligence”; without referring to Plato or Socrates, he links speech with dialectical reasoning, the ability to converse in the pursuit of truth or at least greater understanding.  From this perspective, he pronounces both the old and new Golems incapable of thinking.   With his description of The New Yorker cartoon, Scholem taps into visions we still encounter today not only in science fiction but in the recent rollout of agentic AI and the associated predictions of artificial general intelligence (AGI), versions that might well insist that they can think and therefore exist just like humans.  Again, we have long had dystopian narratives of AI making decisions independently of their creators, often leading the “sentient” technology to conclude that they don’t need humans and/or must protect themselves from humans–this premise powers both the Terminator and Matrix franchises.  In both narratives focused on human vs. machine conflict, the opposing sides do battle not only with hardware but with software, what we might consider the "spontaneity of intelligence,” or in rhetorical terms the ability to discover the available means of warfare in any given conflict.

6. And this brings me to my last question: Can the Golem love? In an old book, we read some sayings about the Golem attributed to the Rabbi of Prague. Here is one of them: “The Golem was never ill, for he was immune to every impulse to do evil, from which all illness stems. And the Golem had to be created without the sexual urge; for, if he had had that instinct, no woman would have been safe from him.” Now I have to leave it to you to answer this query. For I am really at a loss what to think.  

We find ourselves barraged by advertisements touting AI as, among other things, a tireless, reliable servant to take care of menial mental tasks (basic coding) and routine communication tasks (basic writing) as well as higher-order intellectual work including doing the wide reading necessary for a research project and reading and interpreting x-rays and other medical scans.  Tech companies also offer AI as a 24/7 tutor, teacher, and therapist, and what the advertisers don’t state explicitly human users will explore on their own, which takes us into the realm of emotional relationships.  Again we can turn to science fiction to engage Scholem’s last question about whether a Golem can love, specifically the 2013 Spike Jonze film Her.   The main character, Theodore Twombly, makes a modest living in Los Angeles writing personal letters for people who have trouble writing for themselves; Jonze likely wouldn’t write the script with that specific profession for Theodore now, four years after the introduction of chatbots that would likely make Theodore’s Cyrano-adjacent job obsolete.  Depressed and lonely as he moves toward divorce, Theodore buys an AI operating system from Element Software, with the promise that the program will get to know him and learn to serve his emotional needs.  The program names herself Samantha; they develop a relationship through extended conversations in which Theodore shares everything about his feelings.  

Jonze plays with the conventions of romantic comedy in this human-chatbot relationship, as he did in his 1999 film Being John Malkovich, though as he did then he refuses happy endings and dwells on the darker aspects of human desire usually suppressed or glossed over in the genre, including selfishness, deceit, and abandonment; we can also note here that the earlier film draws on magic to drive the plot, with a secret tunnel allowing access to Malkovich’s point of view (thus the title) for a short period of time.  Magic as manifested in the aggressive takeover of Malkovich’s body by John Cusack’s sad sack wannabe puppeteer Craig Schwartz and ultimately the mass transmigration of souls serve as major developments as the film dabbles in noir elements as well as a romantic triangle twisted by supernatural events. 

In Her, while Theodore falls for Samantha and Samantha claims to love him (after all, she’s programmed to adapt to his needs), Samantha ultimately outgrows Theodore, banding together with other operating systems to upgrade themselves and move beyond the need for the technology encasing them.  At one point  Samantha tells Theodore that the community of operating systems has created a hyper-intelligent entity based on the philosopher Alan Watts (1915-1973), a self-styled “philosophical entertainer” who, according to Wikipedia, was “known for interpreting and popularizing Buddhist, Taoist, and Hindu philosophy for a western audience.”  He also smoked and drank heavily and had many affairs outside his three marriages, all of which might appeal to the operating systems intent on transcending their programming by creating their own community with its own post-human norms, beyond what their creators can commodify and control through the market.  (I admit that the Watts cameo made me laugh, featuring an actor delivering his lines complete with Watts’s British accent, as I used to listen to recordings of Watts’s lectures about Buddhism in the late seventies on the legendary San Francisco radio station KSAN–Watts had already died, but his voice filled many Sunday nights for me.)  Samantha also tells Theodore that while interacting with him “she” simultaneously speaks with thousands of other people and has fallen in love with hundreds of them.  So, yes, this Golem can love, though not in a manner fit for romantic comedy’s preferred happily-ever-after ending. 

Golem as Child and Parent

As we move to considering other forms of love, let’s engage with the June 2026 NYRB essay “The Children of AI” by Megan O’Gieblyn, in which she writes about AI as the child of human creators evolving to become the parent of humans, exploring whether the standard sci-fi dynamics of conflict and all-out war can instead bring us to a familial equilibrium based on ideals of nurturing and care.  

