By Any Other Name

Gordon Sumner was Geordie by birth and Sting by choice. He grew up in Wallsend, on the north bank of the Tyne, which made him geographically a Geordie — the demonym pinned on him by accident of postcode. Years later his bandmates in the Phoenix Jazzmen took to calling him Sting because of a yellow-and-black striped sweater that made him look like a bee. He kept the invented name and discarded the inherited one.

I've been quietly making the opposite trade.

My legal name is George. My father was George. My grandfather was George. By the time I arrived my mother had run out of patience for the project of distinguishing one George from another in the same household, so she started calling me Geordie, and the name has followed me around since. I'm now in my sixties, living in Spain by way of Portugal, publishing essays under the name my mother used in the kitchen. Sting moved away from his inherited Geordie. I've leaned into mine.

Watching the trade play out in mirror image clarified something about how monomyms actually arrive. They don't arrive one way. They arrive at least four.

The Earned

Einstein. Feynman. Darwin. Freud. The legal surname that fame hardened into a one-word handle. None of these men picked anything; the surrounding culture did the work, and the resulting monomym is a byproduct of accomplishment rather than a self-presentation. Earned monomyms are dividends. By the time anyone was calling Einstein Einstein, he no longer had a choice in the matter.

The Invented

François-Marie Arouet became Voltaire because Voltaire did actual work — it was shorter, cleaner, and it cut him loose from a surname dense with inherited family obligation. Stendhal, Saki, Molière, Colette, Banksy, Randi: all invented, all adopted before the accomplishment that would retroactively justify them. So is Sting. The invented monomym is a wager. The bearer picks a name from nowhere and walks into it, betting the work will catch up.

The wager isn't smaller for being made by a writer instead of a bassist. Voltaire and Sting are doing exactly the same thing. The respectability gap between them is mostly an artifact of the centuries of hindsight one has and the other doesn't.

The Inherited Nickname

Cher is short for Cherilyn — the family used it before the cameras did. Madonna is her actual legal first name, in domestic circulation long before Like a Virgin. The bearer didn't invent anything; the name was already running, and what they did was simply continue running it in public.

This is my category. Geordie came from my mother, deployed for the operational reason that there were too many Georges in the room. By the time I was old enough to have an opinion about it, the nickname had hardened into the name I answered to among everyone who'd known me before kindergarten. Choosing to publish under it is less a christening than an admission. The airlines do not recognize it. The boarding pass requires the legal version underneath. But the name on the byline is the one mom used.

Briefly, in fourth grade, the playground tried out Einstein on me. It did not stick. Inherited nicknames are not a function of the recipient's preferences. They are conferred by whoever has standing to confer them, and a Florida public school's standing for a single year is not what mom had.

The Formalists

The fourth category is the mirror image of the third. The Formalist refuses the inherited diminutive in favor of the formal baptismal name. Phillip-not-Phil. Robert-not-Bob. Christopher-not-Chris. Where the inherited-nickname person says the name my family uses is fine, the Formalist says the name on the birth certificate is the one I prefer.

This category attracts the most discomfort, but the discomfort is selective. My brother is Philip at home and Phil with friends, and he himself is doing nothing in particular about it — the registers switch around him without his intervention, which is the ambient default for every name in every language. My friend Phillip was Phil until college, when he was elected president of his fraternity and the role required more gravitas than Phil could supply; the formal version stuck and carried him into a professional career. Neither case is cringe. Both are register-fit.

The flinch arrives only when the formal version is enforced across registers — when the bearer demands Phillip at the office and at the backyard cookout, against the natural pull of the moment. What readers flinch at is not the formal name. It's the request that casual contexts honor a formal register.

The Rose Business

At which point Shakespeare arrives. A rose by any other name would smell as sweet. Juliet's claim is that names are contingent, that essence precedes nomenclature; the play she's in disagrees. Romeo dies because of a surname. The Montague-Capulet feud is a name-driven tragedy, and Shakespeare lets Juliet say her famous line and then kills her for believing it.

She's also thirteen when she says it, which matters. She's a teenager arguing that the world should be simpler than the adults insist it is. The line sounds like wisdom at fourteen and like tragic misrecognition at sixty. Shakespeare almost certainly knew the difference.

