Dreams from the Witch House: Female Voices of Lovecraftian Horror

by Joyce Carol Oates

  Dead again. This cycle has been the longest so far at two and a half hours. It was the predators who triumphed this time, if such a word can really be used. The seagull population exploded some time around midnight, killing off the sea fauna. This echoed up the food chain until the snow foxes depopulated the birds below breeding levels, and then died out themselves. In the silent aftermath, the seaweed grew uninhibited until eutrophication set in. All that remained then was the six, blank white squares of dead snow. Maybe permanent balance is a biological impossibility, like the perpetual motion device rendered in flesh. If this is life in fast-forward, then perhaps this is a bleak vision of things to come for us.

  At approximately 3:00, there followed an unexpected addendum: No, just one of an infinity of potential futures. This peculiar digression into Leibniz seems to have accompanied a renewal of faith on Clarke’s part. Some time during the days that preceded her next entry, she appears to have successfully used her model to create her first self-sustaining ecosystem. She writes:

  Seventeen hours now. Until now, the longest it has ever lasted has been three and a half. The cycle is endless and beautiful. There is something timeless and ancient about such a pattern, and it is no longer any wonder to me how so many of the world’s oldest religions had this as their principal focus. Even so, there is something off about it all.

  Later, she adds: Something is still wrong. I fear it is not the program’s function, but its very being. I delight in the cold beauty of a world reduced to first principles; a lifetime in mathematics has invested me with this. But now the world of numbers seems like a prison, into which I have simply created a window. Now entities cry out at me through the bars, their voices mute, baffled. This is Hades, a death world, in which shades of that which once was living now dwell in mindless repetition of what their lives were before. It was I who brought them here.

  Eugenia Clarke’s philosophical disaffection at the completion of her initial project marks the beginning of a curiously silent period in her diary. It is unclear whether any more use was made of her model, but the few sporadic entries appearing during those two weeks deal with rather more mundane concerns. It is hard to be certain what these entries portend, but what they certainly do not indicate is a descent into madness.

  Whatever new thoughts were now resolving themselves in Eugenia’s mind is something we may never ascertain. However, witness accounts indicate that almost all of this time was spent in the library. She seemed to borrow rarely, but it is assumed that much of what she later references in her work was read during this time (no small feat, considering the breadth of her research). Given the nature of her work, it was only a matter of time before such names as Turing, Searle, and Grey Walter would come up in her discussions. Yet her writings after this time also indicate a more esoteric leaning to her tastes. Plato’s Cave is mentioned, along with the theories of mind of Rene Descartes. In time, even such curious luminaries as the Pseudo-Dionysus are quoted at length, while particular attention is paid to the more mystical works of Pythagoras. Whatever she sought, it was evidently beyond the scope of what modern science could offer.

  Yet in spite of what would ultimately transpire, Clarke’s initial return to work was a curiously subdued affair, indicating little of the changes that would later become apparent. It was early in her writings that she first suggested a human addition to her project, but at the time this was little more than a speculation. Inspiration for this had first come from a postscript to Dr. Loughbridge’s book, which told of a hunter whose job it was to help maintain wildlife populations. Loughbridge tells of how overfishing along the coast had driven bird populations into the fjord, disturbing the fragile balance. The hunter’s task was to cull animal populations to a quota, thus ensuring a return to equilibrium. On reading this, Eugenia was at first horrified. Slowly, however, she began to understand its necessity within the implacable logic of the fjord’s brutal daily realities. On her return to the project, she resolved herself on this addition, which would essentially take up the role hitherto played by the user in her simulation. Thus, a hunter was born and his name, she decided, was Jotun.

  It is not clear when her first conscious steps into the sphere of artificial intelligence were taken, but her first explicit mention is one made in retrospect, and coincides with her decision to introduce her model hunter. It came upon first reading about the Grey Walter tortoise, one of the earliest examples of applied robotics. One particular aspect of its description seems to have struck a chord with her, as she writes:

  One of the most remarkable qualities of this little device seems to have come almost as an afterthought. It appears that it is able to detect when it is reaching the end of its battery life, and limit its movements accordingly. This is, if anything, a demonstration of reasoning behavior, if not feeling, in its most rudimentary form. If something can react, then it can surely learn, adding another dimension to its reasoning. I realize now that if I’m to make this final addition to my project, then it must operate on a level separate from the rest of its kin: a program within a program. While the fish and foxes operate according to the fickle logic of algorithms, our hunter must have true reasoning; it must have intelligence.

  The two-week long period of silence that heralded this second phase of her work is followed by a number of marked stylistic changes. Whereas before, Clarke had been rigorously thorough in detailing the minutiae of her coding processes, now only her most significant accomplishments were recorded, with little mention of the intervening work (something that has proven no small annoyance to critics keen on establishing the veracity of her accounts). Instead, the practical business of programing now took place against a backdrop of agonizing deliberation over the more theoretical aspects of her work.

  One of the greatest misconceptions about Eugenia Clarke’s legacy is that she sought to create an AI program with a soul. Though she does give thought to this possibility, at no point did she ever propose bringing such a thing into existence. Nonetheless, in one sense at least, this seems to have been exactly what she would eventually achieve.