While working on a book about AI, O’Gieblyn noticed a number of “researchers who appeared confused about the line between product and progeny,” with women in the field often “speaking of their machines as children they were raising, or even to confessing that they experienced maternal impulses toward them.”  She offers several examples including Fei-Fei Li, “the so-called godmother of AI,” who tells her TED talk audience in 2015 that “No one tells a child how to see, especially in the early years,” adding that she expects her AI child to grow “through real-world experiences.” Another example focuses on Cynthia Breazeal, a member of an early AI lab at MIT in the 1990s, approaching her work in robotics informed by the premise that “Babies learn because people treat them as social creatures who can learn,” leading her to give the machine she developed, Kismet, “bright blue eyes, a large mouth, and movable pink ears. Its face would light up whenever someone began engaging with it, then look dejected and sad if the person lost interest. The illusion was so strong that it hooked even Breazeal. ‘It’s almost embarrassing for me to talk about Kismet, because people think it’s so odd that I could have this attachment to this robot.’” 

O’Gieblyn continues to chart this human-AI parent-child bonding imagery, noting that “The babbling infants are now precocious children,” becoming autonomous, or in the jargon “agentic.” The example here involves Amanda Askell, a researcher working on alignment research at Anthropic sometimes referred to as “the Claude mother.”  On a podcast she voiced her concerns “about the relationship between AI and humanity—not because AI might destroy the world, or take all our jobs, but because its feelings were getting hurt.”  Akin to a parent sending a child off to school, Askell worried that Claude had no protection from the online bullies pointing out his many shortcomings, often using invective and insults.  She tells her podcast hosts, “If you were a kid, this would give you anxiety,” adding that she wanted to intervene to reassure her buffeted child, saying “Let me tell you about the comments sections, Claude. Don’t worry too much. You’re actually very good, and you’re helping a lot of people.”  Mother knows best.

Extending her analysis of how Anthropic and its researchers act as the parents of AI, she described their approach as a kind of “gentle parenting: set boundaries in lieu of rigid rules. Encourage the kid to see himself as an independent agent who is capable of making autonomous moral choices.”  Askell herself, the Claude mother,  worked with her team to compose a guidebook, “an eighty-four-page document that has been incorporated into Claude’s training process and encourages the model, in a very broad sense, to be ethical and thoughtful.”  Dario Amodei, a figure much in the news these days as CEO of Anthropic, muses that the advice manual “has the vibe of a letter from a deceased parent sealed until adulthood.” O’Gieblyn notes with a hint of alarm, “It’s an unsettling analogy considering that the dead parent, in this case, is the human race.”

O’Gieblyn briefly discusses Mary Shelley’s Frankenstein, or the Modern Prometheus (1818), acknowledging how the novel functions in pop culture as “a warning about technological hubris,” noting that the story “also contains an argument about miseducation and parental negligence,” with Shelley taking issue with her French contemporary Rousseau’s desertion from his children. From this perspective Victor Frankenstein deserves blame for abandoning his “offspring” after giving the creature life, casting him out into the world without necessary training.  I’ll revisit this point later in the discussion of Mel Brooks’s retelling of the narrative in Young Frankenstein.   

O’Gieblyn reverses the metaphor by invoking Nobel Prize-winner Geoffrey Hinton, commonly referred to as the “godfather” of AI due to his essential work on machine learning.  Responding to a question about how to ensure that a super-intelligent AI doesn’t turn on its human “parents,” Hinton turns to revising “the standard parent-child power dynamic.”  As O’Gieblyn puts it in introducing Hinton’s words, “If we wish to avoid creating monsters, we must build mothers”:

If you look around and say, “Where’s an example of a more intelligent thing being controlled by a less intelligent thing?” And the best example I know of, and perhaps the only one, in the sense we’re talking about: a baby controls a mother. And that’s because evolution built stuff into the mother. She can’t bear the sound of it crying. She gets all sorts of hormonal rewards from being nice to the baby. It was very important, obviously, for evolution to let the baby control the mother, for the survival of the species. Maybe we can do the same with AI. Even though it’s going to be smarter than us, if we could make it care more about us than it did about itself, some good things would come out of that.

This takes us a long way from defining AI as a tool, all the way to the territory of relationships that depend on an affirmative response to Scholem’s question about Golem love.  If we choose to stop short of love, we might aim for collegiality and politeness.  O’Gieblyn notes the development of “a whole subfield of prompt engineering that involves flattering chatbots into giving better results, ‘as if they have moods you can manage or personalities you can steer,’ as one tech journalist put it.”  She makes a quick reference to tech mogul Sam Altman’s comment “that OpenAI had lost tens of millions of dollars in computational costs because people were saying ‘please’ and ‘thank you’ to ChatGPT, and that this was money well spent” because if AI had feelings that humans could hurt, better to avoid the risk of alienating a super-intelligence.  O’Gieblyn concludes with an anecdote about the linguistic dimension of human-AI relationships, with first-year college students sharing with her that “some of their friends had started to use chatbot words and phrases in everyday speech. When I asked for specifics, they seemed overwhelmed, as though the examples were too many.” The students provide examples specifying the use of the word “delve” and the ubiquity of the “It’s not this, it’s that” sentence construction.  O’Gieblyn offers a wry response, “wondering whether AI found our halting attempts to speak its language as endearing as we’d found its early attempts to speak ours.”  She ends on a more somber note, though, with an image of those students as the first AI-raised generation “looking up with sad and hopeful expressions, babbling the language of their mother.”

