Wago, Kango, Gairaigo, Konshugo: The Four Vocabulary Strata of Japanese
Japanese sorts every dictionary word into four origin-based strata, called 語種 (goshu): wago, kango, gairaigo, konshugo. These are the native, Sino-Japanese, foreign-loan, and hybrid layers.1234 The strata are not a historical footnote. They predict how a word is written, how it sounds in a compound, and which register it carries.
Overview
What 語種 means and why this classification exists
語種 (goshu, literally "word-type" or "word-origin class") is the lexicographic classification of Japanese vocabulary by historical-source stratum. Major Japanese dictionaries treat the four-way split as foundational. Every headword in 大辞泉, 日国, and ニッポニカ is tagged for stratum.1234
The classification exists because Japanese has borrowed heavily since the 5th–8th centuries. It layered Chinese vocabulary onto a native stratum, then layered Western loans on top of both.123456 The strata behave differently in three measurable ways: writing convention (hiragana versus on-yomi kanji versus katakana), phonology (rendaku, accent), and register (casual versus formal versus technical).
The threshold for counting a word as a loan is system-internal assimilation, not origin alone. 改訂新版 世界大百科事典 distinguishes 外来語 (assimilated foreign imports) from 外国語 (foreign words not yet socially adopted): ラジオ counts as gairaigo because the form has been domesticated, while a hypothetical English-faithful レイディオ would not.7
The rough proportions in modern Japanese
The cleanest stratum-by-stratum counts come from 国立国語研究所 (the National Institute for Japanese Language and Linguistics, NINJAL). The key sources are its 1956 magazine-vocabulary survey (現代雑誌九十種の用語用字), reanalyzed by Yamazaki and Onuma in 2004, and its 1994 follow-up survey. These remain the most-cited published breakdowns; no newer NINJAL-grade stratum count of equivalent rigor has been published since.
The 1956 NLRI distribution by distinct word types (異なり語数) and by token count in running text (延べ語数):
| Stratum | Distinct types (1956) | Tokens in running text (1956) |
|---|---|---|
| 和語 wago | 36.7% | 53.9% |
| 漢語 kango | 47.5% | 41.3% |
| 外来語 gairaigo | 9.8% | 2.9% |
| 混種語 konshugo | 6.0% | 1.9% |
Source: 1956 NLRI magazine-vocabulary survey.10
The 1994 follow-up survey shows the gairaigo type-share rising sharply, from 9.8% to 35.8% in 異なり語数.11 Token share for gairaigo rises far less. This indicates that loanwords proliferate more as one-off type entries (terminology, branding, transient vocabulary) than as everyday repeated tokens.
The 言語学大辞典 (Kamei, Kōno, and Chino, eds., 三省堂 1996) gives the konshugo share as approximately 5% of the lexicon, consistent with the 6.0% / 1.9% range from the 1956 survey.12
These NINJAL surveys date to 1956 and 1994. They remain the canonical published stratum counts because no later government-affiliated breakdown of equivalent rigor has surfaced. BCCWJ (2011) is a balanced corpus, but its public-facing summaries do not split by 語種. Treat the percentages as established reference points for the shape of the lexicon, not as live measurements of speech today.1011
The takeaway pattern across the surveys is clear. Kango dominates as distinct words (most dictionary entries are kango). Wago dominates as tokens in running text (the most frequent words are wago particles, copulas, and core verbs and adjectives). Gairaigo has surged in type-share since the post-war period but contributes far less in tokens. Konshugo sits steadily at around 5% as a small but productive category.101112
Why this matters for an N3 learner
The strata are predictive in three ways the learner can use directly.
