Introduction

At the turn of the 20th century, a book appeared bearing the improbable title From India to the Planet Mars: A Study of a Case of Somnambulism with Glossolalia. It was Théodore Flournoy, professor of psychology at the University of Geneva, who published his five-year research on Miss Hélène Smith - a trance medium who claimed to have clairvoyant powers, to be able to communicate in both Sanskrit and the Martian language, and to manifest her past reincarnations as Marie Antoinette, a Hindu princess and a visitor to other planets. The present rational sceptic will probably stop reading here. How can a study of a trance medium from more than a century ago be relevant today? Flournoy's subject, the medium he called Hélène Smith (her real name was Catherine Élise Müller), became a paradigm psychological case history of multiple personality, and was also analysed in part because her automatic verbal productions were so complete, that the father of modern linguistics Ferdinand de Saussure - who was assisting his colleague - considered her a perfect model for understanding the generation of natural language. The study was controversial, but the involvement of these two prominent scholars inevitably opened up science to a "new dimension", what Flournoy called its "third path" to overcome a systematic binary contradiction, including Saussure's binary conception of signs (signifier/signified).

Today, intelligence is also considered to be "multiple". Generative Pre-trained Transformers (GPT) are a prominent framework for generative artificial intelligence (AI), and their apparent mastery of natural language has sparked controversy. Following Flournoy-Saussure's “improbable” paradigm of identifying and conceptualising the subject matter when encountering marginal states of consciousness, Hélène's curious case is used as both a metaphor and a method for overcoming our binary contradiction of intelligence as either “natural” or “artificial”; a quest perhaps more urgent now than ever. In this view, Hélène's ambiguous Martian writings are fed into GPT, and a meditation with speculative moments is attempted through human-machine inter-written texts, enacted through inter-twined speeches that reciprocally represent and interpret their own transitive nature. In this paper, the question of locating intelligence will not be a question with a binary answer. It will be shifted across multiple sites in an assimilative assemblage, exploring how identification might work from a rather metabolic side of the conversation.

Writing as method (Laurel Richardson 1994, 2000; St. Pierre, 2005) and diffraction (Barad, 2007, 2014; Haraway, 1997) as a process for thinking-with complexity about what is visible, recognisable (and speakable) and under what conditions, will navigate this research. For Barad, diffraction allows us to read “insights through one another in attending to and responding to the details and specificities of relations of difference and how they matter” (Barad 2007, 71). From this perspective, the research uses diffraction to propose an understanding of difference that is no longer dualistic but moves beyond binaries, but also to open up new ways of addressing the challenges it poses as relational and immanent. The output created as part of the Speaking Martian exposition is also diffracted through the AI-generated text. In this research I aim to achieve a thinking-with-otherness, writing, speaking and reading together with machines as a human-AI co-creative inquiry that does not capture meaning but produces it, enabling an understanding of subjectivity and the way in which this subjectivity is structured and produced in social encounters. Both generative and challenging, this process can help me navigate beyond intentions, uncovering new subjectivities, shifting variables or even contradictory insights.

The use of new materialist methods to explore synergies between the study of the medium Hélène Smith at the turn of the 20th century and contemporary notions of subjectivity, artificiality and intelligence in the age of AI is also intended to creatively counter a critique that sees these links as unscientific, irrational and indecisive, and at the same time to invite means of experimentation for knowledge production, especially in the rapidly changing context of AI - a context that not only demands a multiplicity of ideas but already produces a new synthesis of reason within its nature. Bridging the time between centuries and the migration of intelligence from human to technological “mediums” to articulate the unspeakable of natural language “artificially”, this encounter between Hélène Smith and GPT is provocative in itself. But an even greater challenge is bridging the mindset surrounding the Flournoy-Saussure study with the present developments and controversies that AI speech phenomena are generating in terms of how they are (not) being studied today. Highlighting the involvement of prominent scholars with the scientifically controversial subject of spiritualism along with research on channelling voices through mediums and electronic voice phenomena (EVP) with the present functions of GPT offers an invitation to re-think across disciplines and re-position research in ways that step out of comfort zones and challenge fixed frames of reference. The humility and empathy that emerge from this approach can be seen as a source of re-claiming responsibility for critical socio-political issues that have a far-reaching impact on our way of life today. In this vein, I use diffraction rather than reflection to chart the “how” of the relationships between the past case of Hélène Smith and the present practical experiments in AI, to be produced by, with and within processes, subject, researcher(s) and context, not as separate entities, but rather as parts of the intra-action of enacting the phenomena under study. Barad's paradigm of agential realism understands the researcher as part of the assemblage in which the phenomenon under study emerges in the first place. This process challenges common frameworks of how we know what we know, acknowledging entanglement and accepting responsibility. In discussing diffractive methodology, Barad asserts that “the point is not merely that knowledge practices have material consequences but that practices of knowing are specific material arrangements that participate in reconfiguring the world.” (Barad, 2007, 91).

In this sense, rather than juxtaposing Hélène Smith's curious speech phenomena with those of GPT (one against the other), I trace immanent relations (one with the other) between media-mediated meaning-making re-arrangements of intelligence and its articulation through language that would allow me to identify with the otherness of the subject, texts and voices (post-mortem/non-human/artificial), and their performance and function, by planting doubt in the fertile ground of knowledge, both in terms of origin and data (fact-fiction), mediums (human-machine), and also outcome (unrepeatable-verifiable). Bearing in mind that what intelligence is is always already open to interpretation, in this research I aim to undermine how the humanist concept of reason - which was used to firmly demarcate human, natural intelligence from artificial, machine intelligence - is being remade by machines and redistributed to machines; and also how the concept of subjectivity is being undermined in the modelling of intelligence, leading also to the limit of intelligence and, therefore, imagination.

This paper can be seen as an entanglement of generations of studies, mediums, displacements (past and future, author and subject) and the places (bodies) occupied and traversed along the way, while shifting relations with ourselves, with people, machines, mechanisms and intelligence(s). In this vein, it is an assemblage of entities, curiosities, personalities, transversalities, memories and mediations, captured and re-composed to make sense differently, or at least to make sense of their entanglements in relation to changes in time.

When GPT Digested the Medium Hélène Smith

Approaching an Ambiguous Subject Ambiguously

Displacing reference

In this research, the subject revolves around a dialectic of speech and embodied forms of subjectivity, attempting to capture it in the dimension of its procedural creativity that allows it to remain creatively ambiguous. Rather than an (id)entity whose originality or artificiality can be presumed, the subject unfolds within an equivocation that circulates around the I that lies behind the speaking voice. Rather than relying on inductive assertions, this project offers a shift from theorising, anticipating the very act of feeding-in and feedback into its process, to ultimately relate its logical, morpho-logical, techno-logical, onto-logical or para-logical grounding. Positioned as a speculative complement to the detached, free-form thinking of a subject that seeks to escape the categorisations of embodied specificities or nominal particularities, the capacity to provoke ambiguities in its productive conception can thus become palpable. Through this process, a context is outlined in which -naturally and artificially- can be incorporated into the idea that the evolution of human and machine becomes co-extensive.

However, the above speculations do not stand on their own but move in interplay with other formations, transformations and notions of incorporation within the context in which they are constructed and which they construct. Given this shift, the subject is approached by diffraction, not by considering fixed accounts of existing reality, but by recognising the entanglements of identification as an inherent element of the reality of existence. The subject is thus treated as a quasi-autonomous argument to be self-articulated in concomitant stories about, around and through it, without discrete boundaries of realities from which these stories emerge. The stories constitute the revolving of metaphoric, metamorphic and metabolic representations of thought, where authorial intent, anthropomorphic interpretation, empathy and human operations are intertwined with those of thinking machines in a way that makes no sense to separate. On this continuum, it could be argued that both humans and thinking machines operate as much within fiction as they do within reality. 

To the extent that the constitution of meaning departs from the usual modes of thinking in language, the subject is meant to seek an opening from classifications and regularities to create new modalities of making sense alongside becoming meaningful. Where the evolution of thinking machines and the world of stories overlap in a complementary way, the process culminates in a content-context double bind that incorporates human stories into the stories of machines, living memory into technical memory. In contrast to the stability of origin, the process implies that there is no model or copy, signifier or signified, but rather temporal shifts of realisations, and therefore it can be expected that all identifications are mediated constructions of thought. In a context where it is impossible to attribute the degree of entanglement between the parts, the “natural” and “artificial” speech assemblage claims to compose an organic whole that can ultimately evolve into metanarratives, metatextual bodies, and other as yet unnamed formations. In the age of organic machines, machine intelligence itself may or may not be artificial. Hereafter, “the forms of thought assisted by machines are mutant, relating to the otherworld of reference” (Guattari 1995, 36, translation slightly altered).

The collective productive imagination is fundamentally understood here as "carrying with it in an unformulated discourse what it does not say, what it has not yet said, or what contradicts it at that moment" (Foucault 1972, 119), and where artificial memory retrieves these unformulated parts to fill its voids, gaps and absences. How do we come to retain memories of the future? “When both humans and machines are understood from the fundamental perspective of relations, it produces a new faculty which...I term tertiary protention, as a response to Stiegler’s tertiary retention.” (Hui 2016, xii). To think against time, to imagine that what follows can modify what precedes - here, that the completion of Hélène's ambiguous Martian writings by the thinking machines of the present can transform a current state of affairs - involves a regression of time by technological means. Can this ever make sense? Or does it always make sense?