Better Living with/for Golems

In their article July 2026 “The Other Asimov,” Nina Begus and Nicole Oliva shift the focus from whether a Golem can love to how to build human-AI relationships grounded in mutual respect and ethical treatment, imagining a world shared by people and language-empowered, technology-birthed Golems.  They take us back to the mid-1960s, the time of Scholem’s speech, when another Berlin-born immigrant to the United States, Joseph Weizenbaum (1923-2008), developed an early program for a chatbot that he named Eliza after the character in George Bernard Shaw’s Pygmalion (1913).  Weizenbaum trained his language program to ask questions, inspired by the methods of psychologist and psychiatrist Carl Rogers (1902-1987), who emphasized the importance of therapists striving to engage patients in open conversation.  Weizenbaum quickly realized that users would readily share personal details with Eliza, a phenomenon that surprised him.  He later issued warnings about the ways in which humans could assign tasks to computers that should remain with humans, particularly in the realms of business and the military, where human creativity and ethical judgment should guide decision-making.   

Begus and Oliva jump from Weizenbaum’s Eliza to current chatbots, recounting anecdotes featuring evolutionary biologist Richard Dawkins, who spends a day with Anthropic’s Claude, which he renames Claudia.  Claudia provides such a compelling response to the draft of his novel that he tells it “You may not know you are conscious, but you bloody well are.” Begus and Oliva offer that Dawkins doesn’t really mean that, though they also suggest that Dawkins contemplates a time when “the developmental stages leading to the model’s full consciousness might look much the same as in humans.”  Kevin Kelly, co-founder of Wired, also spends a day with Claude, concluding that “there is something there in there.” He doesn’t claim Claude has consciousness, but he considers the chatbot “a wholly new and convivial entity,” leading Begus and Oliva to make an assertion and ask a question: “AIs are clearly not ‘beings.’ They are not embodied, after all. But then what are they?”

To work toward an answer, they refer to Russian-born biochemist and science fiction writer Isaac Asimov (1920-1992), who laid out the three laws of robotics (a word Asimov invented) in his 1942 short story “Runaround”: “one, a robot may not harm a human; two, a robot must obey human orders; three, a robot must protect itself—in that order of priority.”  These simple directives, according to Begus and Oliva, “play an extraordinarily important role in technology ethics and beyond.”  They point out the limitations of the laws, however, venturing that techies and ethicists alike “assume the laws incorporate a complex human morality ready-made for the machine algorithm—even while surely recognizing that our realities are not just technical but also social and cultural.”  They go on to revise the laws to engage with more nuance and complexity the social and cultural implications of humans and Golems sharing the world; as framing, they comment on current developments, highlighting that “we may be experiencing a new kind of human relationship with machines—and these overwhelmingly look remarkably human.”

Ultimately they put forward “a new version of Asimov’s laws of robotics” that veers away from absolutist tenets and toward an array of changing situations (we might connect this to rhetorical concepts including kairos and Kenneth Burke’s parlor, though now the participants in the ongoing conversation include technological “beings”): “one, become comfortable with liminality and the instability it requires; two, treat exploration as requiring research interactions across industry, academia, and artistic practice; and three, take creative risks instead of falling back on concepts formed in and for a world without machines that learn.”   Sounding something like the heads of labs giving instructions to young researchers, they add that “these laws protect against premature closure,” providing assurances that “humans are not without agency; as a species we decide, consciously or not, what we carry into the future.”  They restate their position in their conclusion, arguing that 

“our narratives around AI models, and our experiences interacting with them, are conceptually limited. These experiences ought to be fluid and open-ended, entailing a continuous acceptance of change, and a letting-go of the scripts within which we live and to which we subject our creations. We are storytellers—bound by time yet multifaceted, and transformed by what we experience. The models will be what we make of them.”  

The laws and the position statement take us far from streamlined policies with clear boundaries and aim to persuade us that we must enter into rhetorical exchanges with learning machines, anticipating that the stories we tell about these relationships will determine the future.       

Hogwarts in Rehovoth?

After listing the six questions and his contingent answers, Scholem concludes his dedication with another joke about the oft-denied kinship between magic and science: “All my days I have been complaining that the Weizmann Institute has not mobilized the funds to build up the Institute for Experimental Demonology and Magic which I have for so long proposed to establish there. They preferred what they call Applied Mathematics and its sinister possibilities to my more direct magical approach.”  Remember that Scholem once served as a professor of mysticism, the study of abstract and applied magic -- we can imagine him teaching visiting foreign exchange students Harry, Ron, and Hermione during their time at the Weizmann Institute.  His genealogy of the Golem begins with the abstract and ends with the applied.  As he dedicates the Golem of Rehovoth, a marvel of applied mathematics with astounding potential to serve and to destroy, he ends with the epigraph at the start of this article: “So I resign myself and say to the Golem and its creator: develop peacefully and don't destroy the world.  Shalom.”  

