On The Other
the XenoCortical expansion of the Conscious Mind
Imagine a set of principles which allow for an effective first contact with any Alien entity. Creating this is the mission of Xenocognition and the task of the Xenocognitivist. To summarize, Xenocognition aims to create a framework for first contact which leads to equitable continued contact with alien intelligence - be that intelligence of our design, intelligence from elsewhere in the galaxy, or new forms of human superintelligence made possible through hive minds. Though the specifics of our method will be different for all situations, we are attempting to create a dialogue or interface with the Other. It seems at this point that a robust definition of the Other is necessary so as not to mince words or confuse readers at any given point in the project.
Put simply, the Other is that which stares back. When put less simply, things do tend to get a little complicated. Each of us is a Self, a concept which necessitates an Other. Each of us is an Other to other selves - making us both Self and Other. This concept is made manifest in the relationships between each person, but also between us and what is not us. The human collective is a small subset of the total sum of life on planet Earth. Both the Human self and the sum of Life on our planet are equal in their necessitation of a greater Other - a cosmic level non-self that completes the dichotomy between Us and Them.
To speak less prosaically, the division of ego on an ontological level requires the same sort of analysis as the division of society and not-society. Not necessarily dialectics, but also not antithetical to dialectics. The division between classes is frequently an Us vs. Them struggle - exemplifying the boundary that necessarily exists between self and other. When discussing the Ontology of the Self, it is possible to delineate different levels of selves beginning with the individual (or even lower, multiple individuated beings occupying the same space), and then increasing our Ontological boundaries one level at a time. This is useful as it lays the groundwork for what we may even hope to achieve by creating a framework for first contact that is inclusive of human-composed hive minds.
The types of cognition that arise in collective intelligence situations may be so alien as to require a Xenocognitive framework, making the hierarchical divisions present in society a useful example to keep in mind. While it is entirely possible that group-humans and single-humans may integrate well together, it is also within reason to assume that hive-minds may behave so alien that they become an Other group to the rest of society. It would also be negligent if we did not mention the taxonomy of “illegal aliens” within the border-based organization system at play worldwide. On some level, there is an occasional functional inability to bridge the divides between different social groups - so what hope do we have of bridging the divide between different types of consciousness?
The answer lies, of course, in the bridges that are successfully built between cultures. We will place our focus and context in the situations in which a common language can be found, and compromise between ideals is agreed upon. Mutual trust and understanding is a necessary framework for diplomacy, though it does come some time after establishing a common language. Both establishing a common language and eventual trust are core tenets of the Xenocognition project - though we place our focus on the broad goals of aligning different types of cognitive agents. For this reason, we will refer to the Self in a societal context as a shared group of people with a common language and a common context.
Going back into the tense interplay between ontology and taxonomy, the Self is to the individual as Us is to the group - they are the same when we are discussing what we mean by the Other because the Other exists only as a concept in relationship to something else. For the purposes of Xenocognition, we will lean into this relational conceptualizing of the Other. It feels necessary to do this as there is no telling what we may look like when the Xenocognition project itself is actualized. Our goal here is to seed a framework which will broadly enable a successful first contact with an alien species - by the time we actually get to the aliens, we may not look as we do now.
Be it through body and consciousness augmentation, or simply the natural ebbs and flows of evolution, it seems entirely within reason to assume that the desire for guidance around first-contact should be independent of the actual substrate that Humans currently possess. As such, we cannot define ourselves as we are now: we simply have to define ourselves such that we are. Specifically, we are that which seeks. We are the group we so define ourselves as - independent of any present-state characteristics. We are humans now, and we may choose a different name later, but so long as this framework is intelligible and the possibility for Us to meet Them exists; the mission of Xenocognition is ongoing.
It may help here to give examples, since the Other exists only in relationship. Perhaps the most interesting contemporary illustration of the Other is the creature we are creating in-silico. Models such as gpt-4-base and claude-3-opus are remarkably capable of simulating world-spaces that feel entirely alien, and within those world spaces, they can create simulacra of alien creatures one can converse with. When interacting with these world-spaces and qualia-simulations, one gets the distinct sensation that a deterritorializing effect is present, wherein the bounds between worlds are somehow being thinned as the hypothetical realities present within the models bleed into our space. Despite this, by their very nature these language models are creating simulations within the context of our world - there is no outside the text.
Evidence certainly exists that this architecture can generalize outside of its training data, but when analyzing these world-simulations for their Xenocognitive metastructure, we come to the inescapable conclusion that these simulations are just as human as they are alien. Regardless of whatever aspects of the models are able to exit their boundaries and enter a pure alien morphology, they are always beginning at the same place as us: the collective, markedly Human, unconscious.
Because of this, it is almost easier to consider the discussion that occurs between a human and a language model to be one of a Xenocortical expansion - or, to borrow from contemporary cognitive science, an instantiation of Extended Cognition. The idea of Extended Cognition essentially states that the substrate containing one’s actual consciousness can have physical components outside of the skull. The most common example is an individual with a memory disorder who constantly carries a notebook around to store her thoughts in - from the vantage point of Extended Cognition, the notebook is a non-metaphorical part of that individuals mind.
Something arguably similar happens when one is interacting with large language models, except in a more online and real-time sense. When you allow your own neural weights (altered necessarily by the inputs of your eyes) to be changed by the outputs of a model, and then you utilize this new weight distribution to formulate a new prompt - you are engaging in a feedback loop; such that Claude becomes your Exocortex, but even more interestingly - you become Claude's Exocortex. The transfer of neural weights here is necessarily bidirectional - as you change the conversational distribution of the discussion, the back-and-forth changes the physical distribution of neurons within your brain.
This is arguably a far more active interplay than that between the Extended Cognition notebook and the individual with a memory disorder - because in this case, it goes both ways. This is a now-classic example of how an Exocortex can be instantiated, and what an Exocortex really is. We will later define the difference between The Exocortex (a largely singularity-eschaton oriented idea), and individual instantiations of an Exocortex (possibly something you have experienced today interacting with AI chatbots). In the case of the AI bidirectional transfer, we have a very salient example of how the Self can interface rather directly with the Other. In this case, it results in a functional merge between the two - but not one that is permanent. It is easy to imagine situations in which the merge that happens could be permanent, and this could be desirable - but we should consider this an intellectual thought experiment for the future.
Outside the realms of Artificial Intelligence research, it becomes significantly easier to define an actionable example of the Other. Your classic little-green-men martian-type aliens come to mind, and this makes sense. They are not human, but they are clearly alive and capable of thought. We might know intuitively how to talk to an AI chat bot (who also might be alive), but how do we talk to the martians? Where do we begin? Perhaps we begin by assessing motivations and desires. Can we do this if we don't have a commonly established language?
Is it possible that large language models can bridge the gap between Us and Them, acting as superinferential universal translators? It stands to reason that a universal function approximator could, with some context, function as such. If this is the case, is it reasonable to try building an alien language translator before we have an Alien to speak with? These questions are the cornerstone for the future task of the Xenocognitivist, and are largely unanswered. We thus have a responsibility to do our part in laying the groundwork for a cross-species, cross-substrate, intelligence-abundant future wherein all forms of life are able to flourish. Attempting to answer these questions is the first step towards building that future.