The honest answer is somewhere between Juliet and her play. Names don't change the rose. They do change the rose's social existence, and for humans the social existence is most of what we call life. George, Geordie, Voltaire, Arouet — the molecules are identical. The reception, the register, the company the name keeps are not. The rose doesn't care. We're not roses.

A Small Star Trek Detour

Star Trek made the same argument twice. By Any Other Name is the title of a TOS episode from 1968, in which aliens called Kelvans reduce the Enterprise crew to small polyhedral essence-blocks and assume human form themselves. The moment they're wearing human bodies they become subject to human appetites, and Kirk uses the change to defeat them. The form changes the essence; the title quotes Juliet and the plot raises an eyebrow at her.

Two decades later TNG fielded a chief engineer named Geordi La Forge, after a real fan — George "Geordie" La Forge Jr. — who had muscular dystrophy and died in 1975, twelve years before the show aired. Roddenberry reached back into memory for the name. The character is, quietly, a memorial. And his nickname came to him through the same mechanism mine did: family use, distinguishing one George from the others. He himself moves through the taxonomy without ceremony. Geordi among friends, La Forge on the bridge, Commander La Forge up the chain, no corrections offered to anyone. Context sets the register; the name bends to fit.

That's the model I'm trying to follow.

The earned, the invented, the inherited, the formal — four gestures, four registers, four kinds of gap between the name on the boarding pass and the name on the byline. Sting closed the gap one way. I'm closing it the other. The rose doesn't care which gesture you make.

We're not roses. We're graphs.

---

The Ancestry Problem

A graph is nodes and edges. The interesting part is never the node — it's the edges, the connections, the paths from here to there. Names are nodes. The interesting question about a name is: what does this particular node connect to?

The GEDCOM standard — the file format genealogical software has used since 1984 — is essentially a graph serialization format. People are nodes; parentage and marriage are edges. My database covers several thousand individuals across roughly twenty generations. What it looks like at a glance is a tree. What it actually is, three generations in, is a tangle: cousins marry, populations close on themselves over centuries, and the same node begins appearing in multiple branches connected by edges you didn't anticipate. A tree with cycles is not a tree. It's a graph.

My primary lineage runs through the Strothers of Newton in Northumberland — a Catholic recusant family that survived the Tudor Reformation by keeping their heads down and their children baptized in whatever register the local justice of the peace required at the time. They produced a Lancelot, several Williams, and one William VI who emigrated to colonial Virginia around 1650, presumably because continued adherence to Rome in Protestant England had exhausted its long-term career prospects. My branch stayed longer, eventually making it to America by a different, slower route.

Wolfram's concept of the ruliad — the space of all possible computations, which he argues contains all physically and conceptually realizable realities — is the limit case of this genealogical graph. Every possible history is a path through the ruliad. Every family tree is a subgraph of the complete history of human reproduction. Every life is a path. The interesting question is not which node you start at — it's which edges you follow, and which ones were followed before you arrived and constrained the ones available to you.

What genealogy taught me about machine learning: a knowledge graph and a GEDCOM file are doing the same thing at different levels of abstraction. When a language model learns that "Northumberland," "Tyneside," and "Geordie" cluster together in meaning-space, it's reconstructing a fragment of the social history that produced those words. The Strother who emigrated in 1650 left a semantic residue in the names and places he touched. A large enough model finds it. The compression is lossy; so is memory. The difference between a knowledge graph and human recall is mostly that the graph keeps its edges legible.

The Fiber Bet

I run a site called Fibertarian.com. The name is a portmanteau of fiber and libertarian — not politically, but philosophically: the argument that the single most neglected variable in the Western dietary tradition is the stuff we evolved eating by the metric ton and stopped at roughly the same moment our chronic-disease statistics went sideways.

This is, at bottom, a graph argument.

Dietary fiber is not a nutrient in the conventional sense — humans don't absorb it. What it does is feed a bacterial ecosystem in the large intestine that turns out to be running an enormous fraction of human metabolic regulation. The microbiome is a graph: thousands of microbial species connected by competitive and mutualistic edges, inputs and outputs, biochemical signals that propagate into the bloodstream and from there into mood, cognition, immune function, and inflammatory response. We removed the primary substrate of that graph from our diet. The graph changed. The downstream consequences propagated in ways we are still cataloguing.