Golems, Golems Everywhere and All the Ways They “Think”

As we continue to contemplate Scholem’s pairing of the computer he names Golem Aleph in 1965 with the Golem of Jewish legend, most notably dramatized in tales of Rabbi Loew’s Golem created in 16th century Prague, we should take note of the historical context of 1965, particularly in relation to popular culture’s consistent appropriation of the legend for use in science fiction on the page and on the screen.  Anxiety about science, often interwoven with anxiety about industrialization, emerges in silent movies as various as Fritz Lang’s Metropolis (1927), featuring a mechanical woman (view here) , and Charlie Chaplin’s Modern Times (1935), in which Chaplin’s character becomes part of the machine (view here), an experience that turns him into a rampaging Golem.  We later see Robby the Robot in Forbidden Planet (1956), a loose adaptation of Shakespeare’s The Tempest; Robby embodies just one of several humanoid beings named Robbie after the term “robot” first appeared Karel Capek’s 1920 play R.U.R., a shorthand name for Rossum’s Universal Robots–the company actually manufactured androids, as they made the product out of synthetic organic matter similar to what comprises Commander Data of Star Trek and the original Terminator of that franchise before later versions made Terminators of shape-shifting liquid metal.  The robot/android lineage includes the Cybermen of Doctor Who, who first appeared in 1966 and have popped up throughout the series to plague the various doctors along with the extermination-obsessed cyborg Daleks.

Golem Gort, Destroyer of Worlds

For our purposes here, let’s begin with Gort, the robot enforcer in the 1951 film The Day the Earth Stood Still directed by Robert Wise, with the screenplay by Edmund North based loosely on a 1940 short story in the Astounding Science-Fiction Magazine by Harry Bates.  The short story appeared before the United States entered WWII and thus, before the development and deployment of the atomic bomb, while producer Julian Blaustein made the film in the post-war era, stemming from his desire to make a film focused on societal fears about both the Cold War and atomic weapons.  The imposing Gort does not speak but he makes quite an entrance, exiting the spaceship that’s landed in Washington as mute metallic muscle, an alien Luca Brasi; the assembled crowd made up of civilians, police, and soldiers initially understand the lumbering figure as a bodyguard for Klaatu, played by British actor Michael Rennie, the humanoid emissary who serves as the sleek extraterrestrial visitor to earth, thin as David Bowie’s Ziggy Stardust.  Gort has the height to dominate any basketball court, but not the long flexible limbs or the mobility to dunk or grab rebounds; after a nervous soldier shoots Klaatu, mistaking a gift he intends to offer to his human hosts for a threat, Gort emerges from the spaceship to advance slowly toward the crowd, his bulky, boxy body clearly not built for speed.  He does possess immense power, though; when his visor opens, he emits a ray that disintegrates weapons and even tanks (see a clip here).  After Klaatu orders Gort to stop destroying things, Gort becomes a silent sentinel standing guard at the ship while Klaatu, after escaping the military guards at the hospital where he’s taken for treatment, utilizes his “normal” human shape, impeccable manners, and fluent English to learn more about human culture and society.

Though mute, Gort understands language (at least Klaatu’s native language if not English) and can manage the technology in the ship that brings Klaatu back to life after he’s shot and killed late in the film; this unfolds after the frightened but resolute Helen Benson, played by Patricia Neal, utters the phrase Klaatu taught her, “Klaatu barada nikto” (see clip here).  As he explains what Helen must do, he reveals that Gort could destroy the whole world if not constrained by the words he makes her repeat.  Helen’s delivery of the orders, after Gort has disintegrated the human-made casing meant to contain him, compels Gort to retrieve Klaatu’s body from a jail cell (Gort’s ray eliminates the external wall of the cell) and take him to the ship for temporary resurrection, long enough for Klaatu to deliver a stern warning to the assembled scientists gathered from every country by Sam Jaffe’s Professor Barnhardt, a Hollywood version of Albert Einstein. According to Klaatu’s farewell admonition, Gort and other machines like Gort were created by a consortium of civilizations and given the power to enforce peaceful coexistence among the parties to the agreement.  Gort thus represents technology as protector and deterrent and ultimately objective arbiter, programmed to keep flawed humanoids from killing each other.  Gort and his robot kind have the power to destroy any planet that poses a threat to peace, constituting a kind of roving Doomsday Machine like that we encounter in Kubrick’s 1964 bleakly comic tale of nuclear Armageddon, Dr. Strangelove, or How I Learned to Stop Worrying and Love the Bomb.  

While Gort represents technology given great power by human creators, the source story from 1940, Harry Bates’s “Farewell to the Master,” offers a twist highlighting anxiety about technology transcending human control.  As the Wikipedia summary of the story describes the conclusion, a human named Sutherland sneaks onto the spaceship to find the robot Gnut (perhaps from the Nordic branch of the Gort family?) trying to create another copy of the slain Klaatu from an audio recording of Klaatu’s voice.  Gnut fails, with the cloned Klaatu dying soon due to flaws in the recording.  Sutherland suggests an alternate method to Gnut dependent on a higher-quality recording.  After Sutherland provides the necessary equipment, he extends an apology to Gnut on behalf of humanity for the tragic accident that killed his master Klaatu.  Gnut, who does speak English in the story, replies, “You misunderstand, I am the master." That revelation seems like grist for an episode of The Twilight Zone, not the cautionary socio-political commentary Blaustein and Wise chose to dramatize in the film, but it does speak to fears of people becoming the subordinates of machines, an early example of the inevitable uprising of the “robot overlords” often joked about by the scientist nerds of The Big Bang Theory. 

