Sound Series

Centuries of sound change separate modern Mandarin readings from the Old Chinese pronunciations that sound components originally recorded. The drift makes valid links look broken: a learner who meets 批 (pī, “batch; lot”) and finds 比 (bǐ, “compare”) inside it discards the connection without noticing, because the pinyin letters disagree. Reading the whole sound series — every character built on one component — restores the signal: family-wide, the variation collapses into a few related-initial groups, and an authoritative statement that those initials belong together gives the brain permission to use them. Children at a daycare ignored a playground about 30 yards from a busy road until a fence went up; the fence stopped no cars, yet the boundary let them play. Declared relatedness is that fence for phonology — once it stands, new instances of a pattern soak in almost effortlessly, and a trained learner perceives links that untrained native speakers cannot see.

Series membership is tone-blind, because Old Chinese likely had no meaning-distinguishing tones; speech stays tone-strict, because modern Mandarin does. A series where every member matches even in tone — 府 (fǔ) giving 俯 (fǔ), 腐 (fǔ, “rotten; spoiled”), 腑 (fǔ) — is very, very rare. The normal case is 侯 (hóu, “marquis”, a loan from “arrow target”) giving 候 (hòu, “to wait; await”), 喉 (hóu), 猴 (hóu): tones scatter and the family stays legible.

Seven Groups and the Mouth Test

When the rhymes across a series make no sense, the relationship lives in the initials. Seven groups collect the initials that substitute for one another; the test for each is physical — pronounce both candidates and notice where the tongue or lips make contact.

Group	Example series
{t, d}	同 tóng → 銅 tóng, 洞 dòng, 筒 tǒng
{zh, ch, sh}	成 chéng → 城 chéng, 盛 shèng/chéng, 誠 chéng
{j, q, x}	巠 jīng → 經 jīng, 輕 qīng, 勁 jìn, 徑 jìng
{b, p, f}	非 fēi → 菲 fēi, 排 pái, 悲 bēi, 輩 bèi
{g, k, h}	舌 guā (not 舌 shé, “tongue”) → 話 huà, 刮 guā, 活 huó, 括 guā/kuò
{z, c, s}	曾 céng/zēng → 增 zēng, 層 céng, 贈 zèng, 僧 sēng
{m, w}	門 mén → 問 wèn, 們 men, 聞 wén, 悶 mēn/mèn

A series is a sound object: member meanings are irrelevant to the pattern it carries and only obscure it during this stage of study.

• Groups have slack at the edges. ‘b’ and ‘p’ use both lips while ‘f’ touches lip to teeth, yet all three swap within the 非 series; ‘m’ and ‘w’ differ in lip shape and tongue placement but both form at the lips.
• The same alternation runs through Indo-European. Father/Vater/vader hold together across Germanic, fish/Fisch/vis stand against poisson/peces/pesce, brother/Bruder against frère/fratello — the f/b/p swap is general language drift, with Mandarin as one instance.
• Vowels can wobble inside a healthy series. The 舌 guā family alternates -uo and -ua (活 huó, 括 kuò against 話 huà, 刮 guā) while the initials hold the {g, k, h} pattern.
• Irregularities occur and get labelled. 勁 jìn ends in -n where its siblings end in -ng; flagging the exception preserves trust in the rest of the system.

Where Groups Cross

Sound change splits formerly identical initials, so some series straddle two groups. Three interactions cover the cases, each with an English anchor already in the ear.