The writing system tracks the strata. Hiragana and kun-yomi kanji signal wago; on-yomi kanji jukugo signal kango; katakana signals gairaigo. A learner can usually classify a written word by inspection and read it accordingly.123136
The strata carry register. Kango is the Latinate-equivalent layer (formal, written, academic). Wago is the casual, spoken, and emotional layer. Gairaigo is the modern, technical, and fashionable layer. Picking the right word for the register often means picking the right stratum.123136
The strata gate rendaku. Rendaku, or voicing of the second element in a compound, happens preferentially in wago, rarely in kango, and almost never in gairaigo. The same surface compound can change classification (and prediction) by stratum.141516
和語 (Wago): Native Yamato Vocabulary
Definition and historical origin
Wago is the native pre-Chinese-contact Japanese vocabulary. デジタル大辞泉 defines it as "漢語およびその他の外来語に対して、日本固有の語。やまとことば。" (Native Japanese words, distinct from kango and other loanwords; also called yamato kotoba.)1
精選版 日本国語大辞典 gives the same contrast and lists ひと (人), われ (我), ある (有), and ゆたか (豊) as paradigm cases.8 日本大百科全書 notes two sound-pattern hallmarks of the historical wago stratum: no voiced consonants in word-initial position, and short one- or two-syllable elements that combine through compounding.9
改訂新版 世界大百科事典 makes the type-token point explicit: wago occupies a smaller share of distinct-type vocabulary than kango, but in actual token usage wago accounts for more than half of running text.17
Shibatani's English-language definition matches the Japanese references: wago are "those words in Japanese that have been inherited from Old Japanese, rather than being borrowed at some stage." They are typically polysyllabic, with strict CV / V syllable structure.186
How wago appears in writing
Wago is written either in hiragana or in kanji read with kun-yomi. Kun-yomi is the native reading assigned to a Chinese character to match a pre-existing Japanese word.1918 When the same kanji can take an on-yomi for a Sino-Japanese word and a kun-yomi for a native word, the kun reading is the wago route. 人 read as hito is wago; 人 read as jin in 個人 is kango.918
Verbs and i-adjectives are especially characteristic of the wago stratum. They carry okurigana (送り仮名), kana suffixes attached to the kanji stem: 食べる (taberu), 美しい (utsukushii), 思う (omou).
Register: casual, spoken, emotional
Wago dominates the everyday-conversation register. Domains owned by the native stratum include domestic life, the body, feelings, nature, and basic verbs and adjectives.1918 Shibatani's framing makes the register split explicit: "Kango are generally more formal, often restricted to writing, while yamato kotoba are more casual and more often used in speech."186
The token-share dominance of wago in the 1956 NLRI corpus (53.9%) is consistent with this register: the most frequent words in running text are wago, even though kango contributes more distinct entries to the dictionary.10
Worked examples
山の上から海が見える。19
"You can see the sea from the top of the mountain."
山, 海, 花, 食べる, and 思う sit at N5 or N4 level and occupy the daily-life domain that 大辞泉 and 日国 single out for the native stratum.19 The wago "way of saying it" is the spoken-register default. 日国 treats the contrast with kango as register-bearing, not merely lexical.8
漢語 (Kango): Sino-Japanese Vocabulary
Definition and historical layers
Kango is defined by the on-yomi reading principle. デジタル大辞泉 defines it as "日本語の中で、字訓ではなく、字音で読まれる語。また、字音で読まれる漢字から成る熟語." (Words in Japanese read with on-yomi rather than kun-yomi; or compounds composed of on-yomi-read kanji.)2
精選版 日本国語大辞典 reinforces the same reading-based test: kango are 字音語 ("character-sound words"), with the contrast against wago explicit.19
The kango stratum has internal historical layers. The on-yomi system itself runs across three named borrowing periods:
- 呉音 (go-on) brought via Korea in the 5th–6th centuries, prominent in Buddhist vocabulary.
- 漢音 (kan-on) is the prestige reading imported in the 7th–9th centuries from Chang'an, and the default for most kango.