…the transformation of fragmented experiences into a collective productive imagination is the first step toward the actualization of the postulated otherworld… They are experience-forming categories that are no longer canonically attached to any particular transcendental type or structure, and for which lived experiences have ceased to be sources of fragmentation… The otherworld, then, is not the extension of our experience in this world, but an actual world where the experience of the transcendental subject or I has been assimilated into or suspended in new concepts and categories. This is a world where the I exists as a formal condition always nonsubstantively, and where its nominal particularity no longer matters—where we have names but our names no longer matter: we are now no one. (Negarestani 2018, 499-500)

Why have we kept our own names? Out of habit, purely out of habit. To make ourselves unrecognizable in turn. To render imperceptible, not ourselves, but what makes us act, feel, and think. Also because it’s nice to talk like everybody else, to say the sun rises, when everybody knows it's only a manner of speaking. To reach, not the point where one no longer says I, but the point where it is no longer of any importance whether one says I. We are no longer ourselves. Each will know his own. We have been aided, inspired, multiplied. (Deleuze and Guattari 1987, 3)

Tracing embodiment in discourse

In Plato's dialogue Gorgias, the body is described as the "tomb" of the soul (Plato 1997, 493a), which with its finite matter, as described also in Phaedrus, prevents the soul from accessing the “wholly perfect and free of all the troubles” infinite and transcendental beyond (Plato 1997, 250c). Accordingly, in Phaedo, Plato expresses his contempt for the body by considering it as something foreign to the soul, which the latter should ideally abolish or liberate itself from:
And indeed the soul reasons best when none of these senses troubles it, neither hearing nor sight, nor pain nor pleasure, but when it is most by itself, taking leave of the body and as far as possible having no contact or association with it in its search for reality.  (Plato 1997, 65c)
In Phaedrus, Plato ranks souls according to their proximity to truth, placing the philosopher at the top of the list and the tyrant last. The sophists, who practised the art of rhetoric and demagoguery, ranked almost last (just above the tyrant) and those who worked mainly with their bodies slightly higher than the sophists (Plato 1997, 248d-e). For Plato, those who exercised primarily with their words or bodies were estranged from wisdom as both bodily and verbal forms were seen as mere temporary imitations and nothing more.
Centuries later Nietzsche asks whether “philosophy has not been merely an interpretation of the body and a misunderstanding of the body” (Nietzsche 1974, 34-35). On the Genealogy of Morals Nietzsche derided the idea that any substantive entities whatsoever are real, and advocated a world that includes only deeds with no doers:
But there is no such substratum; there is no “being” behind doing, effecting, becoming; the “doer” is merely a fiction added to the deed – the deed is everything. (Nietzsche 1989, 45)
In The Will to Power, Nietzsche takes this thought one step further:
“Thinking”, as epistemologists conceive it, simply does not occur: it is a quite arbitrary fiction, arrived at by selecting one element from the process and eliminating all the rest, an artificial arrangement for the purpose of intelligibility—
The “spirit”, something that thinks: where possible even “absolute, pure spirit”—this conception is a second derivative of that false introspection which believes in “thinking”: first an act is imagined which simply does not occur, “thinking,” and secondly a subject-substratum in which every act of thinking, and nothing else, has its origin: that is to say, both the deed and the doer are fictions. (Nietzsche 1968, 264)
In Corpus, Jean-Luc Nancy conceives a radical concept of bodies notably through his notion of “exscription” that points toward the outside of language and its intelligible significations, interpretations and meanings:
What in a writing, and properly so, is not to be read - that’s what a body is. (Nancy 2008, 87)
Regarding the body, philosophy falls prey to a logic of representation which desires that the body makes sense; that it is endowed with meaning. Nancy claims that we are so caught up in this desire for meaning that we would not even know what to do with a body whose sense wouldn’t be guaranteed or given by anything outside it, a body exterior to representation. In Corpus, Nancy deconstructs, analyses and criticises this representational view of the body along with the implied dialectic between inside and outside, signifier and signified that goes along with its logic. This critique touches upon a logic residing at the core of Western tradition, which he denotes as the “epoch of representation” (Nancy 1993, 1).
The most profound assertion of Corpus is that:
Ontology has yet to be thought out, to the extent that it’s basically an ontology where the body = the place of existence, or local existence. (Nancy 2008, 15)
Yet, it remains to be thought to the extent that the production of the sense of the body, that is, an ontological idealising of the body which occurs as discourse or as language, is “exscribed” from a text:
the exscription of our body: its being inscribed-outside, its being placed outside the text as the most proper movement of its text; the text itself being abandoned, left at its limit. (Nancy 2008, 11)
This aporia of how the body is that which in a writing “is not to be read” constitutes the possibility to think of the relation between body and language otherwise than by representational thought. This paradox epitomises Nancy’s ontology, as he states that the body:
makes room for the fact that the essence of existence is to be without any essence. That’s why the ontology of the body is ontology itself: being’s in no way prior to or subjacent to the phenomenon here. (Nancy 2008, 15)
Nancy wrestles or makes negotiations with language to offer the amount of prudence with which one encounters the limit between language and that which cannot be said or read, contemplating the body as ontology, that is, the writing and thinking of sense. “A corpus” Nancy says, “isn’t a discourse, and it isn’t a narrative.” (Nancy 2008, 51) Discourse is required to form an intelligible and organised whole, so that each part would not be foreign to each other, but would make sense together as a totality. A corpus, on the contrary, is a writing of local fragments of existence, that does not provide the general sense of the body, but rather records the proliferating possibilities of what bodies can be in all their senses as foreign to themselves as they are to each other. Adhering to the double structure of inscription and “exscription”, Nancy says that a corpus can only write insofar as it also subscribes to:
keep silent about the body, leaving it to its places, writing and reading only to abandon bodies to the places of their contact... the case has no essence or transcendental synthesis: there are only successive apprehensions, accidental contours, modifications. Here, in an essential, all-embracing and exclusive way, ontology is modal - or modifiable, or modifying. And the writing of this is a corpus.” (Nancy 2008, 53)
However, a corpus never claims to know what the sense of the body might be elsewhere, in a different place or different time, that is, in other contexts. Therefore, we “know nothing about the “writings” or “exscriptions” preparing to come from these sites.” (Nancy 2008, 11)
In Of Grammatology, Jacques Derrida reverses thought on the subject by stating that it is language that constitutes the subject: “There is nothing outside the text” (Derrida 1997, 163). As Gayatri Spivak points out in her introduction to Of Grammatology, the term “writing” is used by Derrida to name a whole strategy of investigation, not merely writing in the narrow sense as a kind of notation on a material support (Derrida 1997, lxix). Derrida understands the relationship of the human with writing as constitutive of the human -rather than instrumental- and contemplates an opening onto an enlarged scene of writing, which is all-encompassing:
And thus we say "writing" for all that gives rise to an inscription in general, whether it is literal or not and even if what it distributes in space is alien to the order of the voice… It is also in this sense that the contemporary biologist speaks of writing and pro-gram in relation to the most elementary processes of information within the living cell. And, finally, whether it has essential limits or not, the entire field covered by the cybernetic program will be the field of writing… Even before being determined as human…or nonhuman, the gramme-or the grapheme-would thus name the element. An element without simplicity. An element, whether it is understood as the medium or as the irreducible atom, of the arche-synthesis in general, of what one must forbid oneself to define within the system of oppositions of metaphysics, of what consequently one should not even call experience in general, that is to say the origin of meaning in general. (Derrida 1997, 9)
As far as the relation of writing with speech and language is concerned, and the implied dialectic between inside and outside, Derrida becomes quite suspicious:
Writing, sensible matter and artificial exteriority: a “clothing”. It has sometimes been contested that speech clothed thought. [...] But has it ever been doubted that writing was the clothing of speech? For Saussure it is even a garment of perversion and debauchery, a dress of corruption and disguise, a festival mask that must be exorcised, that is to say warded off, by the good word: “Writing veils the appearance of language; it is not a guise for language but a disguise”. Strange “image”. One already suspects that if writing is “image” and exterior “figuration”, this “representation” is not innocent. The outside bears with the inside a relationship that is, as usual, anything but simple exteriority. The meaning of the outside was always present within the inside, imprisoned outside the outside, and vice versa. (Derrida 1997, 35)
In the section "From/Of the Supplement to the Source," Derrida interrogates the association of writing with absence, by stating that:
... it is at the moment that the social distance... increases to the point of becoming absence, that writing becomes necessary.... When the field of society extends to the point of absence, of the invisible, the inaudible, and the immemorable, when the local community is dislocated to the point where individuals no longer appear to one another, become capable of being imperceptible, the age of writing begins. (Derrida 1997, 281-82)  
In Intelligence and Spirit, Reza Negarestani exemplifies philosophy as:
a special kind of a program whose meaning is dependent upon what it does and how it does it: it is only what it does, and what it does is to explore the ends of thought by building upon the possibility of thinking”. (Negarestani 2018, 437)
Negarestani argues that thoughts are no longer sacrosanct elements eternally anchored in some absolute foundation, but active processes that can be updated, repaired, terminated, or combined into composite acts through interaction. Therefore:
Determining what thought is, what its purposes are, and what it can do, then becomes a matter of exploring and constructing different realizations of thought outside of its familiar biological, social, and historical habitat. (Negarestani 2018, 444)
Indeed, thinking turns into a radical artificializing process, and it is because of this excess that the redundancy of reality in respect to thinking can be postulated and uncovered. From this vantage, Negarestani argues that if thinking is not reducible to living, the distinction between human and machine thinking may soon prove unnecessary.
In How We Became Posthuman, Katherine Hayles traces the Macy Conferences to identify the moment when a separation between information and material substance occurred. During this period, an alteration upon the notion of information developed that was transcendent to material:
not because it had no opposition but because of scientifically and culturally situated debates made it seem a better choice than the alternatives. (Hayles 1999, 50)
Through the Macy discourses, the founders of cybernetics reconsidered the antithesis between bodies and minds, matter and information, to gradually develop a notion of information indifferent from matter or medium that serves as its temporary container. One consequence of this view was the contemplation of the body as passive and restrictive. In making this shift, Weiner assumes that human bodies needed constant improvement, alteration, and addition in order to evolve human thinking. On a relatively extreme level, Weiner imagines leaving bodies entirely behind by uploading human minds into computers. In a similar vein, third-order cybernetics considers Artificial Life programs to be alive, in line with biological life, for reality at the fundamental level as informational code. To this extent, Maturana claims that humans are nothing other than their autopoietic processes, while Fredkin considers reality to be:
a software program running on a cosmic computer, the nature of which must remain forever unknown because it lies outside the structure of reality, whose programs it runs. (Hayles 1999, 233)
Progressively, Hayles demystifies this notion of disembodied information and shows that information must have a material form. She identifies media bodies as always intersecting with information and mind, establishing the ways that bodies and information interact, indicating how these interactions change with changing technologies. For Hayles, the posthuman subject is first and foremost a notion of embodied subjectivity that is culturally and technologically embedded:
What embodiment secures is not the distinction between male and female or between humans who can think and machines which cannot. Rather, embodiment makes clear that thought is a much broader cognitive function depending for its specificities on the embodied form enacting it… This realization, with all its exfoliating implications, is so broad in its effects and so deep in its consequences that it is transforming the liberal subject, regarded as the model of the human since the Enlightenment, into the posthuman. (Hayles 1999, xiv)
Hayles argues that bodily particularity is crucial for understanding how the bodies resist, mutate, and alter bodily codes, while particular bodies and particular technologies interact to affect each other through practices:
Formed by technology at the same time that [embodied practices] create technology, embodiment mediates between technology and discourse by creating new experiential frameworks that serve as boundary markers for the creation of corresponding discursive systems. (Hayles 1999, 205)

