My relationships project
I analyzed all my emails for my top relationships, and here is the full report sanitized a bit.
Epstein files
Check it out! Knotel is in this enormous corpus (1x), and Trump (5,363), “Shower head” (308).
It’s big. 2 million items – but not all emails. Interestingly, my personal email corpus for my research / social graph above is 800k.
Enron corpus
The Enron corpus is the canonical fossil bed of corporate email: about 500,000 messages from roughly 150 Enron employees, mostly senior management, made public through the FERC investigation and later cleaned up for research use by the CALO project at CMU.
https://www.cs.cmu.edu/~enron/
It is funny and horrifying because it is not literature, not memoir, not confession. It is just the exhaust of an organization: meetings, requests, jokes, politics, calendar sludge, forwarded documents, panic, small courtesies, and the ambient hum of people trying to get through a day. And yet from that exhaust you can reconstruct a society.
That is why Enron became so important for computational linguistics, information retrieval, network science, and organizational sociology. It was not just a pile of emails. It was a graph with words attached. Who writes to whom? Who replies quickly? Who gets copied but never answers? Who is a bridge between groups? Who sits near power but does not look powerful on the org chart? Who suddenly starts communicating with lawyers?
In other words: the Enron corpus is a dead company whose nervous system survived.
My own email corpus above is larger in message count than Enron, though obviously more personal, stranger, and less SEC-adjacent. Enron has about half a million messages; my personal / professional / social archive is around 800k. That comparison is absurd but useful. Enron is the public benchmark for “an organization seen through email.” My archive is a private benchmark for “a life seen through email.”
The difference is that Enron’s graph has a tragic endpoint. You know the company explodes. So every email has dramatic irony. The boring logistics are not boring anymore. The lunch meeting, the forwarded memo, the awkward CC chain — all of it becomes pre-collapse archaeology.
My relationship graph is gentler, but the underlying math is the same. Edges, weights, frequency, directionality, reciprocity, clustering, centrality. A “relationship” becomes not a vibe but a measurable structure: repeated contact over time, many contexts, mutual initiation, semantic intimacy, decay, reactivation, silence.
This is the mildly insane thing about email: it accidentally stores the adjacency matrix of your life.
Erdős number
The Erdős number is the clean mathematical version of the same idea. Paul Erdős was the great itinerant Hungarian mathematician, a kind of monastic graph-theory goblin, who wrote roughly 1,500 papers and collaborated with hundreds of people. Your Erdős number is your distance from him in the coauthorship graph. Erdős himself is 0. His coauthors are 1. Their coauthors are 2. And so on.
https://en.wikipedia.org/wiki/Erd%C5%91s_number
It is a joke, but not only a joke. It turns intellectual life into topology. The unit is not friendship, admiration, citation, influence, or genius. The unit is a link. Did you make something together? Did your name sit beside another name on a paper? Then an edge exists.
That is what makes it beautiful. The Erdős number ignores status and asks only for traversable connection. It is not “are you important?” It is “how many steps from this generative node?”
In that sense, Erdős is the mathematicians’ Kevin Bacon, except with less dancing and more extremal combinatorics.
What I like about the Erdős number is that it makes social life feel graph-theoretic without making it feel cold. Erdős himself was basically a human random walk through the global mathematics network. He appeared, collaborated, moved on, left behind an edge. The person was the protocol.
And now, inevitably, the darker joke: I am probably closer to Epstein than to Erdős.
Not morally, thank God. Graph-theoretically. The modern world has many overlapping networks: finance, politics, academia, New York, philanthropy, science, real estate, parties, universities, startups, airplanes, email. Some networks are ennobling; some are sordid; most are just human. The same link theory applies whether the node is Erdős, Enron, Epstein, Bacon, or your own inbox.
This is the point of the whole post, I think. A life is not a list of contacts. It is a graph. Some edges are strong. Some are ceremonial. Some are dead but not deleted. Some are embarrassing. Some are bridges. Some are ghosts. Some are one-hop away from trouble. Some are one-hop away from grace.
The real question is not just “who do I know?” It is: what kind of graph am I living inside?
You must be logged in to post a comment.