- 唐音 (tō-on) covers late-period borrowings from Song-Yuan-Ming Chinese, sparse and vocabulary-specific.1920
日本大百科全書 also distinguishes kango by point of geographic origin: ancient borrowings (梅, 絵), Buddhist terms via Chinese (仏, 刹那, 葡萄, 牡丹), terms with various on-yomi origins (蒲団, 椅子, 銭), and Japan-coined Sino-Japanese (日本製の字音語) such as 火事, 返事, and 大根.20 ブリタニカ notes that some Japan-created compounds (自動車, 火事) are sometimes distinguished as 「漢字語」 from 「本来の漢語」.21
How kango appears in writing
Kango is written in kanji read with on-yomi, typically as two-kanji 熟語 (jukugo) with no okurigana between or after the characters.219 The on+on pattern is the unmarked default. Jukugo dictionaries class it as the standard compound shape, with on+kun (jūbako-yomi) and kun+on (yutō-yomi) marked as konshugo subtypes.45
The same kanji often carries a kun reading for the native word and an on reading for the Sino-Japanese word.
| Kanji | Wago reading (kun) | Kango reading (on) | Kango compound |
|---|---|---|---|
| 人 | hito ("person") | jin | 個人 ko-jin ("individual") |
| 学 | mana(bu) ("learn") | gaku | 学校 gak-kō ("school") |
| 山 | yama ("mountain") | san | 富士山 fuji-san |
| 海 | umi ("sea") | kai | 海洋 kai-yō ("ocean") |
Register: formal, written, academic
The kango register is formal, written, and concept-dense. Kango dominates in news, government, law, scholarship, academic prose, and business.220186 Shibatani and 日国 both frame the kango / wago register split as the Latinate-versus-Germanic equivalent for Japanese: paired synonyms across the strata are the productive pattern for register adjustment.1918
NLRI's 1956 magazine survey gives kango 47.5% of distinct word entries, the largest type-share of any stratum, consistent with the kango-as-content-vocabulary register.10
Worked examples
来週から新しい学校に通います。220
"I'll be commuting to a new school starting next week."
日本の経済は緩やかに回復している。220
"Japan's economy is gradually recovering."
会議で激しい議論があった。220
"There was a fierce debate at the meeting."
学校, 政府, 経済, 必要, 議論, 法律, 会議, and 説明 are all on+on jukugo. This is the kango default and the bulk of dictionary headwords in their domains.19 ニッポニカ singles out 火事, 返事, and 大根 as 日本製の字音語 (Japan-coined Sino-Japanese). This shows that the kango stratum is not the same as "borrowed from China": the stratum is defined by the on-yomi reading principle, not by point of geographic origin.20
Wasei-kango: Japan-coined Sino-Japanese
和製漢語 (wasei-kango) are Sino-Japanese words coined in Japan using Chinese-character word-building patterns rather than borrowed wholesale from Chinese. ニッポニカ identifies 火事 (kaji, "fire"), 返事 (henji, "reply"), and 大根 (daikon, "radish") as pre-Meiji examples, and ブリタニカ lists 自動車 (jidōsha, "automobile") and 火事 as canonical cases.2021
The Meiji period (1868–1912) was the productive era for wasei-kango as a tool for translating Western concepts. 西周 (Nishi Amane, 1829–1897) coined 哲学 (tetsugaku, "philosophy") in lectures published in 『百一新論』 (Hyakuichi-Shinron, 1874), along with 論理 (ronri, "logic"), 倫理 (rinri, "ethics"), and 現象 (genshō, "phenomenon").2223
宇田川榕菴 (Udagawa Yōan, 1798–1846) coined the modern Sino-Japanese chemical-element terms 酸素 (sanso, "oxygen"), 水素 (suiso, "hydrogen"), and 窒素 (chisso, "nitrogen") in the pre-Meiji decades.22
Many Meiji-era wasei-kango were re-exported to Chinese. 社会 (shakai, "society"), 自由 (jiyū, "freedom"), 科学 (kagaku, "science"), 経済 (keizai, "economy"), 政治 (seiji, "politics"), and 文化 (bunka, "culture") are now standard modern Chinese terms; the Japanese-language Wikipedia article on 和製漢語 cites over 800 such re-exports including 社会主義 (socialism), 共産党 (Communist Party), and 幹部 (cadre).22
西周は「哲学」という和製漢語を作った。2223
"Nishi Amane coined the wasei-kango term tetsugaku ('philosophy')."