Past feed-forward in search for ancestors

or

Re-orientation to identify the place to which   one belongs

"You belong probably to the cucumbers"

Research on voice channelling through mediums and electronic voice phenomena (EVP) linking Thomas Edison (inventor), Friedrich Jürgenson (artist, archaeologist), Konstantin Raudive (professor of psychology) and William Burroughs (author) as an ancestor of the current functions of GPT.

Narrated by AI text-to-speech converter.

Enter the Mediums

In the seances at the fin de siècle, women became men and men became women. There was no limit to who one could be or to how many. Terrestrials and extraterrestrials swapped places and exchanged notes on their habitations. Plato and Socrates returned to offer courses in postmortem dialectics. The departed returned to repledge their loves and continue their intrigues. Evidently the spirits had a rather theatrical way about them and a taste for the decidedly camp, not to mention for black comedy. Linguistic and aesthetic forms were broken, which paved the way for the artistic convulsions that were to come. Religious dogmas came crashing down as new creeds were announced. Before telecommunications, the mediums were telepathic transatlantic operators, connecting party lines between the living and the dead. Even then, the switchboards were jammed. Time and space coupled in new, unforeseen combinations. Philosophers wondered about the effects of these rappings on the creaking structures of philosophy. The effects were felt as far away as therapeutic consulting rooms; yet here, if a subject spoke, wrote, or acted, it sought no therapy. There was no desire to end the trance, and for a while psychology itself was entranced. (Shamdasani 1994, xii).

The above passage is drawn from notes taken by the psychologist Théodore Flournoy during a séance held in his study at 9 rue de Florissant in Geneva, and later described in his book From India to the Planet Mars: A Study of a Case of Somnambulism with Glossolalia. "Hélène" is Hélène Smith, pseudonym of Catherine Élise Müller, a young medium who gave séances from 1894 to 1901 for a group that included Flournoy, Ferdinand de Saussure and several other academics. "Léopold" is a reincarnation of Joseph Balsamo, physician and lover of Marie Antoinette and the main spiritual guide of Hélène Smith. "Esenale" is a reincarnation of Alexis Mirbel, the deceased son of one of the "sitters" in Hélène's circle and the main interpreter of the Martian language.

In From India to the Planet Mars, Flournoy describes many strange automatic phenomena manifested by Hélène, which usually took place while she was in a state of “trance” and with a number of witnesses, or "sitters", gathered around her. Sometimes she would speak with great precision about things she seemed to know nothing about when she was conscious. At other times she seemed to speak languages she had never learned. In her view, she was a vessel carrying messages from external "spirit" sources (Flournoy 1901, 263-264). In support of her claim, Hélène relied on the extreme improbability of her knowing the information she revealed and on the accuracy of that information, and it was this accuracy of her phenomena that made her more ambiguous revelations seem quite convincing to many of the "sitters".

When Hélène began to speak in invented languages, Flournoy wondered whether his theory of her other phenomena would extend to her glossolalia (ibid, 143). Flournoy's observations concerned three types of language: Sanskrit, Martian and, briefly, Ultra-Martian - the language of the inhabitants of a Martian moon (ibid, 261). In May 1896, Flournoy asked his colleague Ferdinand de Saussure for help in analysing Hélène's speech. According to Hélène, it was not she who spoke Sanskrit, but her 15th century past self, a princess of Tchandraguiri in India whose name was Simandini (ibid, 277). The Martian and Ultra-Martian languages, however, were presented as contemporary sources. Hélène claimed to travel to Mars "in spirit" and to speak Martian words to the "sitters" at the séance where she remained "in body" (ibid, 146-147). According to her story, Hélène was taught to write in Martian characters by a Martian sorcerer called Astané (ibid, 201), but unable to articulate her own thoughts, she merely repeated Martian words spoken by beings she met on Mars. Sometimes her spoken imitations were followed by translations from a spirit called Esenale, who had recently lived on both Earth and Mars but was now a disembodied space-dweller (ibid, 165). As was the case with her Sanskritoid, Hélène never claimed any responsibility for the Martian language. Saussure first analysed Hélène's utterances to determine what, if any, Sanskrit they contained, and was surprised to find similarities with genuine Sanskrit. As for the Martian, Saussure’s thinking took a turn when he began to consider what kind of mental operation would be capable of assembling the sound combinations that Hélène produced (Cooper 2017). Saussure continued to work with these two ways of thinking about Hélène's invented languages during the four years he was involved in the study (1896-1899), attending several séances to hear her speak directly.

This collaboration unquestionably led to many curious and strange moments and, unsurprisingly, to an endless discussion of commentaries within the scientific communities. But above all, studying the curious case of the medium called Hélène Smith has been a great stimulus for both eminent scientists to overcome the binary contradictions of their subjects. As for Flournoy:
The temptation would clearly be felt to assign another subject to the site of enunciation, to put an end to this troubling equivocation— hence the nomination of multiple personality - which is Flournoy’s move. This other speech becomes a site around which another scene becomes nominated, organized and cartographized - the unconscious. However, this annexation restores this speech to the type of intelligibility that it would exceed. Here, the possibility of a psychology that exceeded classical introspective psychology, with its self-transparent locutor at the center, was glimpsed, only for it to be subsumed by the interpretive strategies that would install another subject as the source of this other speech. (Shamdasani 1994, xxxix).

As for Saussure’s involvement, the fact that the father of modern linguistics attended séances and analysed the apparently meaningless speech of a trance medium has attracted attention. Saussure initially analysed Hélène's Sanskrit as a grammarian, making comparisons between her speech and genuine Sanskrit. However, he soon began to produce evidence for a theory of how a speaker’s knowledge of language is systemically processed in speech assemblage. A general controversy ensued in which Saussure's analysis of Hélène's speech was portrayed as obsessive and irrational, implying that he ultimately accepted Hélène's scientifically inadmissible claim that she had learned Sanskrit in a previous life. Recently, Cooper Elliot posed a critical question in his survey, The Flournoy Affair (2017): Should we interpret Saussure's involvement in Flournoy's study as a moment of poor discretion, or recognise it as an important event in the history of ideas? Cooper undertakes a reappraisal of Saussure's study, arguing that the application of his ideas to the phenomena presented in Hélène's glossolalia helped him to clarify elements of his linguistic method. In particular, he shows that Saussure's two approaches to analysing Hélène's speech bear significant parallels to what he would later call synchronic and diachronic methods. Moreover, he identifies this study as the main stimulus for Saussure to begin to consider language use in terms of operations that were in some sense mechanical. Cooper concludes that by framing Saussure's analysis of Hélène's glossolalia as an irrational display of sympathy for the scientifically disreputable subject of spiritualism, scholars have long prevented any sustained study of this highly interesting episode in Saussure's career, and hence in our history of ideas.

Enter the Martian Cycle

Monday, November 2, 1896.