• {g, k, h} ↔ {j, q, x}. The soft g of George sits near Mandarin j while German Georg keeps the hard g. 古 (gǔ) gives 故 gù, 估 gū, 苦 kǔ, 胡 hú on one side and 居 (jū, “dwell”) with 鋸 jù on the other; 夬 (guài, originally an archer’s thumb ring, now sound-only) gives 決 (jué, “decide”), 缺 (quē, “lack”), 快 kuài.
• {d, t} ↔ {zh, ch, sh}. English spells -tion with t and says sh; “watcha want?” fuses t+y into ch. The split mechanism shows in 店 (diàn, “store”): its component 占 (zhàn) anciently began with a d that later shifted to zh, while the slightly different d of 店 stayed put — one family, two modern groups. Series: 單 (dān, a hunting weapon; “single” is a loan) → 彈 dàn/tán, 戰 (zhàn, “war”); 專 (zhuān, “weaving tile; turning”) → 轉 zhuǎn/zhuàn, 傳 chuán, 團 tuán; 兌 (duì) → 稅 (shuì, “tax”), 說 (shuō/shuì, “speak; persuade”), 脫 tuō.
• {g, k, h} ↔ {l}. Rooted in the shape of the Old Chinese syllable: 各 (gè) gives 客 kè, 格 gé, 貉 hé, 胳 gē beside 路 lù, 略 lüè, 落 luò, 絡 luò.
• Let interactions emerge. Once the seven basic groups run automatically, cross-group pairs announce themselves during normal study; memorizing them up front adds load without payoff.

Running the Practice

• Install the seven groups first. Memorize them, say related members aloud, and notice the group every time a new character’s reading relates to its component.
• Resolve doubt with the mouth. When two sounds seem unrelated, pronounce both and check the place of articulation: ‘t’ and ‘d’ share one, while ‘p’ lives at the lips.
• Sing the membership. A verse stringing surnames across the interacting {zh, ch, sh} and {t, d} groups — pairing 張 Zhāng with 湯 Tāng, say — turns group membership into an auditory retrieval structure.
• Reread after the base settles. Early confusion is normal first contact; the same material reads easy once the groups are internalized, and the payoff is plausible pronunciation guesses at unstudied characters.
• Expect the model to explain multi-reading characters. Multiple readings usually branch from one original sense: 盛 chéng/shèng both descend from “to put things into a container”, 彈 dàn/tán from “pellet bow”, 說 shuō/shuì from “to explain”.

Where the Model Breaks

• Surface spelling overrides articulation. The 比/批 refusal repeats wherever pinyin letters get trusted over the mouth test; this is the exact moment the fence exists for.
• The pretty series becomes the expectation. Treating the all-fǔ 府 family as the prototype turns every normal tone-scattered series into apparent counter-evidence; tone variation is the default.
• Tone-blindness leaks into speech. The matching rule ignores tone; spoken Mandarin punishes that indifference immediately.
• Merged glyphs get misread. The 舌 in 話, 活, 刮, 括 abbreviates a distinct ancient character read guā; 舌 (shé, “tongue”) is a separate ancient glyph that collapsed into the same modern shape — a merger that is why the modern form 舌 carries both readings. 括’s own double reading guā/kuò is ordinary {g, k} and -ua/-uo variation inside the series.
• Sound-only components get promoted to meaning. 夬 contributes pronunciation alone in modern Mandarin; hunting for its semantic role in 決 or 缺 manufactures false structure. Likewise the 亻-like stroke in 侯 belongs to the depicted arrow target, while 候 carries a genuine 亻 (rén, “person”).

Related Pages

• How Chinese Characters Work - hub for this cluster and its reading order.
• Sound Components - the single-character mechanism this page scales up to whole families.
• Meaning Trees and Original Meanings - the reading forks of 盛, 彈, 說, 轉 are branches on one meaning tree.
• The IME Method - series-based pronunciation guesses feed its guided-guessing recall protocol.
• The Pipelining Strategy - the intake schedule whose spoken-word days draw on series knowledge.
• Rules of Effective Memorization - the fence principle and derive-rather-than-store rule, generalized beyond characters.
• The Hán-Việt Bridge - the same sound-family reasoning across languages, predicting Mandarin from a Vietnamese reader’s Sino-Vietnamese readings.

Sources

• Outlier Linguistics, Chinese Character Masterclass — commercial course; lesson PDFs kept locally outside this repository. https://www.outlier-linguistics.com/

Open Questions

• Is the seven-group inventory genuinely closed, or do further related-initial groups surface as more series accumulate?
• How far do these series project into other Sinitic languages and into Japanese, Korean, and Vietnamese readings — far enough for cross-language prediction?
• Where should sound-only components like 夬 live in this vault: on series pages alone, or as standalone character entries?