  The first stirrings of theosophical thought in Clarke’s recollections stemmed from a profound disagreement with one of the leading contemporary theories on artificial intelligence. Known as the hard-AI principle, it asserts that on a fundamental level, there is no difference between a complex computer system and human consciousness. This works on the belief that, when deconstructed, what we understand as the human mind manifests as an infinitesimally complex series of logical deductions to derive the appropriate response to a given piece of data. Aspects of the mind, such as emotions or memories, are thus understood as units of this process, designed to facilitate thought. This means that, like a computer, the human mind is on no specific level truly conscious in any real sense. It merely thinks it is. A popular analogy for this process is the Chinese Box theorem, in which an individual sealed within a room (representing the human mind) is anonymously handed slips of paper containing Chinese characters with which they are unfamiliar. They are then charged with matching them to a corresponding set of symbols elsewhere, with no ultimate comprehension of their meaning.

  Clarke’s dispute seems to have arisen more from a profound discomfort with the implications of this theory, rather than any logical counter-argument. It may have been that, should this really be the case, then the vision of a death world, which had so horrified her when seeing her program take form, may have been closer to reality than previously supposed. In response, she seemed to have struck out in search of a theory of her own, and would plunder the writings of authors of whatever discipline she could find for ideas that would support her own. In doing so, her project was rapidly changing from one of science to one of faith.

  Her central belief during this period appears to have been that the logical processes that comprise human thought, when gathered in sufficient complexity, take on a character and substance that is wholly unique, and greater than the sum of its parts. This was expressed through a multitu
de of rephrasings and extrapolations that fill the pages of her diary in these sections. Curiously, it was by her own admission that she observed, with apparent humor, that this principle was the same as that underpinning the ancient science of alchemy.

  Her thinking during these researches is marked by a singular refusal to accept that the minds of animals, even ones of demonstrable intelligence such as apes and whales, are in nature and substance fundamentally the same as those of humans. This is a point she seems to emphasize with an uncharacteristic vehemence. Indeed, it is during one of these recurrent episodes that she makes her first explicit mention of the soul. She would continue implementing this term, and other theological terminology, in an equally unguarded manner for the remainder of her endeavors.

  It is a fact to be appreciated that Eugenia Clarke would eventually find no satisfactory resolution to her problem with hard-AI. What she resolved to do instead was at once a display of her formidable cleverness—and her now evident madness. This resolution would, in time, redeem her faith in the uniqueness of human thought. What she proposed was not to create an AI with a soul; by her own admission, this was impossible within the bounds of modern science. What she would do instead was use her project to create an AI that simulated the actions of a soul-bearing entity. In doing so, she would prove the ultimate superiority of her ideas over the tenets of hard-AI by making her soul-creation the first program ever to convince another human of its own humanity, thereby passing the fabled Turing test.

  However, what Clarke initially proposed as a realistic alternative to an understandably impossible ambition would in turn prove an equally absurd task. Nonetheless, as she set out her specifications for what would define a human with a soul, her confidence in her abilities to mimic one was unfaltering. Clarke’s definition for the human soul is spread out across her work in a series of digressions and internal discussions around an ever-changing theme, drawing on Christian and pagan sources alike. What she settled on seems to have been somewhere between Pythagoras and St. Augustine. Perhaps her most lucid definition is one written on an evening in late July:

  Souls and minds are at once a single entity and two separate bodies. The soul is moved by the rigors of divine providence, and in turn affects the predilections of the mind. Yet the mind is possessed of free will, and has the power to unconsciously influence the path that is chosen for its counterpart. Mind-stuff and Soul-stuff are two different substances, but this is not their main distinction. What truly differentiates the two is that while both are reasoning agents in their own right, minds reason in time, whereas souls reason in eternity. The soul is thus the mind’s gateway out of time and into the realms of divine providence.

  Despite its elegance, scholars of theology have been sparing in their credit for Clarke’s proposal, with many dismissing its arguments as reductive. She was, it is said, paring down the writings from which her work was derived to their most basic elements, and turning what were ultimately moral ideas into a crude logical framework. Yet while the definition itself may not have provided much of a framework for considering the ultimate question of the soul, it was nonetheless sufficient for Clarke’s purposes. With these ideas set in place, she now had working model upon which to base her computerized analogy for the soul-bearing creation. It was to be, in essence, an entity whose actions were neither their own nor wholly the product of their creator, but driven by some ever-changing essence of its own being. Whether Clarke willed it or not, this would remain an apt definition for what Jotun was eventually to become.