「社会」も「経済」も明治時代に作られた和製漢語だ。22
"Both 'society' and 'economy' are wasei-kango coined in the Meiji era."
火事は古い和製漢語の一つだ。2021
"Kaji ('fire') is one of the old wasei-kango."
外来語 (Gairaigo): Loanwords from Outside China
Definition and historical layers
外来語 is the stratum of foreign-origin words assimilated into Japanese. It has one consistent exclusion: kango is filed as its own category. デジタル大辞泉 defines gairaigo as "他の言語から借用し、自国語と同様に使用するようになった語," noting a broad-versus-narrow distinction. In the broad sense, kango is included; in the narrow sense, the term refers chiefly to Western loanwords.3
精選版 日本国語大辞典 places the scope primarily on post-Muromachi non-Chinese loanwords. Modern Chinese loans such as 餃子 are sometimes included, and ancient Sanskrit translations written in kanji (涅槃 nehan) are excluded.24 ブリタニカ excludes kango explicitly because "foreign-language awareness is weak" for Sino-Japanese borrowings; they no longer pattern as foreign.25
The historical layers run across four major donor periods:
| Layer | Century | Examples |
|---|---|---|
| Portuguese | 16th c. | タバコ (tabaco), パン (pão), カステラ (castella), キリスト (Cristo) |
| Dutch (Edo via Dejima) | 17th–19th c. | ガラス (glas), ビール (bier), ゴム (gom), アルコール (alcohol) |
| English (Meiji onward, post-WWII boom) | 19th c. – | コンピュータ, テレビ, バス, ネット (80–90% of gairaigo today) |
| Other donor languages | various | アルバイト (German Arbeit), ランデブー (French rendez-vous), オペラ (Italian), イクラ (Russian) |
Sources: ニッポニカ and the English-language Loanwords-in-Japanese reference article, plus the donor-language inventory in the Coto Academy summary.132627
How gairaigo appears in writing
Katakana is the default writing convention. デジタル大辞泉 states "外来語は一般的に片仮名で表記される" as part of its definitional gloss, and the rest of the lexicographic tradition treats katakana as the gairaigo signal.31326
Foreign words are reshaped to fit Japanese mora structure. The chōonpu (ー) marks long vowels (コーヒー, ビール). Small kana handle special moras (ファ, ティ, ヴァ), and extended-katakana combinations cover foreign-only sounds (ウェ, ディ, ジェ).132526
Older borrowings sometimes appear in kanji (煙草 = tabako) or hiragana (たばこ); the katakana default is post-Meiji and consolidated by the post-war script reforms.1326
Register: modern, technical, fashionable, neutral-to-cool
The gairaigo register skews modern, technical, and rich in connotation. ブリタニカ notes that whether a loan is treated as 外来語 versus 外国語 turns on social acceptance, which is itself register-laden.25
Gairaigo dominates domains such as technology and computing (コンピュータ, ソフト, アプリ, インターネット), fashion (ファッション, スーツ, ドレス), food service (レストラン, カフェ, ランチ), corporate buzzwords (プロジェクト, マネジメント, スキーム), and pop-culture entertainment.1326
The gairaigo register is sometimes status-marking and sometimes the only available word. 国立国語研究所's 「『外来語』言い換え提案」 project (2003–2006) recommended native alternatives for confusing public-sphere katakana loans. The reason was that gairaigo's register split from kango and wago had become a comprehension barrier in official documents and journalism.28
Worked examples
パン comes from Portuguese pão, and コーヒー comes from Dutch koffie. Both are pre-Meiji loans that have been domesticated long enough to feel native to most speakers.
学生時代はアルバイトを掛け持ちしていた。132627
"I worked several part-time jobs in my student days."
アルバイト is from German Arbeit ("work") and belongs to the German layer that entered Japanese through medicine, academia, and military terminology.