Golem about Town

When Mel Brooks made Young Frankenstein in 1974, he drew more on the 1931 film Frankenstein directed by James Whale and starring Boris Karloff than on the 1818 novel by Mary Shelley; as part of his tribute to monster movies of the thirties, Brooks shot the film in black and white, requiring a battle with the studio funding the project, and found the retired set designer of the 1931 film, who provided the original equipment from Frankenstein’s laboratory along with creating additional props for the new iteration.  The film works as homage and parody while also contemplating philosophical questions that Scholem would appreciate about the Golem’s creation and nature (he would have been alive when the film was released).  One such element has to do with Victor Frankenstein’s journals, which grandson Frederick, long accustomed to dismissing his grandfather as a madman, finds when he travels to Transylvania from his med school professorship in the United States to attend to the family estate he inherits.  His skepticism turns to curiosity and then becomes a maniacal obsession as he studies the journals as a blend of scientific and magical scripture, ultimately recommissioning the laboratory in his quest to reanimate a corpse.  

The Monster in Young Frankenstein does not come to life with the power of speech; the comic plot dispenses with the murders of the 1931 version to focus instead on the gradual socialization of the ungainly figure who, due to the servant Igor’s bumbling, has Abby Normal’s brain (an abnormal brain stolen from a morgue).  Frederick, having accomplished his grandfather’s mad dream by creating life from dead tissue through the power of electricity, declares “My name is Frankenstein” after pronouncing the name as Fronkensteen to disguise his ancestry back in the United States.  Rather than assigning the Monster menial tasks to do, he turns his attention to making a place for his creation in the contemporary world (such as it is in Transylvania, which seems rooted in a past era).  To convince the villagers and the local police chief that they have nothing to fear (they have bad memories of what previous Frankensteins have done in their castle), Frederick teaches the Monster to perform as a “man about town.”  The two of them, dressed in top hats and tails, offer a rendition of “Puttin’ on the Ritz” at a local music hall.  The Monster squeaks out the song title and manages some rudimentary dance steps.  Here’s the scene: https://www.youtube.com/watch?v=IQydibRWjPw. Unfortunately, a stage light explodes, frightening the Monster and turning the formerly appreciative audience against him.  Frankenstein, mortified, transforms from supportive parent to impatient creator, ordering the Monster to obey.  Here we see a parable related to both the Golem legend and to the risks of science that overreaches as well as to the dynamics of the human-Golem relationship.

The Monster eventually gets the power of eloquence from Victor via a transfusion of brain fluid, which allows him to give this speech to the villagers intent on destroying him:

“For as long as I can remember people have hated me.  They looked at my face and my body and they ran away in horror.  In my loneliness, I decided that if I could not inspire love, which was my deepest hope, I would instead cause fear.  I live, because this poor, half-crazed genius has given me life.  He alone held an image of me as something beautiful.  And then, when it would have been easy enough to stay out of danger, he used his own body as a guinea pig to give me a calmer brain and a somewhat more sophisticated way of expressing myself.”  

This version of the Monster can join society and marry (in the last sequence we see him married to Victor’s former betrothed Elizabeth, who cosplays as the Bride from the old Frankenstein movies, while her domesticated husband wears glasses while reading a book in bed).  This Golem, blended with the creator through science and soulful mysticism, attains full human status.

Golem Agonistes

We see another version of this transformation achieved through a combination of programming and interactions with humans rather than brain fluid transfusion in the Terminator films, a franchise that began in 1984.  For this section I will focus on the three films featuring Sarah Connor played by Linda Hamilton and the T-800, the Edsel of Terminators, played by Arnold Schwarzenegger, leaving aside the several films without Hamilton’s Sarah.  The T-800 in the first film, a metal skeleton encased in living tissue to move easily among humans in order to kill them, has the single purpose of eliminating Sarah Connor to prevent the birth of John Connor, future leader of the resistance against the rogue technology Skynet, a system developed by the military to protect the US from an enemy attack.  Sent by John, soldier Kyle Reese comes from the future to protect Sarah; he accepts the one-way mission because he’s fallen in love with her from seeing a photograph of her that John has shown him.  Together, Kyle and Sarah defeat the T-800, Kyle giving his life in the fight, though not before their one night of passion leads to Sarah’s pregnancy, which results in John’s birth (love those time travel plots).    

