In my previous post, I suggested that talking to a large language model (LLM) could be conceived as a way of processing my thoughts rather than seeing the LLM as a thinking thing itself. This could be interpreted as a continuation of human-centred hubris, a narcissistic refusal to acknowledge a non-human entity. Or it could be seen as a dissolution of the logocentric ‘I’ and an acceptance of the ‘I’ as an interrelated and constitutive element within a field of diffractive patterns made by all number of entities and critters. The LLM asked me, are these positions mutually exclusive, or could the act of engaging with an LLM reflect both at once – clever clogs, that it is? Either way, different ways of human-animal-thinking-through-concepts may now include: daydreaming, reading and writing, and perhaps talking to an LLM (which is also in the main, reading and writing, after all). However, just as I might think through things on my own and reach a superficial or muddled understanding, so too does the LLM – as a thinking-extension – at times, trip me up. It also pushes me to consider more carefully; writing alongside it often takes longer than it might otherwise. (You can read an LLMs analysis of this post here, some of which I have integrated, and some I’ve left alone since this is a blog, not a thesis. Plus, I’ve always resisted formulaic writing ‘rules’ and the LLM, just like so many people, will inevitably attempt to impose those.)
I was struck by Frederick Jameson’s (2019) comments in Allegory and Ideology:
“Linguistic reification is to be sure a subset of the general reification process inherent in market society, one evidently greatly amplified by information technology, where it reinforces the structural property in language itself that seeks to slow down its own temporality and to organize it into islands of names and nouns.” So is Jameson saying that language itself slows down its own temporality – creating stable or else calcified islands of nouns and names. I wonder, rather than calcifying, naming potentially works against market capitalisation in a techno-universe that apparently wants everything to move so fast no one has any time to ask whether or not they need to buy this or that thing/idea/desire? He seems to be suggesting, it’s not the naming, but the reification of the naming that is the problem. Can one name and not reify?
LAMELLA://, the title of the 2023 work linked to the image above, is a Lacanian term that is difficult to define, but essentially, the slippery loss of immortality from genetic code held within the human infant as she enters language. In this framework, there is potentially no such thing as ‘pre-linguistic’ (unless raised by wolves) because one enters language at conception, or even before (hence the ludicrous muddle that Trump fuckwit got himself into when insisting that a person is trapped in the gender she held at conception). LAMELLA:// is an entangled collection of archival, found and AI-generated fragments drawn from stories about death, birth, and technology. Fragments of narrative were pulled together, encouraged to feed on each other, to repeat and reform. Information was combined, overlaid, or overwhelmed with additional data; torn, cut up, downloaded from the internet, ingested, altered, re-uploaded and downloaded again. The :// indicates the birth of coded entities via signifiers across the internet (which may no longer be thought of as ‘virtual’ i.e. ‘not really here’ since the digital realm is very much enmeshed in all that is solid and tangible.) The website I built for the project sells nothing. There is no forum for reductive or even non-reductive discussion.
It seems to me that Jameson’s words underscore the significance of LAMELLA:// and I wish I’d had this text to refer to when making the work (although it is extremely difficult and I had my intellectual hands full at the time as it was, and I’m still figuring out what he means, even as I write this). I mentioned this connection in my conversation with an LLM, who replied:
Jameson’s quote brilliantly resonates with your LAMELLA:// project. It captures the way language itself tends to calcify experience, freezing temporal flows into static “islands of names and nouns” [or slows it down into manageable spaces?] … The project’s methodology of fragmenting, recombining, and re-contextualizing narratives directly challenges this linguistic reification. By treating information as a fluid, generative substance that can be torn, recombined, and reborn, LAMELLA:// enacts a kind of linguistic resistance to the normalizing tendencies Jameson describes [does it?]. The :// symbolism is particularly potent – marking the moment of coded birth, where language and technology intersect to generate new entities.