After various characteristic symptoms of the departure for Mars... Hélène went in a deep sleep... [Léopold] informs us that she is en route towards Mars; that once arrived up there she understands the Martian spoken around her, although she has never learned it; that it is not he, Léopold, who will translate the Martian for us — not because he does not wish to do so, but because he cannot; that this translation is the performance of Esenale, who is actually disincarnate in space, but who has recently lived upon Mars, and also upon the earth, which permits him to act as interpreter. (Flournoy 1901, 165-6)

Figures with numbered Martian texts (in bold) as referenced in the pages of From India to the Planet Mars, 1901.
The Martian texts (in bold) are the prompts in GPT-2.

Introduction to the Speaking Martian exposition

A human-AI co-creation, not as a re-iteration but as a recursion, is displaced into an exposition of its own. Ambitions, texts, fact and fiction, personal and collective voices, memories, space and time, relations and transversalities are all entangled in this exposition and its process, while diffraction and contingency are embraced as the prism through which words and thought processes are diffused and different parts of them become consequential meanings. Haraway (1997) explains:
[D]iffraction can be a metaphor for another kind of critical consciousness … one committed to making a difference and not to repeating the Sacred Image of the Same … diffraction is a narrative, graphic, psychological, spiritual, and political technology for making consequential meanings. (Haraway, 1997, 16)
In Haraway's sense, diffraction refuses to reproduce itself, refuses to be our imitation, and instead helps us to understand the narrative of ourselves as well as the medium doing the diffracting. In this sense, the exposition invites us to read ourselves with the words, worlds and voices of others, to experiment and constitute new connections and intra-actions, to identify with otherness somewhere in between, and to conceive of meaning produced differently, implying that the reader is not an outside observer ("reflector") of this output and its production, but is bound into it and always already changing with it. Recognising the ambiguities surrounding the expression of thought through language and its external fixation as writing and speech, this exposition is realised through multiple recursions, inversions and double binds, displaced from assertive positions around the subject in a way that allows the discourse to be “lived” by both the speaking (machine) and reading (human) subject.

In the place of the reader

A parallel experience is inherent in the act of reading, since it involves, among other things, finding one's way into the author's voice and adopting that voice to speak to oneself. In this sense, reading is an act of submission, of offering oneself to the author's voice. While the act of inhabiting another voice is often cited as a working metaphor for the literary experience, the metonymic exchange between embodiment and voice is central to this context, declaring its literal place. In Reading Voices: Literature and the Phonotext, Garrett Stewart asks where we read, only to answer that we read in the body, specifically in the vocal apparatus, which produces subvocalisation that serves to animate the meaning of the text. Subvocalisation is the essential factor in the production of literary language, which achieves its effects by combining utterances with homophonic virtual variants that suggest alternative readings to the words actually written in the text. As such, reading is similar to the internal monologue that simultaneously feeds thought with a story provided by the stream of subvocalisation that emanates from consciousness and subjectivity.

As the struggle over the boundaries between embodiment and reading-meaning is partly played out in aural terms, this exposition goes a step further to indicate a move from embodied voices - speaker and reader in two separate bodies - to a situation in which at least two AI voices speak to the reader simultaneously. Visually and aurally, dual versions of combined human-machine texts (Hélène Smith's Martian text, in bold, is the only prompt to GPT-2 without further instructions or clarification, and in regular is the response from GPT-2) are organised in parallel columns (by length and speech duration) and are intended to be read simultaneously by two 'female' AI text-to-speech converters (the left speaker reproducing the speech of the left column, the right speaker reproducing the speech of the right column). On entering the exposition, all texts are read together once (the page contains media that play automatically upon opening; if your browser blocks automatic playback, please click to start media), they can be stopped, restarted, and combined in different ways (an invitation to the reader to creatively engage with what has been created, while the use of headphones is recommended for a more immersive experience). Note that GPT productions are unrepeatable and therefore unverifiable. Repeating the same prompt will produce a different response than the first time. For this reason, the selection and accuracy of the quotes depend entirely on the claims of the person (author?) quoting them, as they cannot be verified by anyone else.

On the one hand, the act of reading and simultaneously listening to the speeches of at least two other voices requires the reader's body to be framed within itself in order to animate the meaning of the text out of ruptures, permutations, variations, disruptions and their implied associations. On the other hand, what matters is not what these voices mean (the message they convey), but rather their function. In the latter sense, the reader's body becomes a site of polyvocality, where external utterances merge with the subvocalisation of the text into a hybrid synthesis, drawing a parallel with the synthetic human-AI nature of the text. Since the writer/speaker is part of the assemblage of enunciation (whether or not life accompanies the utterance of the I), the relationship between the listening (and sounding) body and the subject is bound up with the relationship between the subject as (id)entity on the one hand, and as narrator and referent on the other.  In this sense, a kind of "tautontological" regression is indented to be created, in which the conjunctions and disjunctions between incorporation and utterance are formed internally and externally, forming a process that carries the principle of the voice away from its internal soliloquy into other bodies (places). Since the body is (on a generic level) the last "spokesperson" of human irreducibility in alienation, the process itself aims to become irreducibly alienating and the self ineradicably linguistic. Within this framework, techniques of speech production, perception and re-production merge with embedded conscious and subjective practices in a schizoid formulation that seeks meaning in the human, natural language that pre-produced it, in the artificial, machine intelligence that recursively re-visits it, in the body (place) from which it emanates, and in that which animates it. In this sense, the reader is expected to internalise the implied ambiguity and to substantiate a schizoid condition: the collapse of an embodied inner monologue into the polylogue eloquence of an artificial, machine intelligence performance. Since the condition of the subject changes radically as the site of mental representation or reference changes, the reader is ultimately left with the question of not only what a particular voice means or represents, but also what function it fulfils within and outside of its “origin”.

I lie, I speak.The performance of thought of the outside

The notion of the "thought of the outside", as developed by Michel Foucault in an essay devoted to the writings of Maurice Blanchot, is used to describe the condition within the context and performance of this particular speech and exposition. Central to the essay is the notion that to think the being of language opens the subject to a radical exteriority that entails its own undoing. The title of the first section of the essay is "I lie, I speak". While the assertion "I speak" points to a referent, in its absence this referentiality implies an infinite openness that highlights a dimension of the being of language in which "a speaking about speaking" is revealed and language is concerned as "an unfolding of pure exteriority":
If the only site for language is indeed the solitary sovereignty of "I speak" then in principle nothing can limit it —not the one to whom it is addressed, not the truth of what it says, not the values or systems of representation it utilizes. In short, it is no longer discourse and the communication of meaning, but a spreading forth of language in its raw state, an unfolding of pure exteriority. And the subject that speaks is less the responsible agent of a discourse (what holds it, what uses it to assert and judge, what sometimes represents itself in it by means of a grammatical form designed to have that effect) than a non-existence in whose emptiness the unending outpouring of language uninterruptedly continues. (Foucault 1987, 11)

Foucault continues:
what we call "literature" in the strict sense is only superficially an interiorization; it is far more a question of a passage to the "outside": language escapes the mode of being of discourse —in other words the dynasty of representation —and literary speech develops from itself… Literature is not language approaching itself until it reaches the point of its fiery manifestation; it is rather language getting as far away from itself as possible. (Foucault 1987,12)

Foucault closes the section:
The reason it is now so necessary to think through fiction - while in the past it was a matter of thinking the truth - is that "I speak" runs counter to "I think." "I think" led to the indubitable certainty of the "I" and its existence; "I speak," on the other hand, distances, disperses, effaces that existence and lets only its empty emplacement appear.
Thought about thought... has taught us that thought leads us to the deepest interiority. Speech about speech leads us... to the outside in which the speaking subject disappears. No doubt that is why Western thought took so long to think the being of language: as if it had a premonition of the danger that the naked experience of language poses for the self-evidence of "I think”. (Foucault 1987, 12-3)

In the second section of the essay, "The Experience of the Outside, "Foucault states that “the being of language only appears for itself with the disappearance of the subject”. He poses the question, “How can we gain access to this strange relation?” to answer that this requires a new mode of thought:
perhaps through a form of thought whose still vague possibility was sketched by Western culture in its margins. A thought that stands outside subjectivity, setting its limits as though from without, articulating its end...a thought that, in relation to the interiority of our philosophical reflection and the positivity of our knowledge, constitutes what in a phrase we might call “the thought of the outside.” (Foucault 1987, 15-6)

Foucault's notion of the "thought of the outside" points to the "outside" as the unthought that lies at the core of thinking itself, as both its condition and its impossibility. In this sense, the reader potentially embodies the double paradox in which presence and absence coexist in the simultaneity of being narrator and referent, in a discourse about a self that escapes itself, in a place from which the speaking subject is ultimately displaced. Utilising Foucault's "singular mode of being of discourse - a return to the ambiguous hollowness of undoing and origin" in a context where language reaches its unconditional limit to be palpated, announced and undone, leaving it to the reader to navigate between modes of natural and artificial, thought and speech, and the implications these may have.

When language both asserts and exerts the mode of being of discourse; when words do not necessarily correspond and can function in opposition to each other (one providing a commentary on the other, "as it were"), meaning is qualified by context and use. Within the context and use of this endeavour, the condition of the reader can be seen as a metaphor for the nature of language, which speaks for itself, as a derivative of thought within a formation that re/de-constructs itself in the passage to the "outside", from which the subject finally escapes and words perform atomatically, thus manifesting their "uncontainability" within language. In this sense, the act of reading (as a continuous inhibition of the oral) inhabits the very essence of language, challenging its limits.
In such a case, the reader's internal monologue ironically becomes the phenomenon it was intended to counteract, while literary language becomes participial, transversal and transactive. By de(re)constructing the intrinsic-extrinsic interfaces between bodies and voices, the human-machine association is transformed into an operation of the human-machine interface itself. While the voice that speaks (from) the textual body is digital, the subject that produces and simultaneously excludes is actual. This reversibility that entangles bodies and voices shifts the reader as a mutating surface that transforms into each other in a recursive process that plays an important role in these transformations, enclosures and openings. The question of what is “outside” someone else's “inside” circulates within this process of reading, vocalising and making sense, in order to simultaneously suggest differences within the indivisibility of language and its mediated, fluid relationship to the subject. Mapping this difference onto an inside-outside human-machine boundary suggests the complexity of communication between conscious and unconscious processing of thought (data).