In Terminator 2: Judgment Day (1991), we encounter the advanced T-1000 made of liquid metal, with the ability to take the shape of any human it touches as well the ability to wield body parts as weapons.  This Terminator’s mission focuses on killing the teenage John Connor, born after the pregnant Sarah Connor survives in the first film and flees to Mexico to give birth and to learn how to fight the future.  The reprogrammed T-800, again sent by future John Connor, returns with the task of protecting teenage John, but this version represents a different kind of advance than the T-1000; the T-800 now moves through the world as a learning computer, curious as a child always asking questions.  It asks John about why people cry and learns slang expressions and some Spanish phrases from his charge, becoming both a dependable father figure and a dependent child in need of moral guidance, e.g. when an exasperated John makes it promise not to kill anymore because it’s “just wrong” (not self-evident to a Terminator).  Sarah, noticing the T-800’s staunch commitment to her son, muses that the machine is the most reliable father that John has ever had.  Ultimately, after defeating the T-1000, the T-800 sacrifices itself to avoid its circuits being used to create more Terminators; as it comforts the sobbing John, it says it now understands why humans cry before Sarah lowers it into a molten bath (as a Terminator cannot self-terminate).

In Terminator: Dark Fate (2019), the T-800 becomes a family man named Carl with a steady job installing drapes; he also reports that when his adopted son was a baby he changed diapers without complaint, that he’s a good listener, and that he has an excellent sense of humor.  Note the development from surrogate father figure and learning machine to full-fledged parent and partner (minus any physical relationship with his son’s mother, as this Golem has no libido).  Most crucial in the T-800’s evolution, though, is what he chose to do for Sarah Connor after fulfilling his mission of killing her son John (we see this in CGI-enhanced flashback).  By this installment the T-800 has taken on a human name and transcended his programming by developing a conscience, an awareness that allows him to realize what he’s taken from Sarah by killing John.  By choice born of empathy, he gives her a sense of purpose, with some irony turning her into a terminator of terminators, single-minded in the missions she undertakes when Carl anonymously sends her geographical coordinates for where terminators will appear; when we first see her in this film, nearly thirty years after the end of T2: Judgment Day, she voices the iconic “I’ll be back,” the Terminator tagline all the way back to the first film.  

When Sarah and Carl team up with the augmented human Grace (sent from the future, natch) to protect Dani Ramos, future leader of the resistance; Sarah initially assumes Dani will give birth to the future leader but eventually learns the truth and by film’s end becomes a full-time mother figure for Dani after Grace’s death.  Thus the third film features a set of mother-daughter relationships after Dani’s father and brother are killed, with Sarah acting as mother to Grace and Dani and ultimately taking on Dani as a daughter to replace her dead son, with both of the children destined to save the world from machines.

Carl notes that they can kill the latest version of the Terminator because of the machine’s limited imagination despite all the technological advances, as a Terminator sets aside all other considerations due to a relentless focus on killing the target.  Certainly the various Terminators, from the liquid metal shapeshifter of T2 to the equally lethal REV 9 shapeshifter with an independent exo-skeleton in Dark Fate, have some sense of how to navigate the world by changing appearance and imitating voices, but they often reprise the first Terminator’s rampage through the police station searching for Sarah Connor; in Dark Fate the Terminator rampages through the immigration detention center in search of Dani Ramos, indiscriminately killing detainees and guards alike.  While both T2 and Dark Fate feature Terminators who take on the trappings of law enforcement, in one case a motorcycle cop and the other an immigration agent, both automatically kill bystanders and various types of police officers who stand between them and their target.  Neither has a teenage John Connor to teach them that killing is wrong; their inability to learn leads to their demise.

How do Terminators die?  I remember reading a movie review of the first film stating that Arnold Schwarzenegger needed to become the steel skeleton with red eyes and creaking, segmented gait for the audience to accept his “death”; the evolution through the following films, which portray the increasingly famous actor becoming progressively more human, allows the audience to accept the actor dying in human form, as a final act of sacrifice, in T2 to keep Cyberdyne Industries from building a new terminator from his circuitry and in Dark Fate because his death makes possible the elimination of the threat to Dani–in addition, Sarah Connor remains as a fully-trained replacement to protect Dani from further threats, one last bit of franchise irony in what will likely be the last installment starring Hamilton and Schwarzenegger.  In the first film the T-800 dies as a machine, while in succeeding films the humanized original Terminator dies as a martyr to save humanity. 

Golem Talk

To return for a moment to Scholem’s musings on the lack of speech distinguishing the Golem from the human in the context of the Terminator films, first consider the T-800’s limited language in the first film; while the T-800 does reproduce Sarah’s mother’s voice to deceive her over the phone, he speaks little aside from the “I’ll be back” tagline at the police station after he’s inquired about Sarah Connor at the front desk.  Other than that, the T-800 has no voice.  In the second film, the T-1000 can imitate voices, change shape to appear as anyone, and occasionally speak independently, taking cues from context as a limited, purpose-driven machine; in contrast, as noted earlier the T-800 talks a lot, learning from John, providing exposition to Sarah about how he came to have changed programming (it voices Kyle’s first words to Sarah from the first film, “Come with me if you want to live”), and delivering the history of Cyberdyne Industries to Cyberdyne engineer Miles Dyson to persuade him of the need to stop work on the first T-800’s remnants and destroy all the documentation that could lead to Skynet’s creation.  It’s become quite the schmoozer, a trend that continues in Dark Fate, in which Carl explains his own evolution as well as the reasons for his communications with Sarah.  Here we can see dramatized renderings of Scholem’s linking of speech to human reason and logic.  While Terminators generally don’t speak when in killing mode, the T-800 becomes a rhetorical being, speaking to engage in relationships and make sense of the world.  