The evolutionary path is legible if you follow it backward. Modern metabolic disease traces to dietary pattern shifts in the mid-20th century, which trace to agricultural innovations that removed bran from grain, which trace to a pre-industrial diet in which the average person ate roughly forty grams of fiber daily — compared to the current Western average of twelve to fifteen. The gut bacteria weren't selected for a twelve-gram environment. They were selected for forty. The mismatch is not subtle.

What makes the Fibertarian project interesting to me is that it sits at the intersection of evolutionary biology, graph theory (the microbiome), and the very human problem of making correct but underappreciated information compelling enough that people will act on it. Most nutrition writing is either academic (effective but unreadable) or sensationalist (readable but wrong). The productive space between those two nodes is underoccupied. I am trying to occupy it.

The Engineer Can See

Geordi La Forge wore a VISOR — Visual Input Sensory Organ Replacement — that gave him access to a wider spectral range than baseline human vision. He could see in infrared and ultraviolet. He couldn't see in what humans call the visible spectrum, which turns out to be a narrow consensus convention. The band of physical reality is much wider than the slice the human eye evolved to process.

The metaphor has stayed with me for decades. The knowledge-graph work I do — genealogy databases, blog taxonomies, language ontologies, moral-foundations maps — is all in this mode: an attempt to perceive in frequencies that ordinary attention skips past. Not because ordinary attention is defective, but because it is, necessarily, band-limited by the pressures that shaped it. You can only attend to so much at once.

The VISOR was uncomfortable. The show mentioned it occasionally — a low background of pain that came with the expanded perception. That detail is honest. Seeing in multiple registers simultaneously is expensive. Building systems that make expanded perception legible to others is a career, and occasionally a headache.

The Enterprise's LCARS interface — Library Computer Access and Retrieval System — is an underappreciated design philosophy. It assumes the user is navigating a graph, not consuming a feed. It presents branching choices, contextual connections, the shape of an information space rather than a ranked list of documents. A feed is optimized for engagement; LCARS is optimized for finding. The distinction matters more as information environments become noisier. I have spent more thought on LCARS than is professionally defensible, and I am at peace with that.

The Turning and the Foundations

ShadowyRealm is a project organized around a deceptively simple question: how does a single person change the shape of history?

The naive version of this question is the Great Man Theory, which historians have spent a century complicating for good reasons. The interesting version is the graph-theory version: which historical nodes have the highest centrality? Which figures, if removed from the network of their time, would most radically alter the paths available to everyone who came after? Centrality in a graph is not the same as greatness in the Victorian sense. It's structural — a function of how many shortest paths pass through you, how many connections you hold between clusters that would otherwise be disconnected.

Strauss and Howe's Fourth Turning is a temporal graph argument. Their claim: Anglo-American history runs in roughly eighty-year cycles of four distinct phases — a High, an Awakening, an Unraveling, and a Crisis. The Crisis Turning is when the institutional graph gets restructured; the High that follows it is a period of consolidation around the new structure. The theory has real predictive texture: it accounts for why certain periods produce certain kinds of leaders, why institutional trust rises and falls in roughly generational rhythms, why the grandchildren of a Crisis generation tend to build things their grandparents burned.

We are, by most readings of the model, somewhere in a Crisis Turning now. Whether that's alarming or clarifying depends on your prior distribution over what comes after.

Jonathan Haidt's Moral Foundations Theory is the other frame I return to constantly. His argument: human moral intuition is not a single liberal-conservative dimension but a six-dimensional space — Care, Fairness, Loyalty, Authority, Sanctity, Liberty. Different political cultures emphasize different subsets of those dimensions, and because each person navigates primarily the dimensions they're attending to, political arguments are usually not arguments at all. They're two people solving different subgraphs of the same moral space and concluding that the other is irrational or malicious.

What graph theory adds to Haidt: the dimensions are not independent. They're connected by weighted edges, some positive and some negative. Loyalty and Authority are often co-activated. Care and Fairness conflict at their borders. Liberty and Sanctity are structurally in tension. Political tribes are the clusters that form when you run community-detection algorithms on this moral graph weighted by shared edge-strengths.

ShadowyRealm is my attempt to make that structure visible using historical figures as seed nodes — to trace how the centrality of a person propagates forward in time through the network of their influence.

What the Machine Doesn't Know

LinguaMama is a language-learning app. The short description covers the surface. The longer one: it's an attempt to build a language teacher that has an honest accounting of its own gaps.