駅前で待ち合わせるのがランデブーの定番だ。1327
"Meeting in front of the station is the classic rendezvous."
ランデブー is from French rendez-vous, illustrating the small but durable French layer in the cuisine, fashion, and art domains.
ミルク (from English milk) sits in the same sentence as the kango synonym 牛乳. This is a typical Japanese pattern: the gairaigo cousin coexists with a native pairing and signals modernity, casualness, or fashion.
The 1994 NLRI magazine-vocabulary survey shows the gairaigo type-share rising to 35.8% from the 1956 figure of 9.8%, confirming the post-war boom in distinct katakana loan entries.11 The boom is asymmetric: gairaigo tokens grow far more slowly than gairaigo types. This indicates that loanwords proliferate as terminology and branding rather than as everyday-frequency words.1011
Wasei-eigo: English-looking words Japan invented
和製英語 (wasei-eigo, "Japan-made English") are Japanese expressions built from English-origin elements. Their meanings or forms do not exist in English. Academic linguistics classifies them as pseudo-loanwords or pseudo-anglicisms, meaning loan-like words created inside the borrowing language (Miller 1998; Miura 1998; Loveday 1996; Kay 1995, all cited via the Wasei-eigo reference article).29
Worked examples cited in the Wasei-eigo article and the gairaigo lists include スキンシップ (sukinshippu, "skin" + "kinship," meaning "physical affection between intimates"), カンニング (kanningu, from English "cunning," meaning "cheating on a test"), and ハンドルキーパー (handoru-kīpā, "handle" + "keeper," meaning "designated driver," with "handle" itself meaning "steering wheel" inside Japanese gairaigo).29
Other widely cited wasei-eigo include サラリーマン (sararīman, "salaryman" meaning "salaried office worker"), ベビーカー (bebī-kā, "baby car" meaning "stroller"), ナイター (naitā, "nighter" meaning "night baseball game"), and パソコン (pasokon, a clipped portmanteau of "personal computer"). The bulk of wasei-eigo coinage falls in the post-Meiji period, with a sharp boom after the Second World War.26
Wasei-eigo morphemes, or word parts, are loan-origin (English), but the resulting words are Japanese-only. They pattern as gairaigo in writing (katakana) and as Japanese in meaning.29
混種語 (Konshugo): Cross-Stratum Hybrids
Definition: any word combining two or more strata
Konshugo is the category for productive cross-stratum coinages. デジタル大辞泉: "異なる言語に由来する二つ以上の要素が結合してできた単語." (A word formed by combining two or more elements drawn from different languages.) Cited examples: 手製 (wago + kango), 長ズボン (wago + gairaigo), 邦文タイプ (kango + gairaigo), カフスボタン (gairaigo + gairaigo).4
精選版 日本国語大辞典 gives the same definition and enumerates the standard subtypes: 漢語＋和語 (重箱, 湯桶, 組合員, 運動する), 外来語＋和語 (長ズボン, カットする), 外来語＋漢語 (原子エネルギー, スリル感, スト権), and 外来語＋外来語 (シュークリーム, エレキギター).5
The category is the most under-taught of the four strata for foreign learners. In dictionaries, it is treated as a primary 語種 category, not a leftover category.4530 The 1956 NLRI survey gives konshugo 6.0% of distinct types and 1.9% of tokens; the 言語学大辞典 modern-Japanese figure is approximately 5% of the lexicon.1012
Pattern 1: kango + wago (重箱-style and beyond)
The 漢語＋和語 subtype is the 重箱読み (jūbako-yomi) reading pattern at the lexical level. The first element is on-yomi (kango), and the second is kun-yomi (wago).45
Worked examples from 大辞泉, 日国, and the Wikipedia 混種語 entry: 重箱 (jū-bako), 台所 (dai-dokoro), 番組 (ban-gumi), 残高 (zan-daka), 本屋 (hon-ya), 役場 (yaku-ba), and 運動靴 (undō-gutsu).4530
In 台所, 台 dai is on-yomi (kango) and 所 dokoro is kun-yomi (wago).