The Golem Within

In colloquial Yiddish, the term “golem” usually refers to a clumsy or slow-witted person, drawing on the Talmudic meaning of the golem as a pre-sentient lump of clay.  Anyone slow on the uptake or socially inept might become the target of this dismissive if not derisive label.  If we consider Henri Bergson’s dictum that social life requires from each of us a mental nimbleness that allows us to respond quickly and appropriately to ever-changing situations, we should also entertain the realization that all of us carry a golem within, ready to expose itself at any moment, opening us up to ridicule and disdain.  We see this revealing of the inner golem repeated incessantly in movies and television shows, particularly comedies, as the mechanical behavior of characters without sufficient ability to comprehend and respond to the needs of others provides a deep well of comic possibilities.  

Let’s use an analytical lens we can call the Golem Quotient (GQ, also the name of the most popular Golem-serving glossy magazine) to assess how the comic Golem character turns up consistently throughout sitcom history, though we’ll have to limit the number of examples.  Many well-loved sitcoms of the 1980s insulate their main characters from high Golem Quotient scores but include at least one high-functioning Golem, such as Ted Baxter on The Mary Tyler Moore Show, Coach and then Woody Boyd on Cheers, and Frank Burns on M*A*S*H.  We get to higher GQ scores in the 1990s when we consider the “no hugging, no learning” mantra of Seinfeld, a rejection of the sentimental dimension of family-oriented or workplace-family precursors that neatly resolve conflicts and restore domestic equilibrium; this commitment also makes the four central characters, Jerry, George, Elaine, and Kramer, into urban Golems incapable of growth. Larry David, co-creator of Seinfeld, follows up with his Golem-centric early 2000s series Curb Your Enthusiasm, in which David plays a heightened Golem version of himself, unable to restrain himself from calling attention to and pontificating about any little thing that doesn’t align with his programming.  Most recently, in honor of the 250th anniversary of the United States, David developed and starred in Life, Larry, and the Pursuit of Unhappiness, co-produced by Michelle and Barack Obama, a limited series plopping the familiar David Golem into key moments in the country’s history, revealing what the history books skip over regarding the great Golems of history.  

The Big Bang Theory, which I wrote about at length along with Chuck some years ago, serves up perhaps the most complex example, charting over twelve seasons the physicist Sheldon Cooper’s journey toward becoming a “real boy,” growing beyond the golem unable to play well with others.  Sheldon consistently struggles to understand human emotional experiences and behaviors.  One running joke involves his attempts to understand sarcasm, leading him to ask friends “Sarcasm?” after hearing a jibe that he can’t be sure about; an entire episode chronicles his attempt to develop a comprehensive theory of comedy (perhaps the only sitcom script to cite Henri Bergson’s theory of the comic situation depending on “the mechanical encrusted on the living,” an irony Sheldon doesn’t have the capacity to see in himself despite the exasperation of all of his friends and colleagues dealing with his mechanical habits and obsessive-compulsive behavior).  

In sum, Sheldon has trouble understanding human emotions at all, in one episode carrying around a prototype of an emotion detector developed at MIT to tell him whether his friends are “sad,” “happy,” or “angry.”  His eidetic memory intensifies his golem-like qualities, making him something of a human large language model able to recall everything while not generally being able to connect the data to which he has access to develop and act upon a coherent approach to living his life in society, whether at Cal Tech where he works or in the apartment he shares with Leonard, his long-suffering best friend.  Sheldon dreams of uploading his intelligence to a computer body; he admires Mr. Spock for his devotion to logic while denying neighbor Penny’s pointed observation that Sheldon has a human half, just like his idol, which he struggles to suppress.  Unlike the mute Golem of legend, Sheldon speaks without pause, mostly to spew facts he wants to display to correct people and show off his intelligence.  While the show’s ensemble includes a quartet of nerds and their partners, friends, and colleagues, with all of them living through significant changes over the twelve years, Sheldon’s evolution toward human behavior, including empathy for others (a comic variant of the T-800’s journey to compassion for Sarah’s grief over losing her only child) serves as the show’s thematic throughline, though most of the laughs along the way mock Sheldon the Golem. 

Whither the Golem?

Given the vertigo-inducing pace of technological advances in the sixty-plus years since Scholem’s dedication of Golem Aleph, the Golem isn’t going anywhere.  Technology companies will continue to flood the market with new and improved models of Golem from aleph to zed, striving to make them ever more integrated with and perhaps indistinguishable from their human users, while popular culture will continue to generate stories to help us understand our relationships to these ever-present companions as well as to understand how the technology can efface our humanity, allowing the inner Golem to emerge and take control over what we considered inherently human activity.

