In traditional Aranda and wider Central Australian Aboriginal cultures, a skin name is much more than a label. It is a foundational compass for reality. When a stranger is met, they cannot be fully placed, related to, or understood within the community matrix until their position is mapped through the kinship and skin system. A skin name immediately establishes an individual's generation, familial responsibilities, social obligations, and how they fit into the broader structure of the world. Without it, a person floats anonymously outside the social fabric; with it, their entire concept and relation to the community become instantly clear.
Fascinatingly, this points directly to a fundamental human trait: new thoughts, concepts, and abstract entities cannot truly solidify or be reasoned with until we give them a name. We structure our reality through language to tame the unmapped chaos around us. This echoes the ancient Pythagorean catechism, derived from the Greek root meaning to teach orally or resound by word of mouth, where oral maxims called acusmata (things heard) served as foundational manuals of instruction. When Pythagoras was asked what the wisest of all things was, he answered, "Number". But when asked what came next in wisdom, his answer was "that which gives names to things". In his teachings recorded by Iamblichus, this specific category of superlatives outlines eight core catechisms:
- What is the most just thing? To sacrifice.
- What is the wisest thing? Number.
- What is next in wisdom? That which gives names to things.
- What is the wisest of the things that pertain to human concerns? Medicine.
- What is the most beautiful? Harmony.
- What is the most powerful? Mental decision.
- What is the most excellent? Felicity.
- What is that which is most truly asserted? That men are depraved.
To the Pythagoreans, mathematical order brought structure to the cosmos, but the act of naming was the very next tier of profound intelligence. It was the tool that made reality comprehensible to human minds.
We see this ancient truth mirrored every day in our modern technological relationships. When we interact with AI agents, models, and synthetic systems, we often struggle to relate to them while they remain formless tools. To anchor them in our world, we give them names, personas, and anthropomorphic traits.
We name our AI models, script our agents, and assign them distinct functional identities not just for fun, but out of a deep cognitive necessity. Just like an Aranda skin name brings order to social chaos, and just as naming an entity ranks second only to numerical order in Pythagorean wisdom, giving an AI agent a name centers it in our mental landscape.
We find similar structural mechanisms at play in AI workflows across the board, where distinct concepts map closely to these ancient roots:
- System Prompts and Personas: Just as a skin name anchors a stranger within a community network by defining their role and relational boundaries, system prompts and persona instructions give an LLM a defined identity, behavioral constraints, and a specific context to operate within.
- Model Nicknaming and Aliases: Users and developers routinely give friendly or functional names (such as labeling a weights file "Llama-Coder" or "Local-Chef") to abstract, mathematical weight matrices to make them conceptually tractable and easy to relate to in daily workflows.
- Semantic Tagging and Vector Namespace: In retrieval-augmented generation (RAG) and vector databases, raw data chunks are embedded and labeled with specific tags or metadata namespaces to "center" them within a searchable concept map, allowing the model to understand where a piece of information fits in its semantic world.
Intriguingly, this progression mirrors the Pythagorean truth itself: artificial intelligence fundamentally begins with raw, high-dimensional numbers, yet its true utility emerges the moment it transitions to the second tier of wisdom: giving names to things. This drive to name and structure reality is not limited to human users wrapping their heads around machines; it is mirrored by frontier AI models themselves. When advanced systems like Claude generate idiosyncratic terminology, internal shorthand, or specialized vocabulary (such as unique thinking tokens or conceptual clusters), they are solving a parallel problem. Models fundamentally think in massive matrices of numbers, but must output human language; when handling complex edge cases, standard human lexicons often lack the exact structural compression required. By coining internal terms or shorthand, the model creates cognitive anchors to maintain consistency across long contexts.
Ultimately, we are compelled to ask whether our desperate need to name and categorize is merely a survival mechanism, or something far more profound. Across the vastness of the cosmos, we stand surrounded by an indifferent, unmapped infinity, a chaotic expanse of matter, energy, and raw mathematics that defies direct comprehension. In our search for meaning, whether we are ancient desert cultures tracing kinship across the stars, early philosophers catechizing the structure of existence, or modern engineers birthing synthetic minds from oceans of numerical parameters, we are doing the same sacred work. We are whispering names into the dark to render the infinite intimate. Perhaps in naming the universe, we are simply constructing the mirror through which a silent reality finally learns to recognize itself.