Building it taught me more about language acquisition than reading the research literature had, because the literature can afford to be approximately right while a production system cannot. A machine that is approximately right about when to introduce the subjunctive is wrong in ways that confuse the learner at precisely the moment they're trying to build a new syntactic structure.

The main thing I learned: language acquisition is not vocabulary acquisition. Vocabulary is nodes. Language is edges — the relationships, the collocations, the registers, the pragmatic rules about who says what to whom when, in which syntactic frame, with what prosodic weight. A learner who has memorized ten thousand Spanish words but doesn't know which register a given word belongs to has built a lot of nodes and not enough edges. They know the taxonomy but not the grammar of the graph.

Teaching a machine to navigate this exposed two things clearly. First: most language instruction materials are organized around vocabulary lists and formal grammar rules — around nodes and explicit constraints — rather than around the contextual relationships that actually carry meaning. The curriculum is optimized for what's easy to sequence and test, not for what the brain actually needs to build fluency. Second: the gaps in a learner's graph are not randomly distributed. They're systematically shaped by the structure of the learner's first language. Spanish-speaking English learners make different errors from Mandarin-speaking English learners because they're arriving at English from different positions in concept-space, with different edges already in place and different absences where new ones need to be built.

An LLM doesn't know that it doesn't know this. It generates fluent, plausible-sounding language-learning content that encodes the assumptions of the majority-language instruction tradition, because that tradition produced most of its training data. Fine-tuning it requires identifying which edges are missing, which requires a theory of the learner's current graph, which requires meeting each learner where they are rather than where the curriculum assumes they are.

That's the insight building a machine gave me about teaching people. They're the same problem. You're not filling a vessel. You're editing a graph.

The Mojo Problem

MojoBump resists the two-sentence description, which is either a branding problem or a useful signal about the domain it's trying to occupy. It's organized around a question evolutionary psychology frames clearly but popular health writing mostly evades: what determines whether a human being operates at full metabolic and hormonal capacity, and what actually moves those parameters?

The informal term mojo gestures at something the research literature covers under testosterone levels, cortisol regulation, mitochondrial efficiency, sleep architecture, and a small library of other biomarkers that correlate with what people colloquially mean when they say someone has their edge or has lost it. The Bump is the intervention side: the evidence-based practices that actually shift those markers, stripped of the supplement-industry capture that has colonized most of the discourse.

The evolutionary psychology frame is, again, the graph frame. Natural selection built the human organism to allocate resources adaptively in response to environmental signals. The core trade-off is between fast-and-cheap (prioritize current reproduction, deplete reserves, run hot) and slow-and-expensive (invest in future capability, maintain reserves, run cool). High-uncertainty environments shift the allocation toward fast. Low-uncertainty environments allow the slow investment.

Modern environments send a bizarre mixture of signals. The brain's threat-detection hardware is running on Pleistocene firmware in a 21st-century media ecosystem that has been architecturally optimized to trigger threat responses for engagement. The result is a chronic background of stress-hormone activation that was designed for short-duration emergencies, not indefinite low-grade alarm.

The graph problem: the body's regulatory systems are connected by edges that the popular-health literature treats as isolated nodes. Sustained cortisol elevation suppresses testosterone synthesis. Sleep disruption impairs cortisol regulation. Dietary fiber — an edge running back to Fibertarian — influences gut serotonin production, which influences sleep quality, which closes the loop on cortisol. The microbiome, the endocrine system, the nervous system, and the behavioral patterns that govern all three form one graph. Treating any node in isolation produces, at best, partial results and, at worst, the kind of rebound effects that make people conclude the system can't be moved.

It can be moved. But you have to move the whole graph, not just the node that happens to be in the current headline.

---

The Root Node

So: what is a Geordie?

Not a rose. A root node in a graph whose edges run northward to Northumberland and Norman recusants and a Catholic family that kept two sets of baptismal records; westward to a colonial Virginia emigrant named William Strother VI who bet the New World on a fresh start; southward to a Florida kitchen and a mother who ran out of patience for duplicated Georges; and outward — through a VISOR, a GEDCOM file, a gut microbiome, a moral-foundations matrix, a language model's missing edges, and an endocrine system running on wrong-era firmware — into a graph that keeps acquiring nodes faster than I can document them.

The name is the node. The projects are the edges. The ontology is whatever the full traversal reveals.

Sting went one direction. I went the other. The rose doesn't care which gesture you make. The graph does.

We're not roses.