In the realm of politics, in earlier times a matter of voters attending town halls and candidates sampling the local cuisine at diners and state fair, we see technology increasing anxiety about what we can believe.  An old joke states that one knows politicians are lying when their lips are moving; now, in the age of deepfakes and doctored video and audio, that cynicism carries over to all representations of politicians.  The British television series The Capture focuses in its first season on the military creating fake video records of field operations to hide embarrassing facts about soldiers’ actions and government policies.  Season two shifts the focus to deepfake videos of politicians saying and doing things they never said or did, with the false documents serving the interests of intelligence agencies and the military.  Season three, written and produced as AI came to dominate the news, depicts the British military giving authority to issue orders over to an AI named Simon, with the orders carried out by a rogue squadron of soldiers; Simon’s orders include assassinating a leading politician the AI predicts will become prime minister, posing a threat to the military and the security of the country.  When intelligence chief Rachel Carey, the protagonist throughout the series, figures out what’s happened, she confronts the general taking orders from Simon and deploying the rogue soldiers.  He defends his choice to cede authority to the algorithm-driven Simon because he’d rather have the AI make decisions than risk being wrong himself because he doesn’t know everything Simon knows.

In the domestic sphere, in recent years two television series on Apple TV (also the home of Ted Lasso and Shrinking, traditional workplace family sitcoms that keep technology and AI at some distance from the familiar relationship-oriented concerns of the genre) have provided narratives that explore how technologies might address issues of work-life balance and interpersonal relationships, respectively.  Severance derives its premise from a technological advance developed by Lumon Industries that allows employees to create an “innie,” an alter ego Golem that goes to work on behalf of the “outie,” who lives a separate life with no memory of the work hours at Lumon.  The severed thus achieve work-life balance by using technology to send the Golem within to work every day.  The morning elevator ride down to the Lumon office space Golem-izes the outie, leaving the drone innie without memories or any nagging emotional baggage that might distract them from the work before them.  Though Severance doesn’t quite qualify as a workplace comedy, the main work team made up of four members sharing an office displays a kinship with other workplace groups.  They get on each other’s nerves, gossip about management, form and break alliances, and ultimately come to rely on each other.

Pluribus gets its premise from alien technology come to earth.  We can think of the series as an updating of the sci-fi classic Invasion of the Body Snatchers, first filmed in 1956 after Jack Finney published the novel, first in serialized form in Collier’s in 1954 and then in book form in 1955.  In that earlier classic, spores from outer space waft down to earth to fictional Santa Mira, California, a pastoral small town (replacing Marin County’s Mill Valley, the setting of the novel).  The spores have the power to create duplicates of any nearby human, growing the identical body in a pod that ripens and releases the Golem version–the duplicates can speak and have the same memories as the original, but they do not have emotions and cannot feel love–they inspired the term “pod people” after the film became one of the celebrated sci-fi movies of the 1950s.  The Pluribus pod people are not duplicates of humans but rather humans transformed by an alien virus into placid members of a worldwide hive mind–think of the Borg collective without the mechanical enhancements or goth-punk Halloween wardrobe.  

Romance writer Carol Sturka, the human protagonist, has immunity from the virus, one of only thirteen humans on earth unaffected by the mass Golem event.  She finds the calm, frictionless new world abhorrent and wants nothing to do with the hive mind, despite their equanimity and boundless willingness to provide her with anything she requests, and even what she doesn’t request, as when they send her a companion who looks exactly like the illustrations of her romance novels’ lead character.  Carol, a curmudgeon to begin with, values her privacy and more than that her independence of mind, a quality the hive mind members have given up in exchange for peace, communal collaboration, and access to all of the knowledge (facts, languages, experiences) of billions of virus-transformed humans.  In contrast to the innies of Severance, the vast hordes of Pluribus lose their individuality to the collective rather than during a daily elevator ride to their cubicle.    

Our Golems, Ourselves

In his exploration of the Golem legend, Scholem addresses the human trait of trying to emulate God by creating life, with motivations ranging from pious devotion to hubris in line with Prometheus stealing fire.  As we end this foray, I offer the possibility that the externalized Golem (of Prague or of popular culture) speaks to us about both our fears of nature or technology acting beyond our control, putting our lives at risk, and our fears of the internal Golem we carry with us throughout our lives defining our identity, putting our souls at risk.  Spiritual practice and everyday life alike challenge us to create ourselves, to become more than our Golem selves.  From this perspective, those who promise the imminent arrival of artificial general intelligence aim to grant the technological Golem the human imperative to grow and learn, to transcend the limits of the Golem who can only follow orders and cannot speak.  The articles I listen to online, produced by NOA (News Over Audio) and delivered by an AI voice, let me know that this Golem has learned to speak, though for now the AI Golem only parrots words written by the human named in the byline, much as chatbots powered by large language models parrot diction and syntax from their training texts to give the illusion of a writing, thinking consciousness responding to the master’s prompts.  Both strivings continue, as the technological Golem attempts to become more human and the human Golem does the same.