In 残高, 残 zan is on-yomi (kango) and 高 daka is kun-yomi (wago).
Pattern 2: wago + kango (湯桶-style)
The 和語＋漢語 subtype is the 湯桶読み (yutō-yomi) reading pattern at the lexical level. It has kun-yomi first and on-yomi second.45 Worked examples include 湯桶 (yu-tō), 場所 (ba-sho), 夕刊 (yū-kan), 手本 (te-hon), 身分 (mi-bun), 荷物 (ni-motsu), 雨具 (ama-gu), 家賃 (ya-chin), and 手製 (te-sei).4530
Verbal hybrids belong to the same family. 運動する (undō suru) and カットする (katto suru) attach a Sino-Japanese or English noun stem to the native verb する.5
公園の前の場所で待ち合わせよう。4
"Let's meet at the spot in front of the park."
In 場所, 場 ba is kun-yomi (wago) and 所 sho is on-yomi (kango).
In 夕刊, 夕 yū is kun-yomi (wago) and 刊 kan is on-yomi (kango).
Pattern 3: wago / kango + gairaigo
Gairaigo crosses with both native strata. The gairaigo-only subtype mixes two donor languages within one item.
和語＋外来語: 歯ブラシ (ha-burashi, "tooth" + "brush"), 粉ミルク (kona-miruku, "powdered milk"), 輪ゴム (wa-gomu, "rubber band"), ガラス窓 (garasu-mado, "glass window"), 長ズボン (naga-zubon, "long trousers").530
漢語＋外来語: プロ野球 (puro-yakyū, "pro baseball"), 原子エネルギー (genshi-enerugī, "atomic energy"), カップ麺 (kappu-men, "cup noodles"), ソーダ水 (sōda-sui, "soda water"), ガス器具 (gasu-kigu, "gas appliance"), スリル感 (suriru-kan, "thrill"), スト権 (suto-ken, "strike right").530
外来語＋外来語: シュークリーム (French chou + English cream), エレキギター (Dutch elek- + English guitar), レーズンパン (English raisin + Portuguese pão), ヘアサロン (English hair + French salon), テーマパーク (German Thema + English park).4530
Three-stratum hybrids exist too. 原子エネルギー is a clean kango + gairaigo case, but it sits alongside genuine three-way mixes. 半そでシャツ (han-sode-shatsu) combines 半 han (kango) + そで sode (wago) + シャツ shatsu (gairaigo) inside a single item, and the Japanese-language Wikipedia article on 混種語 lists 駅前ビル, 客寄せパンダ, カメラ小僧, マンション管理組合, and 大型観光バス as further three-stratum konshugo.30
子どもに歯ブラシで歯を磨かせる。530
"Have the child brush their teeth with a toothbrush."
歯ブラシ = 歯 ha (wago) + ブラシ burashi (gairaigo); a wa + gai compound.
プロ野球 = プロ puro (gairaigo) + 野球 yakyū (kango); a gai + kan compound.
原子エネルギー = 原子 genshi (kango) + エネルギー enerugī (gairaigo); a kan + gai compound.
シュークリーム = シュー shū (French chou) + クリーム kurīmu (English cream); a gai + gai compound across two donor languages.
Pattern 4: gairaigo stem + Japanese verb morphology
A productive sub-pattern combines a clipped gairaigo noun stem with the verbalizing suffix 〜る. The result is a godan verb conjugated as if native. The product is konshugo at the morpheme level (gairaigo root plus Japanese inflection) rather than the compound level.5
The canonical case is サボる (saboru, "to slack off" or "to cut class"), formed from サボ (clipped from サボタージュ, from French sabotage) plus 〜る. The verb conjugates as godan (サボらない, サボった) and behaves entirely as a Japanese verb.2631
Modern productive coinages include ググる (guguru, "to google" = グーグル + る), バズる (bazuru, "to go viral" = バズ from English buzz + る), トラブる (toraburu, "to have trouble"), and タクる (takuru, "to take a taxi"). The pattern remains productive in 21st-century slang.26
Other gairaigo-stem plus native-morphology compounds use する instead of 〜る: カットする (katto suru, "to cut"), キャンセルする (kyanseru suru, "to cancel"), アップする (appu suru, "to upload"). 日国 lists カットする as a 外来語＋和語 konshugo of the same family as 運動する.5
分からない言葉はググればいい。26
"If you don't know a word, just google it."