  Just how successful she was in realizing these curious ambitions is hard to say—not least because the whole of this second incarnation of her program appears to have been deleted within the two weeks following its creation. From Clarke’s diary, we know what the hunter looked like (a rather unimpressive anthropoid form pasted into a diary entry in late August) and that it seems to have been capable of speech, at least through a number of stock messages. After six weeks of strenuous hacking, it appears to have been completed. She would then spend much of the subsequent days marveling at her creation. She writes:

  I watched him resume his lonely vigil, walking the ice-blasted rocks and dark shores of his domain in quiet contemplation. He is patient. He knows the land and is satisfied with his simple lot in life. Furthermore, he cares deeply for the birds and the foxes, and the manifest creation he sees all around him. He is like Francis of Assisi, or the ascetic first fathers in the deserts, or even like Adam wandering the forests and valleys of paradise. I asked him, why does he hunt? He recognized the question, and gave his response in blunt and predictable terms. I told him he should shoot a gull in the northernmost corner of the fjord, which I indicated. He refused, not saying that he wouldn’t, nor couldn’t, but merely should not, even though it would not (I knew) affect the overall balance either way. He said it was simply wrong. Later, once the seasons changed and the cycles adjusted, he shot that very bird, leaving its chicks defenseless. It had, it seems, become acceptable that this bird die.

  Critics have identified the events of an evening in late July as the major turning point in the already doubtful soundness of Eugenia’s mind. It appears to have come at around 22:00, when she began to experience difficulties accessing certain files on her hard drive. Fearing that all her work of the last few months might somehow be lost, she copied the entirety of it onto a collection of disks in order to transfer them to the library computers. The events of that late-night exodus onto the university campus are sketchy, but she seems to have uploaded her program with little difficulty. It was dawn when she eventually returned home, having spent the night in silent communion with the fjords. This would prove to be only the first in a series of strange incidences.

  Further changes have been identified in her writing at this time, wherein she first began to talk about dreams. However, these discussions would veer more towards the concept of dreaming itself rather than detailing any of her own. This is understandable in context, as the nature of dreaming is a significant consideration to any study of the human mind. Yet her new preoccupation with nocturnal visions brought with it a troubling ambiguity for readers, uncertain now as to whether many of the more surreal details that followed were actually true.

  Critics have attributed this apparent departure from reality to the fear experienced when problems started to emerge with her program. Such was the attachment she now felt with the program that, it is said, the shock at seeing it malfunction must have been like the loss of a child, the pain of which could only be assuaged by giving it a latent fantasy life. This often came with the assertion that the malfunctioning of her own computer had in fact wiped the Jotun program entirely, and that its revival in the library had never really taken place. Yet if this were truly the case, then it was hardly a merciful fantasy she would eventually create for herself.

  Returning to the library after a break of two days (during which her diary is suspiciously quiet), Clarke found the program once more behaving strangely. She writes:

  I fear there may be another problem with the program. I had thought at first it may be a virus, as I suspect was the cause of the original problems (a reboot of my home computer appears to have sorted this out), though I’ve never known a virus to act like this. Life in the fjords is as active as ever. Seaweed abounds, the fish are lively, seagulls wheel through the skies and foxes are ever vigilant, but Jotun is just standing there, his gun limp at his side. I assumed this may have been some unusually perfect moment of balance, which called for no further action on Jotun’s part, but it seemed to just go on. No equilibrium can last forever; my algorithms won’t permit it. Indeed, he remained just so, even as the seagulls began to exhaust the fish and rotten seaweed pollute the water. I issued the order to kill a seagull and he duly complied, but did nothing more.

  Later that week, she describes how Jotun finally broke his spell of inactivity and embarked upon the wholesale slaughter of the snow-foxes. In the absence of predators, the seagulls lay waste to the fish populatio
n, exterminating every last one. Though this was a known outcome of several of the potential situations Eugenia had predicted, what happened next surprised her. Without their primary source of food, the seagulls set upon one another. Cannibalism was not something she had programmed, nor, however, was it something at odds with their basic behavior. Thus, in the absence of any pre-determined response, their own instincts were the same as their hunters. Eugenia did not wait to see the gruesome outcome, though she noted that Jotun had resumed his previous state of repose.

  It is difficult to be certain whether these had been the last such incidences of strange behavior. Passing references hint toward at least two further occurrences, but how they compare to those of the previous entries can only be surmised. In the wake of the second incident, Eugenia’s diary becomes frustratingly vague. It is clear that she attempted no further programing work, or any of her previously energetic researches into philosophy of the mind. She was evidently reluctant to dwell on the anomalies, perhaps afraid to give too much thought to their implications. But in light of what she would later describe, it is during this time that her understanding of what she was seeing began to change.

  Initially, she saw the anomalies as mere malfunctions: errors in the program causing conflicts of interest that resulted in irrational behavior. Eventually, however, she considered the possibility that Jotun had not simply been programmed incorrectly, but was somehow consciously working to alter its own coding in ways Eugenia no longer understood. She had mentioned earlier that she had given Jotun the capacity to learn, but indicated that this would be limited to the information relevant to his task. What was happening now was something wholly unprecedented and there seemed no limit to what it could become.

  This period of fearful inertia was broken on 28 August. It seemed that Eugenia had once more resumed her habit of taking long walks. That morning saw her wander beyond the city reaches and into the grounds of the medieval abbey at Newstead. There, among those ancient stones, she seems to have found new resolve. That evening, at around 21:00, Eugenia Clarke recorded the following entry in her diary: