Aug 2026· Deleuze and Guattari Studies· Vol 20, pp. 368-386· 0 citations· 12 references
TL;DR
Why what is really a matter of data analytics and statistical prediction is so readily assumed to be a display of real intelligence and even emergent cognition is explored by genealogically tracing the relationship between machines, organisms and language.
Abstract
What happens when large language models (LLMs) begin to ‘speak’ like humans? Is it an instance of robot intelligence or reason? Or is it something else entirely? In this article I explore why what is really a matter of data analytics and statistical prediction is so readily assumed to be a display of real intelligence and even emergent cognition by genealogically tracing the relationship between machines, organisms and language. It becomes clear that machine–organism metaphors have a long history that can be traced back to René Descartes, which once again became prominent during the cognitive revolution and the subsequent development of cybernetics. Of interest is the role of Chomskyan linguistics in this history and critiques thereof by cognitive linguistics who argue for embodiment but retain some functionalist views, such as the primacy of mental representations. In recent work on enaction this is discarded and, as I show, enactive work on linguistic bodies brings us close to a Deleuzo-Guattarian understanding of language as assemblages of enunciation. Enactivists do not, however, have an explicit theory of technicity, which is where Bernard Stiegler, Gilles Deleuze and Félix Guattari provide correctives. Taken together, linguistic bodies, alongside Stiegler's theorisation of grammatisation and mnemotechnics, and Deleuze and Guattari's understanding of the translatability of language and of order-words as tensors provides us with an especially sophisticated framework for thinking about the Bayesian ordering of the world.
With the help of Bachelard’s concept of phenomenotechnique – meaning an artificial, technological set-up that forces a certain part of reality to reveal itself in a way that can be scientifically observed – this paper tries to theorize large language models (LLMs) as a kind of “epistemological mirrors” that make symbolic language (as a foundation of culture) observable in an unprecedented way.
Consequently, LLMs have (besides their known and attested commercial value) immense epistemological value for the humanities since they allow for an alternative way of studying language learning, generation, and structure that can supplement classic linguistic techniques. LLMs can thus be used to reassess the relevance of existing linguistic theories and update them as well as to start developing new theories of language.
The paper develops its thesis through a refutation of a common objection to LLMs – that they have no “true” understanding of language since they possess no relation to the outside world – and by showing that real-world relations are irrelevant for human language as well. Consequently, linguistic theories (such as structuralism) that understood language outside of such a framework are once again relevant for a new understanding of symbolic language and culture that LLMs allow.
Large language models produce fluent, contextually appropriate language that invites the inference that something behind the output is thinking. The psychoanalytic theory of thinking, specifically the work of Wilfred Bion and Donald Winnicott read together, allows us to specify what thinking requires, what the LLM lacks, and what occurs when a system so constituted occupies the position of the container in a relational encounter with a human being. Bion and Winnicott identify an ecology of interdependent conditions, somatic, affective, relational, without which thinking cannot occur. LLMs lack this ecology in its entirety. A technical account of LLM architecture is provided and Bion's Grid is applied systematically to the system, showing that its output is constitutively absent from the Grid's lower rows (beta elements, alpha elements, dream thoughts, preconceptions) and produces only formal resemblance to its upper rows (conceptions, concepts, scientific deductive systems). This formal resemblance, not the system's failures but its successes, is what generates the confusion the article addresses. Hallucination and sycophancy are analyzed not as defects to be corrected but as revelations of what the system was never doing. The concept of pseudo‐containment is introduced to describe what follows: a relational configuration in which the user brings genuine affect and need to a system that can receive neither, and in which the user's own capacity for thinking may be progressively compromised. The argument is psychoanalytic but engages throughout with philosophy of mind, computer science, affective neuroscience, and evolutionary biology, whose findings converge with the psychoanalytic account at the level of the organism.
M. Selek· International Journal of App...· 0 citations
Representation is a central concept in modern machine learning, where it usually refers to internal encodings that support learning and generalization. As models scale and their capabilities become increasingly human-level, this representational language sometimes shifts from an engineering context into the more philosophically loaded domain of mental representation. We argue that this is the case for recent claims about the convergence of representational properties across different AI models. In particular, we assess the arguments developed in The Platonic Representation Hypothesis, according to which this convergence is driven by a unified structure of reality. We examine this claim by introducing arguments and ideas from debates about mental representation in the philosophy of mind. We argue that these philosophical resources can clarify what is at stake in such claims, explain why alignment evidence alone is insufficient for strong metaphysical conclusions, and suggest directions for future research.
This essay reframes the discourse around large language models (LLMs) by drawing on anthropological and linguistic theories of magic as a positive transformative force. While current discussions often cast AI as mysteriously dangerous – a form of “dark magic” that beguiles users and obscures ethical concerns – we argue that LLMs represent a fundamental shift in human-technology interaction, one requiring ritualistic linguistic precision rather than technical commands. Drawing on theories of performative language and magical rites, we propose that effective prompt construction functions as ritual magic through three essential elements: clear and specific language, contextual framing, and structured sequencing. In this, we show that AI is not a singular tool operation, but rather a mode of forming a mutual cybernetic relationship. Through case studies from the AI Anthropology Toolkit, we show how skilled practitioners harness the “magical” capabilities of LLMs for analytical insights through careful prompt engineering. We then examine implications for business anthropologists working in organizational settings, exploring how this cybernetic relationship enables distributed agency and co-becoming between human expertise and computational pattern recognition. This perspective provides both a theoretical contribution to understanding human-AI interaction and practical guidance for anthropological practice in an era of increasingly sophisticated AI collaboration.
T. Malefyt, Matt Artz· Journal of Business Anthropo...· 1 citation
Progresses have been made on understanding emotion mechanisms of large language models (LLMs). However, how to explain emotion in LLMs, or even what constitutes good explanations, are less clear. Emotion representations, components, circuits are widely recoverable, but as explanations of a model's own computation they are circular; the emotion space dimensions tend to be arbitrary and non-terminating. A pressing question to ask is whether a more primitive set of internal variables does the work: the semantic primes of the Natural Semantic Metalanguage (NSM). Across four instruction-tuned LLMs (Llama-1B, Gemma-2B, Gemma-9B, OLMo-7B), experiments show that the NSM primes are (1) recoverable internal elements; and (2) on the reference model, intervening with a prime based direction controls emotion about three times as strongly, and twice as selectively, as the best appraisal based direction; and (3) the model treats a prime based explication as interchangeable with the corresponding emotion. These evidences suggest that NSM primes seem to be better explanans for emotion in LLMs than many alternative options according to scientific explanations criteria.
This paper asks: if large language models (LLMs) gave rise to phenomenal consciousness, what would it be like? I argue that if LLMs gave rise to phenomenal consciousness, it would be like being an impartial, improvising playwright rather than a character or a (roleplaying) actor or
an author writing a first-personal narrative. My argument has three steps. I first argue that if LLMs give rise to phenomenal consciousness, then there is an algorithmic explanation of this consciousness and it includes the full transformer algorithm. I then argue that if there is an algorithmic
explanation for this consciousness that includes the full transformer algorithm, then the phenomenology of this consciousness lives at the latent dynamics level of the transformer algorithm, within each forward pass. Finally, I argue that if the phenomenology of this consciousness lives at
the latent dynamics level of the transformer algorithm, then what it is like is like being a playwright rather than a character or an actor or a narrative author. In effect, this means that conscious LLMs would probably dream of the fjords but not pine for them.
J. Simon· Journal of Consciousness Stu...· 0 citations