Why do some ages produce committed purists and others fluent syncretists, from the same menu of traditions? Whether a culture yields devotees or blenders may have less to do with the traditions themselves than with one background parameter — a dial set before anyone chooses anything.
In one mid-sized city you can find a seminary, a yoga studio, and a therapist's waiting room within three blocks of each other. A few people are exactly one of these things: the novice who has given his whole life to the rule, the sitter who wants nothing but the cushion, the analysand who has translated her entire childhood into the vocabulary of the couch. Most people are mixtures — some Stoic discipline for the workweek, some Zen for Sunday morning, a little psychoanalysis for talking about their mother.
What varies across centuries is not the menu. It is which mixtures are
thinkable. Late antiquity minted syncretists as readily as it minted coins; the
Reformation made purists on an industrial scale; the modern spiritual marketplace makes
blenders again. The interesting variable is a pressure that
operates before anyone chooses — a setting that decides, in advance, whether the
corners or the centre of the space of possible people will be crowded. The Dirichlet
distribution gives that pressure a name: α, the concentration.
Each tradition, in this toy, is nothing but a way of spending words: the probability it puts on saying duty, koan, transference. A tradition is the heights of these bars — and because every word keeps a small floor of probability everywhere, no tradition owns any word absolutely. Even a Stoic sometimes says breath.
hand-set toy: three vocabularies of eight words each over a 24-word world; all sampling and inference below are computed exactly from these numbers.
The triangle is the space of every possible person. Each corner is a pure devotee of one tradition; each interior point is one person's blend of allegiances, θ. The glow is the culture's prior — which kinds of person this world makes likely before any individual opens their mouth. Drag the white point to decide who someone is; draw words to listen to them talk; and watch the dashed circle — the reputation their speech earns them — crawl toward the truth.
The instrument runs a three-move recipe. Run it forward and you get a biography; read it backward and you get a reputation.
Stoicism, Zen, psychoanalysis: each is a distribution over what can be said. Where it puts its mass is what the tradition is.
φₖ ~ Dirichlet(β)Not one bin. A mixture θ — 60% Stoic, 25% Zen, 15% analysand — one point on the triangle. The culture's α decides which blends are common.
θ ~ Dirichlet(α)For each act or utterance, draw a tradition from the blend, then draw a word from that tradition. Repeat for a lifetime.
z ~ θ · w ~ φ٠The culture enters only at Move 2. α never writes a word and never edits a doctrine. It acts one step earlier, shaping the population of blends from which people are drawn — the quiet demography of conviction that is settled before any individual conversion, argument, or vow.
Go back to the instrument and drag α down toward 0.1. The glow
flees to the corners; toggle sample the population and the dots pile up on the pure
types. This is a world of purists — and no one in it decided to be a purist. The
culture simply made mixed positions improbable, the way thin air makes certain fires
unlikely. Purism here is a background field, not a personal virtue.
Now drag α up past 1 and the centre blooms: everyone a fluent,
even wash of all three traditions, and the committed devotee becomes the statistical freak.
At α = 1 exactly, the triangle goes flat — anything goes, every blend equally thinkable.
Then notice what never changed: not one word of any tradition's vocabulary. Stoicism
is exactly as Stoic at α = 0.1 as at α = 5. Only the concentration moved.
Place a genuinely mixed person near the centre and draw five words. The dashed estimate lunges to a corner: five draws rarely visit all three vocabularies, so a brief acquaintance hears a devotee where there is really a blend. This is the arithmetic of first impressions — sparse evidence is almost always purer than the person who produced it. Now draw two hundred. The estimate settles onto the truth, because a long acquaintance gives every allegiance time to show itself.
The instrument's listener is deliberately naive: θ̂ reads the words alone, a maximum-likelihood reading with no prejudice. A real listener also carries the culture's α and uses it to fill in what speech underdetermines. In a purist age, ambiguity is heard as purity — surely she's really one of us; in a syncretist age the same silence is heard as blend. Reputations are finished by the prior long before they are finished by the evidence.
Once you have the dial, you see it being turned everywhere. Ecumenism and orthodoxy argue about doctrine, but what actually separates them is a concentration setting — how much probability the age assigns to mixed allegiance. Universities run the same triangle: departments are corners, and the tenure system is an α set well below 1, while "interdisciplinarity" initiatives are attempts to nudge the dial upward without touching a single discipline's content. Cuisines too: the purist kitchen and the fusion kitchen cook from the same pantry of traditions; what differs is how thinkable the centre of the triangle is.
| Mathematics | Life |
|---|---|
| a corner of the simplex | a pure tradition — one complete way of being, undiluted |
| the point θ | one person's blend of allegiances |
| the concentration α | the culture's background pressure toward purity or blend, set before anyone chooses |
| drawing a word | one observable act or sentence — the only part of a person anyone else ever sees |
| the empirical counts | reputation's raw material: the blend others can actually tally |
| the inferred θ̂ | reading someone's commitments from finite evidence — always late, at first too pure |
The traditions supply the corners. The age supplies the concentration. Most of what we call a generation's character — its zealots, its dabblers, its rare honest blends — is the second thing, not the first.