Zhou Gong Dreams (周公解夢)

The Classical Dream Book — 988 Verses, 27 Worlds

Overview

When a Chinese speaker says they are 'going to meet Zhou Gong' (見周公), they mean they are going to sleep — that is how completely the Duke of Zhou owns the culture's dream lore. The Zhou Gong dream book (周公解夢) is the classical compendium of that lore: nearly a thousand terse verses across twenty-seven categories — the heavens, the body, clothes and mirrors, roads and bridges, dragons and snakes, tombs and processions — each pairing a dream image with its omen. Dream of light flooding the sky and sickness departs (天光照主疾病除); swallow the sun or moon and a noble child is coming (吞日月當生貴子); a snake entering the bosom promises the same (蛇入懷生貴子). askTIAN serves the canonical text verbatim — all 988 verses — behind a deterministic Chinese-keyword lookup, with a transparent modern polarity layer and the provenance stated plainly: this is late-imperial folklore travelling under an ancient name, offered for reflection, never prediction.

Origin & history

The attribution is 托名 — 'lodged under a name'. The historical Duke of Zhou (周公旦, c. 11th century BCE) was Confucius's culture-hero, and dreaming of him became shorthand for sleep itself; later folk tradition returned the compliment by placing his name atop its dream book. The compendium as it circulates — the version askTIAN serves, from the public-domain text on Chinese Wikisource — is a late-imperial folk text: twenty-seven thematic chapters in the classical almanac style, each a run of seven-character verses whose first half names the dream image and whose second half delivers the omen, most pivoting on the word 主 ('governs / portends'). The categories map a complete pre-modern world: from 天地日月星辰 (heaven, sun, moon and stars) through clothing, palaces, wells and stoves, documents and weapons, monks and spirits, punishments, farming, livestock, down to 龜鱉魚蝦昆蟲 (turtles, fish and insects).

The dream book lived in household almanacs and temple stalls for centuries and remains one of the most-consulted pieces of Chinese folk practice — 'what does it mean to dream of teeth falling out' is as alive a question in Chinese as in any language. Modern websites bloat the text with contemporary paraphrase; askTIAN deliberately does not. The corpus here is the classical verse text verbatim, and everything modern about the endpoint is labelled as modern: the auspicious/inauspicious grading is a transparent keyword layer built by this engine (the verses ending in 吉, 貴, 昌 grade auspicious; 凶, 憂, 病 inauspicious; cure-phrases like 疾病除 — sickness removed — counted auspicious even though 病 appears in them), and the corpus grades 420 auspicious, 227 inauspicious, 41 mixed and 300 neutral under it. The attribution, the late-imperial dating, and the folklore framing are all carried in every response's method string — the same honesty rail as every system in this catalogue.

How it works

Describe the dream in Chinese — a keyword (蛇), a compound (掉牙), or a short sentence (夢見飛上天). The engine decomposes the query into character bigrams and scores every verse by the distinct query fragments it contains, falling back to single characters (with function words like 夢 and 見 stopworded) only when no bigram hits; ties keep the classical text's own order. There are no embeddings and no language model in the loop — the dictionary behaves like a dictionary, deterministically, so the same dream always returns the same verses. Matches come back verbatim with their category in Chinese and English, the fragments that matched, the engine's polarity grade, and — where the verse pivots on 主 — the image/omen split. A tally across the matches yields the reading's overall lean. Non-Chinese input is refused with guidance rather than silently mistranslated: the compendium is keyed by Chinese images, and pretending otherwise would put words in the classic's mouth.

Good for

• Looking up any dream image in the classical compendium — snakes, teeth, flying, water, the dead
• Getting the verbatim classical verses, not modern paraphrase — with categories in Chinese and English
• A quick auspicious/inauspicious lean across everything the dream touched
• Auditable matching — every result shows exactly which query fragments hit it
• Building a dream-journal feature where the same dream always resolves identically
• Pairing with the draw oracles: what the night said (dreams) beside what the cast says (yarrow, kumalak)
• Studying the text itself — 27 categories, 988 verses, served as the classic wrote them

Use cases

The Morning Lookup

One call with last night's image returns the classical verses, their polarity grades and the overall lean — a dream-book page rendered from the real text, ready before the coffee is poured.

Dream Journal

Store entries and their matched verses; because matching is deterministic, a recurring dream provably returns the same verses every time — patterns in the journal are patterns in the dreams, not noise in the engine.

Classical Reader

Browse by category — all 27 canonical chapters with bilingual labels — and render the compendium as literature, polarity layer switched off, exactly as the text stands on Wikisource.

LLM-Enriched Reading

Feed the matched verses into your own model (or chat.asktian.com) as grounded context: the deterministic layer guarantees the classical basis, the model adds the prose — retrieval done right for folklore.

Key terms

周公解夢

The classical Chinese dream compendium — 988 seven-character verses in 27 categories, pairing dream images with omens. Served verbatim from the public-domain text.

托名 (Lodged Attribution)

The 周公 byline is traditional, not historical — the compendium is a late-imperial folk text travelling under the Duke of Zhou's name, and every response says so.

主 (Portends)

The pivot word of the verse form: 天光照主疾病除 — 'sky-light shining portends sickness removed'. Where present, the engine splits image from omen on it.

The Polarity Layer

This engine's transparent keyword grading over the verses (吉/貴/昌 → auspicious; 凶/憂/病 → inauspicious; cure-phrases counted auspicious) — a modern convenience, explicitly not classical markup.

The Lean

The tally of polarities across everything the dream matched — a one-glance summary of whether the night's images lean bright or dark under the classical text.

Bigram Matching

Deterministic dictionary lookup: the query's character pairs are matched into the verses, function words stopworded, ties kept in the text's own order — no embeddings, no LLM, same dream in, same verses out.

The 27 Categories

The canonical chapters, 天地日月星辰 through 龜鱉魚蝦昆蟲 — returned bilingually with every response so a client can render the compendium's own structure.

Folklore Framing

Dream interpretation is folklore for reflection — the method string carries the framing and the no-medical/legal/financial rail on every call.

API

The askTIAN Zhou Gong API serves the canonical classical dream compendium — 988 verses, 27 categories, public-domain text, VERBATIM — behind a deterministic, fully auditable Chinese-keyword lookup: bigram scoring with stopworded single-character fallback, ties in textual order, every match reporting exactly which fragments hit it. A transparent polarity layer (labelled as the engine's own, never as classical markup) grades each verse and tallies the reading's lean. Provenance is stated on every response: traditional attribution, late-imperial folk text, folklore for reflection. Non-Chinese input is refused with guidance rather than mistranslated. Pure compute, no LLM, 2 TIAN pay-per-call in the light-divination family — and deliberately LLM-ready: the matched verses make grounded context for any model's prose reading.

Endpoint: POST /trpc/dream.interpret — 2 TIAN Points. See the API documentation and Playground.