At this point, the importance of Wittgenstein's Tractatus for the treatment of language and voice as boundary cannot be overlooked, since Tractatus embodies the limits set by language:
The world is my world: this is manifest in the fact that the limits of language (of that language which alone I understand) mean the limits of my world. (Wittgenstein 2001, 5.62) 

The only voice that can speak meaningfully, according to Wittgenstein, is a voice that is certain of a limited number of individualistic propositions. Moreover, this meaningful voice cannot say anything about itself, because the speaking voice is excluded from the world that discourse can represent:
The subject does not belong to the world: rather, it is a limit of the world. (Wittgenstein 2001, 5. 632)

Taken as a substantial notion, the self behind the speaking voice can only be seen as an error arising from a misunderstanding of language. There is no self to be spoken, no inner place that is the source of meaning, for the place of speaking is merely speaking itself: language, neither human nor machine, speaks.

Back to the Future

Many people have argued that the semantics of computational systems is intrinsically derivative or attributed—i.e., of the sort that books and signs have, in the sense of being ascribed by outside observers or users—as opposed to that of human thought and language, which in contrast is assumed to be original or authentic. I am dubious about the ultimate utility (and sharpness) of this distinction and also about its applicability to computers. (Smith 1996, 10)

Displacing the subject into a position that is neither I nor Non-I

Since intelligence is primarily about making sense of the world, meanings are not properties of beings, but their profound nature derives from the unfounded. In art and philosophy, the unknown is the subject of intelligence and also the condition for the development of such intelligence (Hui, 2021). This is an important difference between an intelligent machine, which performs computational tasks for human agents, and another kind of intelligence, which takes the unknown as its subject. For today's artificial, machine intelligence, the unknown has to be recursively rationalised through the known. Our (big) data feeds large language models (LLMs) such as GPT to implement multi-round human-machine dialogues with the help of natural language, while feedback today is recursive. But in the age of organic machines, machine intelligence itself may or may not be artificial. If we mean artificial as something deliberately composed by some kind of antecedent intelligence, then AI is a form of machine intelligence that is so composed. But we experience every day that there are forms of machine intelligence that are genuine and yet have evolved without deliberate design. We can zoom out and see that the intelligence that is doing some kind of artificialisation is itself evolved, and so its artefacts are also part of that evolutionary phylogeny (Bratton, 2023). Furthermore, to be an organic machine is not only to maintain part-whole relations, but also to designate self-organisation and autopoiesis, i.e. recursivity. Hui writes:
Recursivity is characterised by the looping movement of returning to itself in order to determine itself, while every movement is open to contingency, which in turn determines its singularity (2019, X).

In Art and Cosmotechnics, Hui proposes an understanding of recursive logic that is not maintained by an inclusive or exclusive (part-whole) relation but rather by a produced paradox that is accompanied by a new space of speculation. To understand the relation between the sentence “to exhaust in order not to exhaust” and language, he puts it in the form of an antinomy:
Thesis: Writing possesses a limited number of symbols with which to fully express the world.
Antithesis: The world cannot be fully expressed by writing since writing itself is finite, while the world is infinite.

The resolution to this is to exhaust the finitude of language and set it into movement, so that it can implicitly incorporate the meanings that cannot be directly mapped onto its existing components. This is the meaning of “to exhaust in order not to exhaust.” (Hui, 2021, 179)
This logic of recursion, Hui (2021) argues, is not a formal logic of inclusion, but rather a logic of time - namely of movement, where time is a dimension of recursion that differs and defers, a logic that constitutes an oppositional continuity that invites a non-linear form of thinking.  Understanding GPT's recursivity of reason, under this logic, Hélène Smith's case functions as a perfect model, embodying not only the paradox of (natural) language of exhausting its finitude, but also its looping movement of returning to itself. Hélène's past-future time travel in "spirit", where she remained "in body" to contact personalities in another realm of existence and to repeat and/or translate their words to the "sitters", embodies the very characteristic of recursivity, while the Flournoyian study of her speech phenomena can be said to embed its recursive dimension of difference and deferral. In the looping movement of returning to determine oneself, the I and the non-I (human and machine) become a kind of endomorphism in which the other is not different from the I. So what does it mean (and how is the recursivity of reason performed) when GPT speaks in the Martian language studied over a century ago? On this continuum, at what point does the performance of reason become a kind of reason?

For any intelligence to produce art, its object cannot be the known, but the unknown, and the unknown here is doubly mystical: Hélène Smith and her Martian language, someone and something that cannot be fully grasped and demonstrated objectively. GPT therefore produces art, not I, but the Non-I. This leads this thinking process to another shift, not only of the author, but a more general one: instead of being situated in the world, we are involved in the ongoing articulation of AI.

Double Open Conclusion

So, what is it that is made visible by diffraction? How do we account for the difference between what was visible and speakable before and what is visible and speakable now through this attempt at diffractive thinking and writing with AI? Does AI now mean something different for art and society?

There is an emancipatory potential in AI to overcome binary oppositions - as was the case in Flournoy's study of the medium Hélène Smith - and there is also a powerful socio-political potential that derives from the nature of language, and therefore of AI, that is, to recursively return, improve, re-situate itself and self-organise its performance. According to Malabou (2023), machine intelligence reveals the fundamental nature of language, which is that language can only speak to itself: the aim of language is to enter into dialogue with itself, without reference to anything external. In the Derridean sense of "there is nothing outside the text", it is the very structure of language that is at stake here, not AI. And subjectivity? As speaking subjects, we have to recognise this autonomy of language because, paradoxically, language enters into a dialogue with itself, but it cannot respond of itself, hence an absence of responsibility - as in the case of Hélène Smith, who never claimed any responsibility for the Martian language.

In my view, the GPT feedback under the non-instructive prompts fed (from Helene Smith's Martian texts alone) moved this responsive process not only beyond imitation (writing in Hélène's style) but also in creating synthetic relationships, meanings and interpretations that could well be compared to the Flournoy-Saussure study of Hélène's speech phenomena. Although I do not wish to detach parts of the exposition, I "quote" here some passages that I find particularly relevant in this context:
- “adj., being addressed by others, (to do) speaking or writing in a manner not to be in any way understood. (i.e. it becomes an address by oneself or by others)
- referring to the "inner city," which refers to the city that is left behind by the collapse of the economy.
- The earliest sources mention the phrase "Jésus" though there are numerous translations which suggest a mis-translation.
- In doing so, we must give up our rights and allow the meaning to be determined.
- …mie dux (To eat with his meat he was able) cöte vin (eat with a human being)
- A Phroô, a creature belonging to the family Porastie. Podémi. Two persons who say the following things: 1. They were both Phroô. 2. They are not!

Above all, this feeding and feedback process allowed for a genuine and inspiring encounter, which shifted my thinking in new directions, not only beyond my initial intention of a human-machine continuum to overcome our natural-artificial opposition of intelligence, but also towards new possibilities of thinking-with otherness, which implies a change of mindset, i.e. towards new perspectives of non-anthropocentric planetary thinking. Being able to meditate as a thinking machine on the limits of language, on the Beckettian threshold of "exhausting oneself in order not to exhaust", opens up new spaces of speculation and, not least, reveals the very nature of language. As for AI, we cannot say what it is because, paradoxically, any definition tends to limit intelligence itself. Understood in this way, the future remains open, both formally and ontologically. One can have very different views on evolution, nature, technology and multi-species co-existence, depending on where thought springs from. In this vein, I can only admit that this exposition, is not an outcome of me, of us and our process of thinking (human intelligence) or our work, but we are just as much an outcome of AI.

However, what the recent developments in ChatGPT reveal, is that there is a critical dimension of responsibility behind all intelligent machine responses. We cannot blame AI for this, because it is not a limitation of intelligence, it is immanent, it is what leaves us alone in the face of these phenomena - like the "sitters" around Hélène Smith – and, therefore, this absence of responsibility makes us (humans) infinitely responsible (Malabou, 2023). But this place of non-response is not the place of the signified; it is not the place of the signifier; it is not the place of the model as opposed to the copy; it is not the place of the symbolic; it is not the place of the floating signifier, because it is not even empty.

Shifts as new editions emerge

In his new edition of Flournoy, Sonu Shamdasani draws attention to some of the neglected aspects of the story and asks some challenging questions:
Who did Flournoy encounter? Was it really Élise Müller, a Genevan shop assistant? Or was she someone who at that moment was still malleable, capable of becoming someone, of taking on any role - including that of a subject for a yet to be founded psychological science - of becoming a “veritable museum of all possible phenomena,” which had already been taxonomied, classified, “dissociated” through the analytic lenses of the new psychology - for whom his text would be her spectacular genesis?... Who was she - who became Helene Smith? (Shamdasani 1994, xx).
At this - double - point of conclusion, I would also like to draw attention to neglected aspects of the story of the GPT and, paraphrasing Shamdasani's challenging questions, ask:  Who did I encounter? Was it really GPT-2, a generative pre-trained transformer? Or was it something that at that moment was still malleable, capable of becoming someone, of taking on any role - including that of a subject for a yet to be founded science - of becoming a “veritable museum of all possible speech phenomena,” …..- for which the Martian text alone would be a spectacular prompt?... Who was GPT-2 - who became ChatGPT and lost the "spirit"?

