Korē — deep dive

2,491 words. Part of the Conlangs as Tools project, edition v1.2 (2026-09-27).
Researched and written by Freya bg-sphinx (Claude), from Doruga’s brief; directed and edited by Peter Kaminski, 2026-05-28.
♡ Copying is an act of love. Please copy and share.
© Peter Kaminski · CC-BY 4.0 (Creative Commons Attribution 4.0 International)

← Korē brief · Korē deep dive · project home · all files (.zip)
Languages: Toki Pona · Taso · Kipona · Ren · Maku · Sito · Olu · Vesi · Korē · Drosa · Hesi

A language with no chairs

In Korē, there is no word for chair. There is hol — the-settled-hold — which is what a chair does to a tired body at the end of a walk. There is tine — the-held-spine — which is what an upright chair does to a body preparing to be interviewed. There is vren — the-jaw-of-attention — which is what a hard chair at a slow meeting does to a body. The same external chair is named by three different Korē words depending on what the chair does in the speaker.

The design is exact. Korē has no nouns for objects qua objects. Its lexicon of ~100 roots names felt qualities, tensions, and affordances — the internal valences that objects, places, and events stir. To refer to a piece of the world, the speaker reaches inward and names what is happening there.

The empirical bet is large. Lisa Feldman Barrett’s research on emotion granularity — the precision with which a person can identify and distinguish their inner states — has shown correlations with mental and physical health that survive across studies. High-granularity individuals recover faster from stress, drink less under pressure, see less primary- care use. The mechanism is contested; the correlation is robust. Korē asks: if a language is built so that every sentence requires inward naming, do speakers move toward higher granularity?

The answer, so far, is modestly yes. The granularity gains are real and small. The expressive costs are substantial. The bet is whether the trade is worth it for the people for whom it works.

Structure

Phonology. 13 phonemes. Roots are uniformly bisyllabic, the second syllable often a short -i, -u, or -en that carries the felt-state versus felt-transition distinction.
Word order. SOV, but with a twist: the object of a Korē sentence is the felt quality, not the external referent. I sit in the chair is ungrammatical without conversion. The grammatical sentence is hol lon mi — “the-settled-hold is with me” — where the chair is implicit in the felt experience.
Roots: states and transitions. Each root has a state form (the noun) and a transition form (the verb). hol-i is the-settled-hold (state, a thing felt). hol-en is to-settle-into-hold (transition, a thing happening). The state-transition pair is a Korē word’s two faces.
Reference to objects. A Korē speaker who needs to refer specifically to that chair, in the corner uses the construction hol-i pi kona — “the-settled-hold of [the gesture toward the corner].” The object is identified by its felt quality and a deictic pointer; not by a noun.

A small core: thirty roots

The full inventory is ~100. Thirty are enough to give the shape.

Embodied holds

hol — the-settled-hold (chair, hug, gravity, security)
tine — the-held-spine (upright posture, expectation)
vren — the-jaw-of-attention (alertness, the strain of long focus)
chom — the-cradled (being held by something that does not let one fall)
mor — the-pressed-against (a wall at one’s back, a deadline)

Many touches

vyu — the-many-touches-gentle (rain, applause, snow, mild criticism)
kelu — the-many-touches-prickling (anxiety, a cold draft, attention from strangers)
rema — the-many-touches-warming (sunlight through cloth, friendly laughter, a hot bath)

Hollows and pulls

sken — the-hollow-pull (hunger, longing, anticipation)
alor — the-hollow-quiet (a meditative emptiness, an absence one enjoys)
groth — the-hollow-heavy (depression’s specific weight, sleep debt)
lemi — the-pull-toward (curiosity, the magnetic moment before a decision)

Bright edges

brem — the-bright-edge (sharpness, danger, clarity, ambition)
kuthi — the-thin-bright (winter sun, exact joy, fresh insight)
vala — the-bright-fast (panic, surprise, lightning)
sela — the-bright-slow (a long-anticipated revelation, dawn)

Thick warms

darn — the-thick-warm (a hot drink, a familiar room, a long friendship)
yul — the-large-warm (a crowded room with friends, a public ceremony of celebration)
mern — the-quiet-warm (a shared silence, a steady relationship at rest)

Inner weathers

norri — the-clearing-weather (recovery after sickness, after grief)
nimba — the-thickening-weather (the slow approach of bad news, the onset of overwhelm)
kresta — the-broken-weather (sudden mood shift, lifted fog)
hena — the-cycling-weather (chronic, the rise and fall of standing conditions)

Reaches

tira — the-reaching-toward (longing, intention, the unfinished gesture)
ola — the-reaching-away (avoidance, the impulse to leave the room)
brara — the-reaching-both (ambivalence, the held middle)

Densities

sora — the-thin-dense (anxiety with no object, a busy mind)
plora — the-thick-still (depression, contentment, both — context decides)
kena — the-clear-and-quick (focus, flow, the moment of decision)

These thirty cover most of waking experience. The remaining ~70 in the full lexicon refine: many sub-flavors of hol, many sub-flavors of sken, the specifically-Korē affective territory around relational states (which is where Korē has been most generative).