予約をキャンセルしました。5
"I cancelled the reservation."
Why konshugo proves the strata are alive
Konshugo coinages keep happening. Each decade brings a fresh batch of サボる-pattern verbs and a fresh batch of wago + gairaigo and kango + gairaigo compounds. The NINJAL 1994 follow-up survey's spike in gairaigo type-share (from 9.8% to 35.8%) feeds straight into konshugo, as new loans get attached to native morphology.1130
In dictionary treatment, 大辞泉 and 日国 list konshugo as a primary 語種 category with formally enumerated subtypes, not as a residual class.45
The four-stratum system describes how speakers feel each stratum's flavor. A kango-only term reads as formal, a gairaigo-only term reads as modern, a wago-only term reads as native, and a hybrid lets the speaker dial register by recombining flavor sources.13518
The register and domain effect
The kango / wago register pair
The productive pattern across the lexicon is paired synonyms: one kango entry and one wago entry for the same referent. The kango entry is formal, and the wago entry is casual.1918
| Referent | Kango (formal) | Wago (casual) |
|---|---|---|
| meal | 食事 shokuji | 食べ物 tabemono |
| question | 質問 shitsumon | 問い toi |
| speed / velocity | 速度 sokudo | 速さ hayasa |
| beginning | 開始 kaishi | 始まり hajimari |
| going home | 帰宅 kitaku | 家に帰る ie ni kaeru |
Sources: 日国 and Shibatani for the productive synonym-pairing pattern.192018
The register asymmetry mirrors English Latinate versus Germanic ("commence" against "start," "purchase" against "buy," "comprehend" against "understand"). 改訂新版 世界大百科事典 makes the comparison explicit: wago carries everyday function, while kango carries scholarly and abstract function.17186
Domain ownership
Each stratum dominates certain core domains.
- Kango owns government, law, academia, news, business, medicine, and formal correspondence. 47.5% of distinct dictionary entries in the 1956 NLRI survey are kango: the bulk of content vocabulary.10
- Wago owns daily life, kitchen, body, emotion, nature, basic verbs and adjectives, particles, and copula. 53.9% of tokens in running text are wago: the bulk of high-frequency function and core-meaning words.1018
- Gairaigo owns technology and computing, fashion, food service (especially Western cuisine), sports, entertainment, corporate buzzwords, and branding. By the 1994 NLRI survey it had reached 35.8% of distinct types.1126
- Konshugo sits across domains, with strong representation in compound food terms (カップ麺, ソーダ水, 粉ミルク), product categories (歯ブラシ, ガス器具, 半そでシャツ), and sports and media (プロ野球, スリル感).530
When gairaigo replaces an existing wago / kango word
The lexicon often carries both a native pairing and a katakana loan for the same referent. The loan signals modernity, casualness, or fashion.31326
| Referent | Gairaigo (modern) | Kango / wago alternative |
|---|---|---|
| milk | ミルク miruku | 牛乳 gyū-nyū (kango, standard) |
| rice | ライス raisu | ご飯 gohan (wago, everyday) / 米 kome (wago) |
| cancel | キャンセル kyanseru | 取消 torikeshi (wago, formal/written) |
| hotel / inn | ホテル hoteru | 旅館 ryokan (kango, Japanese-style) |
Sources: 大辞泉, ニッポニカ, Loanwords-in-Japanese reference article.31326
NINJAL's 「『外来語』言い換え提案」 project (2003–2006) recommended Japanese-side alternatives for opaque public-sphere gairaigo because the connotation shift was working against comprehension in government documents and journalism. The project's existence is the cited evidence that gairaigo replacement is a register choice with real comprehension consequences.28
A working register chooser
When the lexicon offers more than one option for the same idea, stratum usually decides the choice.