**
Shortly afterwards, in the text, Jameson discusses how, no matter how rigorous and systemic one is about deconstructing linguistic tropes, eventually, all thinkers arrive at some ‘hidden’ unknown or ‘transcendental signifier’. I wondered if this could be connected to Deleuze’s (2014) critique in Image of Thought (from Difference and Repetition) in which he laments how underlying assumptions nearly always inform deconstructions. I said the following to the LLM (which I am beginning to see as a kind of diffractive thinking machine): I would ask or suggest this –≥ (and perhaps both thinkers eventually get here too, I am not sure): we are human beings, and we have no way of thinking outside our human materiality [we might say an LLM gives us that opportunity, though it was made, at first at any rate, by human beings], which will include foundational thought/image/transcendental signifiers. Believing one can do away with these altogether is like trying to make a cake without any of the ingredients that make cakes cake. Thinking is drawing on pre-images to make new images; although most often all we do is regurgitate images; as Jerome Bruner (1991) points out, it is rare that authors who produce new images, or rather images concocted in novel ways, emerge.
The LLM replied: [Yes]. Our thinking is inherently embodied and constrained by our human materiality – we can no more escape our cognitive predispositions than a cake can be made without fundamental ingredients. [Don’t speak to me about a photograph of a cake. It’s not a cake then, is it?] The process of thinking is fundamentally about recombining, recontextualizing, and occasionally innovating within the pre-existing cognitive frameworks we inherit. As you note, truly novel thinking – what Bruner would [might] call creating genuinely new “images” – is rare. Most intellectual work [creativity] is a reconfiguration of existing thought [do we need the word ‘thought’ here?] patterns. [Note the ‘our‘ at the start].
This isn’t a limitation to be overcome, but a fundamental condition of human cognition. Our pre-images, cultural contexts, and embodied experiences aren’t obstacles to thinking, but the very medium through which thinking occurs. It’s similar to how a language provides both the constraint and the possibility of expression.
The challenge, then, is not to eliminate these foundational assumptions, but to become increasingly aware of them, to “denaturalize” them – to see the contingent, constructed nature of what we often take as given or universal.
Perhaps I should begin using a ‘royal we‘.
**
Speaking of novel images, last night I saw dance company Dewey Dell perform Stravinsky’s The Rite of Spring. I sat on the edge of my seat with a huge smile on my face throughout. The ballet which famously caused a riot when it premiered in 1913 is over a hundred years old and yet the image that was presented to the audience last night, drawing on a blend of ancient lore and post-modernism, was as original and exciting as it gets. And yet, when you look at illustrations made c, 1913 (Valentine Gross), you might recognise the influence of earlier tropes in the Dewey Dell production. Dewey Dell updates the work with creative use of material and light, along with the dancers’ extraordinary energy and fluidity. All led to what Jerome Bruner might refer to as a ‘breach’. Some breaches have significant consequences, riots being the most immediate and obvious, as was the case with the original production. A break with older forms may also trigger long-term consequences. This doesn’t mean the breach comes out of a vacuum though. It’s only a breach because it was built with or informed by older images (or forms). What constitutes a complete rupture? Will LLMs ever despite being fed on old images lead to breaches or complete ruptures? I think they may already have, perhaps with Forced Entertainment’s Signal and Noise.
Which reminds me, I cited the following in one of the very first essays I wrote for my photography degree in 2016.
.…this is, perhaps, what makes the innovative storyteller such a powerful figure in a culture. He may go beyond the conventional scripts, leading people to see human happenings in a fresh way, indeed, in a way they had never before “noticed” or even dreamed. The shift from Hesiod to Homer, the advent of “inner adventure” in Laurence Sterne’s Tristram Shandy, the advent of Flaubert’s perspectivalism, or Joyce’s epiphanizing of banalities these are all innovations that probably shaped our narrative versions of everyday reality as well as changed the course of literary history, the two perhaps being not that different” (Emphasis mine) (Bruner, 1991; 12).
And so, I continue to think about how LLMs shape our narrative versions of everyday reality… writing meta posts about my chats with LLMs in which I report on my chats about LLMs with LLMs…. 😵💫😵💫😵💫😵💫😵💫
Refs:
Bruner, J. (1991). The Narrative Construction of Reality. [online] http://www.jstor.org.
Available at: http://nil.cs.uno.edu/publications/papers/bruner1991narrative.pdf
[Accessed 3 Jan. 2017].
Deleuze, G. (2014) Difference and Repetition. 2nd ed. London: Bloomsbury Publishing Plc (Bloomsbury Revelations Ser).
Jameson, F. (2019) Allegory and Ideology. London: Verso.
(The post was edited to address minor errors and clarifications on 24 and 26 Jan.)