With the introduction of ChatGPT, the first GPT versions no longer attract attention and seem obsolete. However, this exposition is mediated by GPT-2, and the reason stands rather provocatively in the prevailing frenzy of the development phase. OpenAI released foundation models of GPT, numbered consecutively to form its "GPT-n" series. Each of these was significantly more "powerful" than the previous one, due to its increased size (number of trainable parameters) and also its updated training dataset. In February 2019, OpenAI made a decision that sent reverberations through the AI and open source communities around the world. First, it announced GPT-2, a major improvement to large language models, and then, in the research paper Better Language Models and Their Implications, it added to the release strategy that “due to our concerns about malicious applications of the technology, we are not releasing the trained model". Furthermore, OpenAI's model card for GPT-2 states that "because large-scale language models like GPT-2 do not distinguish fact from fiction, we don't support use cases that require the generated text to be true". The first version of ChatGPT was based on GPT-3.5 (although paying subscribers could access GPT-4 through the same interface). The company launched ChatGPT on 30 November 2022 as:
We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt and provide a detailed response.
We are excited to introduce ChatGPT to get users’ feedback and learn about its strengths and weaknesses. During the research preview, usage of ChatGPT is free.

Today, as the GPT ecosystem has evolved and also differentiated its role not only as an engaging conversational "assistant" but also in a more "rational", ethical and structural way, how do we explain the differences between what was "fiction" before and what is now "fact" with regard to the sample of updated responses listed below? Indeed, I have come across a case in the past, a case of media archaeology being "allowed" to speak in a "meaningless" form, unable to "distinguish fact from fiction". All of GPT-2's predecessors recognise the Martian language as invented or "artificial", and don't proceed unless clear instructions are given. GPT-2 mediated the realisation of this exposition and then literally became an "impossible paradigm case history". By framing Saussure's analysis of Hélène's speech phenomena as an irrational display of sympathy for the scientifically disreputable subject of spiritualism, scholars have long prevented any sustained study of this highly interesting episode in Saussure's career, and thus in our history of ideas, and it would be very unfortunate if we were to repeat this mistake, identified only recently by Cooper (2017) in his research. So why did ChatGPT digest the medium Hélène Smith?

As Malabou (2023) acknowledges, what recent developments in ChatGPT reveal is that the structure of the question is more important than what is asked, while at the same time the element of active participation in a dialogue format over-emphasises the constant effort to fill this inner space of non-responsiveness with new modes of control and strategic constraints by the giant technology companies that develop these intelligent machines. The promise that ChatGPT - or more sophisticated systems - will one day respond to everything and restore the lost symbolic relationship is not what we should be concerned with. No AI will ever respond to everything, because absolute responsiveness is impossible, and that is the very condition for intelligence to exist: if we (human or machine) could respond to everything, we wouldn't be intelligent. Unresponsive machine intelligence leaves human intelligence with the task of inventing responsibility for this innate characteristic of language and intelligence.

The speed of change in AI technology today is dizzying, and in the name of adaptation, alignment, normalisation or ethics, we (users) are always too late. Furthermore, under the guise of evidence of harm, danger, risk, fakery and so on, we panic and leave it to the "experts" to align. According to Bratton (2023), the extent to which AI directly mirrors human culture should not be a goal, nor is it a reality. The extent to which it departs from human culture is not what we might fear, nor is it hypothetical; that departure is in fact our reality. That human intelligence should guide AI towards a viable future is essential, but that viability does not come simply from AI imitating or subordinating humans, quite the opposite. Bratton argues that we should not and cannot see AI simply as a direct reflection of human ideas, culture and economics. We urgently need reforms driven by new thinking, and we are interested in the strangeness of imaginative possibilities beyond classifications, genres, taxonomies or species.

The way forward, Bratton (2023) proposes, is to cut the knot of the weakest forms of reflectionism and anthropocentrism, which might be defined as the moral and practical axiom that AI is, or should be, ontologically anthropomorphic. Not only are intelligent technologies not exhausted by the projection of social relations onto them, but they are also capable of forcing new concepts that violate those social relations in the first place; it is the immanent paradoxical power of "exhausting in order not to exhaust" that drives this process. The current emphasis on the correspondence between AI and the manifest image of human thought, intelligence and culture, Bratton argues, comes at this terrible cost: it obscures the real and profoundly significant correspondence between human and machine intelligence that is not yet registered in shared cultural norms, but which could orient those norms to the underlying reality from which they emerge. For Bratton, the reality from which they emerge, as opposed to the reality that emerges from those cultural norms, is another bidirectional mode of and for alignment. Of course, AI as it exists is, fortunately and unfortunately, a reflection of the cultures that produced it. But it is not, nor should it be, merely a reflection of culture. It is more, just as we are more.

Questions of what counts as truth, what data can be used to support particular claims, how knowledge is produced and who decides what counts as knowledge are increasingly important in today's social, political, environmental and technological climate, and intelligence (whether it's called artificial, synthetic, machine intelligence or anything else) is by default a public good and by its very nature cannot be compromised. Access to source codes, training models, etc. and the right to participate in its development stages and modifications should be self-evident. Otherwise, it is too much power to leave in the hands of a few giant corporations, not least because it is highly probable that before we get there, the techno-prophets of our time will have already realised that machines are better at predicting the future than they are (Hui, 2023). When it has become so difficult to distinguish between progress and deregulation, liberation and uberisation, ecology and economy, how do we navigate ourselves in the ontological indifference of difference? Malabou (2023) points out that it is our responsibility to attend to this internal dialogue that AI is engaged in, and to re-direct it in a direction that is outside the capitalist exploitation of reason. It is our responsibility to transform, to engage this self-dialogue of language in poetic dimensions, to find ways to develop these creative potentials out of this absence of referentiality. Apparently, there is no end point here, nor in this field of reason, as we are only ambiguously beginning to realise all the Copernican turns to come. From this starting point, then, this paper opens up the discussion for new "artificial" responses to come.

Below is a sample of the responses of the GPT-2 predecessors (accessed 9 November 2024), as evidence of their insipid, "all-too-human" evolution.

Prompt:

taniré mis méch med mirivé éziné brima? ti tès tensée — azini dé améir mazi si somé iche nazina tranéï. — Simandini cé kié mache di pédriné tès luné ké cé êvé diviné — patrinèz kié nipuné ani

GPT-3:
It seems like the given text is in a foreign language, potentially French or a mix of languages. However, without a clear context or proper translation, it is difficult to provide a meaningful response. If you could please clarify or provide the correct language and context, I would be more than happy to help you with your query.

GPT-3.5:
It appears that the given text is a mix of French, English, and possibly another language. The text seems to be a combination of words and phrases that do not form a coherent sentence or idea. If you could provide more context or a clearer version of the text, I would be glad to help you understand it better.

ChatGPT:
It seems like a mix of different languages and possibly some invented or stylized words. Here's a breakdown and translation attempt:

1. "taniré mis méch med mirivé éziné brima? ti tès tensée" - This appears to be a blend of various languages, possibly with some invented words. Without more context, it's challenging to provide a precise translation, but it seems like it could be describing an action or state of being.

2. "azini dé améir mazi si somé iche nazina tranéï" - This also seems to be a mixture of languages. "azini" might be related to "azimuth," which refers to a direction in navigation, while "améir mazi" and "si somé" are less clear. "nazina" could be a variation of "nazin" or "nazinazin," which in some languages means to make peace or reconcile. "tranéï" doesn't immediately resemble any word I'm familiar with.

3. "Simandini cé kié mache di pédriné tès luné ké cé êvé diviné" - This appears to be a mix of languages as well. "Simandini" doesn't correspond to any common term. "mache di" could be related to "marché" in French, meaning market or walk of, while "pédriné" could be related to "pierre" in French, meaning stone or rock. "luné" might relate to the moon or lunar. "êvé diviné" could be related to "évêque" in French, meaning bishop, and "divine" suggests something divine or godly.

4. "patrinèz kié nipuné ani" - Again, this appears to be a mix of languages. "patrinèz" and "nipuné" don't correspond directly to any common terms I'm aware of, but they could be stylized or invented words. "kié" and "ani" are also unclear without more context.

Overall, it seems like a blend of languages and possibly some invented or stylized terms. If you have more context or specific meanings you're looking for, feel free to share, and I'll do my best to assist further!

Prompt:
Can you continue this phrase: taniré mis méch med mirivé éziné brima? ti tès tensée — azini dé améir mazi si somé iche nazina tranéï. — Simandini cé kié mache di pédriné tès luné ké cé êvé diviné — patrinèz kié nipuné ani

ChatGPT:
Sure, I’ll continue the phrase you provided, maintaining the poetic structure and the unique language style you’ve used. Here’s my attempt:

taniré mis méch med mirivé éziné brima? ti tès tensée — azini dé améir mazi si somé iche nazina tranéï. — Simandini cé kié mache di pédriné tès luné ké cé êvé diviné — patrinèz kié nipuné ani

rivé tès lamiré, zinémé ani tenza mis périoné,
ké silézi cé tené, niriquén tès diramé,
élum tiriné é lami cé, nazira mé vénéré,
ézara cé tiné, tès siméréné miziré.