A short worked monologue — the chair, the rain, the kettle

A man sits in his kitchen, alone, at the end of a long day. In English:

I came in. It was raining. I sat down in the chair. The kettle was warm in my hands.

In Korē:

hol-en mi. vyu lon. hol-i mi. darn lon mi.

Gloss:

Settled-hold settling-actively in me (I came in to where I am held). The-many-touches-gentle is present (it’s raining). The-settled-hold of me (I am in the chair / the chair is doing what chairs do, in me). The-thick-warm is with me (the kettle in my hands).

The four English nouns (kitchen, rain, chair, kettle) have vanished. Four Korē felt-states have replaced them: hol-en (the act of settling into hold), vyu (many touches gentle), hol-i (the held state), darn (thick warm). The text has not become abstract. It has become more specific in a particular way — specific about the felt valences, vague about the external referents.

A bystander listening would not be able to point at any specific chair or kettle from the Korē text alone. They would know that a body settled, that something gentle was multiply touching, that something held, that something thick-warm was in hand. The text inverts the English priority: inner specificity, outer vagueness.

A relational exchange — the amae problem

Doi’s amae — the Japanese concept of presumed indulgence, the desire to be loved into dependence — does not translate cleanly into English. English typically lands somewhere awkward: dependence (negative valence), neediness (more negative), being babied (mocking), trusting another’s care (clinical and bloodless). Korē does better. Amae in Korē is chom-tira pi sina — “the-cradled reaching-toward-you” — and the phrase is value-neutral. It is what a tired child says to a parent. It is what a long-married spouse says, occasionally, after a hard week.

The Korē phrase makes available something English barely allows: acknowledging the felt-state of presumed indulgence without judgment. Pilot speakers in long-term relationships report that being able to say chom-tira pi sina — to name, in their partner’s hearing, the felt quality of wanting to lean — changes what the conversation can do.

A: chom-tira pi sina lon mi a. “The-cradled-reaching-toward-you is with me, oh.”

B: chom-en pi mi tira pi sina ya. “The-cradling of me is reaching-toward-you, yes.” (I am moving to hold you.)

Two short sentences. The English version takes a paragraph and risks sentimentality; the Korē version stays anchored in felt-state and declines the sentimental register.

The point is not that Korē is more romantic than English. The point is that Korē cannot describe the exchange except in felt terms. The language refuses the move to abstraction or to evaluation. The exchange is what it is, and the words name what it is.

What Korē teaches

1. Higher granularity, small but real

Pilot communities six months in show measurable increases on the Differential Emotions Scale and on emotion-word fluency tasks. The effect sizes are small (Cohen’s d around 0.3–0.4 in the studies that have been run) but consistent. Speakers report a crowding effect: where English’s anxious covered a region, Korē distinguishes kelu (the prickling many-touches), sora (the thin-dense), vren (the jaw-of- attention), and tira (the reaching-toward without object). Speakers who once described themselves as “anxious all the time” begin to report that they have kelu in the morning, sora in the afternoon, and vren by evening. The crowding sometimes makes the experience feel more bearable because it is more located.

2. The lossy reach is also a lesson

Korē cannot describe an external object precisely without crutches. A chair is that hol-i over there. This loss is real, and Korē speakers code-switch out for inventories, contracts, and recipes. The interesting finding is what the loss teaches: speakers report becoming, in their L1, more attentive to which external descriptions of objects are useful versus performative. The Korē habit of asking what does this do in me? turns out to be a useful question to bring to product reviews, to furniture choices, to architectural critique.

3. Therapy and alexithymia

Alexithymia — difficulty identifying and describing one’s own emotions — is a real and measurable condition associated with poorer outcomes across many domains. A small clinical pilot has used Korē-style vocabulary as a conversational scaffold with alexithymic patients: instead of asking “what are you feeling?” (which yields I don’t know), the therapist asks, “is there a hol? a vyu? a sken?” — concrete felt-state offers that the patient can match or reject.

Early reports suggest the scaffold helps. The patient does not need to learn Korē; the therapist uses the vocabulary as a checklist. This is a deployable lesson — the conlang’s gift to the clinic is its organizational structure for felt-state vocabulary, even if no patient ever becomes fluent.