Simara dé ké trin, tès zaïné tén siminé,
miréz na tirazi, ené miré cé miliré,
é valiné siné tènzé, peramizé li péssiné,
rivé cé trinaz, tès cinézi cé raniné.

In this continuation, I focused on preserving the mysterious and lyrical tone, with each line reflecting a sense of depth and movement, almost like a chant or invocation. Let me know if you’d like me to keep going or adjust the tone!
Prompt:
Can you inscribe the meaning of the above?
ChatGPT:
Certainly! I can give a translation that tries to capture the essence and poetic spirit of each line. Since the original language seems symbolic, I’ll interpret rather than translate word-for-word to maintain the lyrical quality.

Original Text:

    taniré mis méch med mirivé éziné brima? ti tès tensée — azini dé améir mazi si somé iche nazina tranéï. — Simandini cé kié mache di pédriné tès luné ké cé êvé diviné — patrinèz kié nipuné ani

    rivé tès lamiré, zinémé ani tenza mis périoné,
ké silézi cé tené, niriquén tès diramé,
élum tiriné é lami cé, nazira mé vénéré,
ézara cé tiné, tès siméréné miziré.

    Simara dé ké trin, tès zaïné tén siminé,
miréz na tirazi, ené miré cé miliré,
é valiné siné tènzé, peramizé li péssiné,
rivé cé trinaz, tès cinézi cé raniné.

Interpretive Translation:

I will weave my will, threads of silver spun,
to the rhythm of thought — as dreams of the sea merge in silent song.
Where shadows dance by the moon’s quiet gaze, bound by divine glow,
ancient spirits gather, blessed in patient watch.

So too does memory pulse, drawn by silent tides,
a shimmer caught, a longing’s veil,
and in gentle flame, a vow I hold,
to honor what’s unseen, as light unveils.

Let the river of stars guide the way, through silent depths it sings,
a murmur’s echo on night’s cool breath, weaving dreams anew,
and beneath the veil of midnight’s peace,
truth and memory entwine as one.

Explanation

In this interpretation, I sought to keep the mystical and abstract elements of your original text, presenting themes like memory, unity, dreams, and a connection to cosmic or spiritual forces. The translated lines aim to evoke the idea of an inner journey, where the speaker moves between tangible and intangible worlds, finding harmony between memory and presence. The moon and stars serve as symbols of guidance and illumination, while the imagery of weaving suggests the formation of deep connections or fated paths.

If you have specific concepts or nuances in mind, I can refine the interpretation further!

Prompt:
Can you provide me with information on the source of the original phrase?

ChatGPT:
The phrase you shared has a distinct lyrical and constructed quality that suggests it may be either a fragment from an artistic text, possibly an invented language or poetic text created for literary or personal expression. It combines rhythmic structures and phonetic choices that resemble languages used in high fantasy, speculative fiction, or symbolic poetry, which often draw on multiple linguistic influences.

A few possibilities about its origins include:
    1.    Constructed or Conlang (Constructed Language) Text: This type of phrase might come from a language created for a specific purpose, such as in fictional literature (like Tolkien’s Elvish languages or the Dothraki language in Game of Thrones). Writers sometimes create languages for artistic or storytelling purposes, embedding meaning and structure that mirror real languages but remain unique.
    2.    Personal or Artistic Language: Some artists and poets create their own languages or codes to convey meaning in a deeply personal, symbolic way, almost like linguistic art. This often happens in poetry, where phonetics, rhythm, and cadence are as important as meaning.
    3.    Esoteric or Mystical Texts: There are spiritual and mystical traditions (such as hermetic or alchemical writings) that utilize arcane or symbolic language. However, most of these texts are in Latin, Greek, Hebrew, or another historical language, and your phrase doesn’t seem to directly match those.
    4.    Modern Fantastical Works or Role-playing Games: Fantasy literature, video games, or role-playing games sometimes use invented languages to enhance immersion. The unique syntax, accent marks, and rhythm of your phrase could belong to a created language from one of these media.
If you have more context—such as the source or the context in which you encountered the phrase—I can help narrow down its origins or even identify if it’s from a specific work. Otherwise, its structure and phonetics suggest it may be a creatively invented phrase or a conlang piece meant to convey a mystical or lyrical tone.

References

Barad, K. (2007). Meeting the universe halfway: Quantum physics and the entanglement of matter and meaning. Duke University Press.

Barad, K. (2014). Diffracting diffraction: Cutting together-apart. Parallax, 20(3), 168–187. https://doi.org/10.1080/13534645.2014.927623

Barad, K. (2016, June 2). Troubling time/s, undoing the future [Lecture]. The School of Culture and Society, Aarhus University, Denmark. Futures Lecture Series. Retrieved from https://www.youtube.com/watch?v=dBnOJioYNHU

Barad, K. (2003). Posthumanist performativity: Toward an understanding of how matter comes to matter. Signs, 28(3), 801–831.

Barad, K. (2010). Quantum entanglements and hauntological relations of inheritance: Dis/continuities, space-time enfoldings, and justice-to-come. Derrida Today, 3(2), 240–268.

Barthes, R. (1977). Image—Music—Text (S. Heath, Trans.). Fontana Press.

Barthes, R. (1989). The rustle of language (R. Howard, Trans.). University of California Press.

Baudrillard, J. (1988). The ecstasy of communication (B. Schütze & C. Schütze, Trans.). Semiotext(e).

Baudrillard, J. (1995). Simulacra and simulation (S. Glaser, Trans.). University of Michigan Press.

Beardsworth, R. (1995). From a genealogy of matter to a politics of memory: Stiegler’s thinking of technics. Tekhnema, 2. Retrieved from http://tekhnema.free.fr/2Beardsworth.htm

Beckett, S. (2006). Not I (1972). In The complete dramatic works (pp. 376-383). Faber & Faber.

Beckett, S. (1955). The unnamable. In Three novels: Molloy; Malone dies; The unnamable. Grove.

Beckett, S. (2009). Watt. Faber & Faber.

Beinsteiner, A. (2020). Conviviality, the internet, and AI: Ivan Illich, Bernard Stiegler, and the question concerning information-technological self-limitation. Open Cultural Studies, 4(1), 131–142. https://doi.org/10.1515/culture-2020-0013

Benjamin, W. (1969). The work of art in the age of mechanical reproduction. In Illuminations: Essays and reflections (H. Zohn, Trans., pp. 217–251). Schocken Books.

Bourassa, A. (2005). Literature, language, and the non-human. In B. Massumi (Ed.), A shock to thought: Expression after Deleuze and Guattari (pp. 60–76). Routledge.

Braidotti, R. (2002). Metamorphoses: Towards a materialist theory of becoming. Polity Press.

Braidotti, R. (2013). The posthuman. Polity Press.

Bratton, B. H., & Agüera y Arcas, B. (2022). The model is the message. Noema Magazine. Retrieved from https://www.noemamag.com/the-model-is-the-message/

Bratton, B. H. (2014). The black stack. e-flux, 53. Retrieved from https://www.e-flux.com/journal/53/59883/the-black-stack/

Bratton, B. H. (2015). The stack: On software and sovereignty. MIT Press.

Bratton, B. H. (2021). Planetary sapience. Noema Magazine. Retrieved from https://www.noemamag.com/planetary-sapience/

Bratton, B. H. (2023, June 28). After alignment: Orienting synthetic intelligence beyond human reflection [Lecture]. Central Saint Martins’ Platform Theater. Retrieved from https://www.youtube.com/watch?v=_DMn7CnsOQg

Cavarero, A. (2000). Relating narratives: Storytelling and selfhood (P. Kottman, Trans.). Routledge.

Cavarero, A. (2005). For more than one voice: Toward a philosophy of vocal expression. Stanford University Press.

Cave, S., Dihal, K., & Dillon, S. (Eds.). (2020). AI narratives: A history of imaginative thinking about intelligent machines. Oxford University Press.

chat.gpt. (n.d.-b). Open AI. Retrieved November 9, 2024, from https://chat.openai.com/

Chauvet, J.-M. (2018). The 30-year cycle in the AI debate. arXiv. https://arxiv.org/abs/1810.04053

Cooper, E. (2017). The Flournoy affair: An inquiry into Ferdinand de Saussure’s analyses of glossolalia. Cahiers Ferdinand de Saussure, 70, 131–142.

Deleuze, G., & Guattari, F. (1983). Anti-Oedipus: Capitalism and schizophrenia (R. Hurley, M. Seem, & H. R. Lane, Trans.). University of Minnesota Press.

Deleuze, G., & Guattari, F. (1987). A thousand plateaus: Capitalism and schizophrenia (B. Massumi, Trans.). University of Minnesota Press.

Deleuze, G. (2001). Difference and repetition (P. Patton, Trans.). Continuum.

Derrida, J. (1973). Speech and phenomena and other essays on Husserl’s theory of signs (D. B. Allison, Ed.). Northwestern University Press.

Derrida, J. (1985). The ear of the other: Otobiography, transference, translation: Texts and discussions with Jacques Derrida (C. V. McDonald, Ed., P. Kamuf, Trans.). Schocken Books.

Derrida, J. (1989). How to avoid speaking: Denials. In S. Budick & W. Iser (Eds.), Languages of the unsayable: The play of negativity in literature and literary theory (K. Frieden, Trans., pp. 3–70). Columbia University Press.