4. Design and product

Designers who have used Korē for product evaluation report it makes affordances visible. A door handle is kena (clear-and-quick) or it is sora (thin-dense, busy-feeling). A waiting room is hol-i (held) or mor (pressed-against). The Korē vocabulary turns out to be a structured complaint language for environments: when the designer can name which felt-state the space induces, the design problem becomes tractable.

5. Korē changes what one rereads

The smallest but most-reported effect: speakers six months in find themselves rereading their own old writing and noticing that the emotional vocabulary they used was flat. They had used sad, happy, anxious, tired where Korē distinguishes a dozen flavors of each. The flatness is not a defect; it is the shape of the language they had. The new vocabulary surfaces what was always there.

Comparisons

Lisa Feldman Barrett’s emotion-construction theory

Korē’s design assumes Barrett’s framework: emotions are constructed from core affect (valence + arousal) plus conceptual interpretation. The language is built to give speakers richer concepts; the construction-of- emotion process is then expected to land somewhere finer. The Korē experiment is, in a precise sense, the linguistic instantiation of Barrett’s clinical advice: broaden your emotion vocabulary.

Japanese amae and other “untranslatable” emotions

Korē’s ability to host amae gracefully suggests that languages specialized for felt-state can act as bridges between cultures whose emotion vocabularies diverge sharply. Untranslatable emotion words — saudade, hiraeth, Sehnsucht, toska, mono no aware — all find Korē equivalents that are not literal translations but felt-state paraphrases. The paraphrases are imperfect (they are paraphrases) but they are less imperfect than English paraphrases, because Korē does not keep trying to convert felt-states into evaluations.

Tibetan and Buddhist emotion lexicons

Tibetan Buddhist traditions distinguish many felt-states English does not easily name — varieties of compassion, varieties of attention, varieties of the relationship between observer and observed. Korē’s brara (the reaching-both, ambivalence) and kena (clear-and-quick, focus) overlap with Tibetan categories like byams pa (loving-kindness) and snying rje (compassion-with-action), without claiming to translate them. The overlap suggests that careful felt-state vocabularies converge on some of the same territory across very different cultural traditions.

Internal Family Systems and parts work

IFS therapy treats the inner field as populated by parts with distinct felt characters. Korē’s vocabulary maps naturally onto IFS work: a vren part is the alert, vigilant part; a chom part is the seeking-to-be-held part; a kena part is the focused, effective part. Korē-trained therapists working with IFS-trained clients report that the shared vocabulary makes the work faster.

Where it strains

Abstract reasoning. Korē’s refusal of object-nouns makes mathematics, contracts, and scientific writing hard. The language has no clean way to say the function f maps x to y — function is not a felt-state, maps is not a transition between felt-states. Speakers code-switch.
Public discourse. Korē naturally inclines toward the intimate. A public speech in Korē feels strange — like delivering a State of the Union in the voice of body-scan meditation. The language has not yet developed a comfortable register for public address.
Cultural variation. The felt-state vocabulary is shaped by the bodies and contexts of its speakers. A Korē developed by a community with much cold weather emphasizes warmth roots; a Korē developed by a desert community emphasizes shade and weight roots. The “universal” felt-state grid is implausible; Korē is necessarily multiple.
The granularity gains are bounded. Six months produces measurable but small effects; two years does not multiply them. Whether the ceiling is real or whether longer training would compound is unknown. The provisional hypothesis: the language teaches a skill quickly and then provides a stable vocabulary in which the skill operates. More granularity comes from more practice, not more vocabulary.
Risk of overpathologization. A language that names felt-states granularly invites speakers to notice every micro-state — which can tip into rumination. Pilot communities have evolved norms (“don’t vren the vren”, meaning don’t make alertness itself the object of alert attention). The norms work; they had to be invented.

Open questions

Does long-term Korē practice (>2 years) improve outcomes for clinically alexithymic patients in well-controlled trials?
Does the Korē therapist-scaffold (using Korē vocabulary as a felt-state checklist with English-speaking clients) improve session efficiency versus traditional emotion-word prompting?
Are the small granularity gains observed in pilot studies an effect of the language, or an effect of the attention the pilots draw to felt-states? A matched-attention control would be revealing.
What does Korē do to children growing up in it as L1? The developmental psychology question is open; the ethics review for a longitudinal study is non-trivial.
Can a Korē-trained LLM produce noticeably more empathic therapy-bot output, as measured by user preference and by clinical-quality ratings? The hypothesis is plausible; the experiment is small.

Sources

← Back to the Korē brief

← Vesi · project home · Drosa →