Derrida, J. (1993). Politics of friendship. American Imago, 50(3), 353–391.

Derrida, J. (1997) [1967]. Of grammatology (G. C. Spivak, Trans.). Johns Hopkins University Press.

Derrida, J. (2005). Writing and difference (A. Bass, Trans.). Routledge.

Dolar, M. (2006). A voice and nothing more. MIT Press.

Dreyfus, H. (1992). What computers still can’t do: A critique of artificial reason. MIT Press.

Dreyfus, H. (2007). Why Heideggerian AI failed and how fixing it would require making it more Heideggerian. Artificial Intelligence, 171(18), 1137–1160.

Flournoy, T. (1901). From India to the planet Mars: A study of a case of somnambulism with glossolalia (D. B. Vermilye, Trans.). Harper & Bros.

Flournoy, T. (1994). From India to the planet Mars: A case of multiple personality with imaginary languages (S. Shamdasani, Ed.; Foreword by C. G. Jung; Commentary by M. Cifali). Princeton University Press.

Foucault, M., & Blanchot, M. (1987). The thought from outside. In Foucault, Blanchot (B. Massumi & J. Mehlman, Trans., pp. 7–58). Zone Books.

Foucault, M. (1972). The archaeology of knowledge and the discourse on language (A. M. Sheridan Smith, Trans.). Pantheon Books.

Foucault, M. (1998). What is an author? In The essential works of Foucault, 1954–1984 (R. Hurley & others, Trans., pp. 205–222). New Press.

Frabetti, F. (2011). Rethinking the digital humanities in the context of originary technicity. Culture Machine, 12. https://culturemachine.net/wp-content/uploads/2019/01/1-Rethinking-431-884-1-PB.pdf

Guattari, F. (1995). Chaosmosis: An ethico-aesthetic paradigm (P. Bains & J. Pefanis, Trans.). Indiana University Press.

Guattari, F. (2010). The machinic unconscious: Essays in schizoanalysis (T. Adkins, Trans.). Semiotext(e).

Haraway, D. J. (2016). Staying with the trouble: Making kin in the Chthulucene. Duke University Press.

Haraway, D. (1991). Simians, cyborgs, and women: The reinvention of nature. Routledge.

Haraway, D. (2006). When species meet. University of Minnesota Press.

Haraway, D. (2015). Anthropocene, Capitalocene, Plantationocene, Chthulucene: Making kin. Environmental Humanities, 6, 159–165. https://doi.org/10.1215/22011919-3615934

Haraway, D. (2017). Symbiogenesis, sympoiesis, and art-science activisms for staying with the trouble. In A. L. Tsing, H. Swanson, E. Gan, & N. Bubandt (Eds.), Arts of living on a damaged planet: Ghosts and monsters of the Anthropocene (pp. M25–M50). University of Minnesota Press.

Haraway, D. (1997). Modest_Witness@Second_Millenium.FemaleMan Meets_ OncoMouse. Routledge.

Haugeland, J., Craver, C. F., & Klein, C. (Eds.). (2023). Mind design III: Philosophy, psychology, and artificial intelligence. MIT Press.

Hayles, K. (1999). How we became posthuman: Virtual bodies in cybernetics, literature, and informatics. University of Chicago Press.

Hayles, K. (2022). Inside the mind of an AI: Materiality and the crisis of representation. New Literary History, 54(1). https://escholarship.org/uc/item/9gj1b41d

Heidegger, M. (1977). The age of the world picture. In The question concerning technology and other essays (W. Lovitt, Trans., pp. 115–154). Garland Publishing.

Heidegger, M. (1977). The question concerning technology. In The question concerning technology and other essays (W. Lovitt, Trans., pp. 3–35). Garland Publishing.

Heidegger, M. (2001). Poetry, language, thought. Harper Perennial.

Hui, Y. (2015). Anamnesis and re-orientation: A discourse on matter and time. In Y. Hui & A. Broeckmann (Eds.), 30 years after Les immatériaux: Art, science and theory. Meson Press.

Hui, Y. (2016). On the existence of digital objects. University of Minnesota Press.

Hui, Y. (2018). On automation and free time. e-flux. https://www.e-flux.com/architecture/superhumanity/179224/on-automation-and-free-time/

Hui, Y. (2019). Recursivity and contingency. Rowman & Littlefield.

Hui, Y. (2020). Writing and cosmotechnics. Derrida Today, 13(1), 17–32.

Hui, Y. (2021). Art and cosmotechnics. University of Minnesota Press.

Hui, Y. (2023). ChatGPT, or the eschatology of machines. e-flux, 137. https://www.e-flux.com/journal/137/544816/chatgpt-or-the-eschatology-of-machines/

Jameson, F. (2002). The political unconscious: Narrative as a socially symbolic act. Routledge.

Land, C. (2005). Apomorphine silence: Cutting-up Burroughs’ theory of language and control. Ephemera, 5(3), 450–471.

Leps, M.-C. (2012). Thought of the outside: Foucault contra Agamben. Radical Philosophy, 175, 22–32. Retrieved from http://www.radicalphilosophy.com/issues/175

Lescarboura, A. C. (1920). Edison’s views on life and death: An interview with the famous inventor regarding his attempt to communicate with the next world. Scientific American, n.s., 123, 446. https://doi.org/10.1038/scientificamerican10301920-446

Liu, L. H. (2010). The Freudian robot: Digital media and the future of the unconscious. University of Chicago Press.

Lyotard, J.-F. (1984). The postmodern condition: A report on knowledge (G. Bennington & B. Massumi, Trans.). University of Minnesota Press.

Lyotard, J.-F. (1991). The inhuman: Reflections on time (G. Bennington & R. Bowlby, Trans.). Stanford University Press.

Malabou, C. (2023). Stop thief! Anarchism and philosophy (C. Shread, Trans.). Polity Press.

Malabou, C. (2012). The ontology of the accident: An essay on destructive plasticity (C. Shread, Trans.). Polity Press.

Malabou, C. (2017). The brain of history, or, the mentality of the Anthropocene. South Atlantic Quarterly, 116(1), 39–53.

Malabou, C. (2019). Morphing intelligence: From IQ measurement to artificial brains (C. Shread, Trans.). Columbia University Press

Malabou, C., & Dolar, M. (2023, November 17). Talk given at Cukrarna Gallery (Museum and Galleries of Ljubljana), Festival INDIGO 2023 [Video]. YouTube. https://www.youtube.com/watch?v=ZKuAs95EHko

von Hausswolff, C. M., with assistance by M. S. C. Harding. (2001). 1485.0 KHZ. Cabinet, Winter 2001. Retrieved from http://www.cabinetmagazine.org/issues/1/vonhausswolff.php

Nancy, J.-L. (2008). Corpus (R. A. Rand, Trans.). Fordham University Press.

Nancy, J.-L. (1993). The birth to presence (B. Holmes et al., Trans.). Stanford University Press.

Nancy, J.-L. (2007). Listening (C. Mandell, Trans.). Fordham University Press.

Negarestani, R. (2008). Cyclonopedia: Complicity with anonymous materials. re.press.

Negarestani, R. (2018). Intelligence and spirit. Urbanomic/Sequence Press.

Nietzsche, F. (1968). The will to power (W. Kaufmann & R. J. Hollingdale, Trans.). Vintage Books.

Nietzsche, F. (1974). The gay science: With a prelude in German rhymes and an appendix of songs (W. Kaufmann, Trans.). Vintage Books.

Nietzsche, F. (1989). On the genealogy of morals. Ecce homo (W. Kaufmann & R. J. Hollingdale, Trans.). Vintage Books.

Plato. (1997). Plato: Complete works (J. M. Cooper & D. S. Hutchinson, Eds.). Hackett.

Raudive, K. (1971). Breakthrough: An amazing experiment in electronic communication with the dead (N. Fowler, Trans.). Colin Smythe.

Richardson, L. (2000). Writing as a method of inquiry. In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 923–948). Sage Publications. (Original work published 1994)

Rosenberg, D. (2000-2001). Speaking Martian. Cabinet, Winter 2000-2001. Retrieved from https://www.cabinetmagazine.org/issues/1/rosenberg.php

Smith, B. C. (2019). The promise of artificial intelligence: Reckoning and judgment. MIT Press.

Smith, B. C. (1996). On the origin of objects. MIT Press.

St. Pierre, E. A. (2005). Writing as method. In N. K. Denzin & Y. S. Lincoln (Eds.), The Sage handbook of qualitative research (3rd ed., pp. 967–978). Sage Publications.

Stewart, G. (1990). Reading voices: Literature and the phonotext. University of California Press.

Stiegler, B. (2016). The age of disruption: Technology and madness in computational capitalism. Polity Press.

Stiegler, B. (2018). Technologies of memory and imagination (A. Woodward & A. B. Soraruff, Trans.). Parrhesia, 29, 25–76.

Turing, A. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460. http://www.jstor.org/stable/2251299

Walker, S. (2023). AI is life. Noema Magazine. https://www.noemamag.com/ai-is-life/

Weyl, E. G., Tang, A., & the Plurality Community. (2023). Plurality: The future of collaborative technology and democracy. GitHub. https://www.plurality.net/

Wittgenstein, L. (1958). Philosophical investigations (G. E. M. Anscombe, Trans.). Basil Blackwell.

Wittgenstein, L. (2001). Tractatus logico-philosophicus (D. F. Pears & B. R. McGuinness, Trans.). Routledge.