- Pick the wago synonym for casual, spoken, and emotional contexts (talking with family, intimate writing, narrative description of everyday life).186
- Pick the kango synonym for formal, written, and technical contexts (news, government, business correspondence, academic prose).1918
- Pick the gairaigo synonym for modern, technical, fashionable, or marketing contexts (tech, fashion, food service, branding, contemporary pop-culture).26
- If a katakana loan is opaque to the reader, NINJAL's 言い換え提案 recommendation is to substitute a kango or wago equivalent in public-facing writing.28
Why this matters for reading and listening
Reading prediction: the on'yomi default for jukugo
大辞泉 defines kango as 字音で読まれる漢字から成る熟語 (jukugo composed of on-yomi kanji). 日国 gives the same on-yomi-driven definition.219 The on+on default is the dictionary-level expectation for a two-kanji compound; jūbako-yomi (on+kun) and yutō-yomi (kun+on) are the marked konshugo exceptions.
The 1956 NLRI kango share of 47.5% (types) feeds straight into this prediction: a random two-kanji compound is likely to be kango with an on+on reading.21910
Rendaku is gated by stratum
Rendaku (連濁) is the voicing of the second element in a compound. It applies preferentially in wago compounds, rarely in kango compounds, and almost never in gairaigo compounds. The English-language Rendaku reference article, citing Vance (2008) and Irwin (2005), gives approximate frequencies as follows.
| Stratum of compound | Approximate rendaku rate |
|---|---|
| Wago + wago | ~90% in eligible compounds |
| Kango + kango | ~10% (Sino-Japanese mononoms ~20%) |
| Gairaigo | Near zero |
Sources: Vance 2008 and Irwin 2005, via the Rendaku reference article.141516
These rates are widely accepted academic approximations rather than the result of a single primary corpus count. Treat them as the established range, not as exact published percentages.
Lyman's Law adds a sound-pattern gate: rendaku is blocked when the second element already contains a voiced obstruent (/g/, /z/, /d/, /b/). This applies across strata but interacts with the stratum gate.1415
Classic worked examples by stratum:
- 折り紙 ori-gami (wago + wago; rendaku fires, kami becomes gami).
- 学校 gak-kō (kango + kango; no rendaku, kō stays kō).
- 歯ブラシ ha-burashi (wago + gairaigo; no rendaku, burashi stays burashi).1415
The practical implication is simple. If you see a two-element compound and the second element voices, you are looking at wago (or at minimum a kun-yomi side). If it stays voiceless across an eligible context, you are looking at kango or gairaigo.1415
Listening: the kango / wago register cue
News, government announcements, and academic lectures pile up kango. Dense on+on jukugo signal formal register.1018 Domestic conversation, casual narrative, and emotional speech run heavy on wago. They use more kun-yomi and hiragana-rendered vocabulary.1018
The 1956 NLRI token-share dominance of wago (53.9%) is the running-text texture of everyday conversation. The kango type-share dominance (47.5%) is the dictionary texture of formal content vocabulary.10
Good to know
Not every katakana word is a loan
Katakana is the gairaigo default, but the script also marks emphasis, onomatopoeia, and 標準和名. 標準和名 are standardized Japanese names for plants and animals in biological writing. イヌ is not foreign-origin. It is the wago 犬 written in katakana for emphasis or for biological-name convention. The definitions in 大辞泉 and 日国 hinge on foreign origin, not on script. So the script-to-stratum mapping is a high-probability default rather than a rule.32413
Not every on-yomi reading is kango
Treating every kanji compound read with on-yomi as kango misses two important exceptions. 大人 is read otona as jukujikun (a whole-word kun-yomi assignment), making it wago, not kango. Ateji items like 寿司 sushi and 煙草 tabako use kanji phonetically for an underlying wago or gairaigo word.203
The kango definition is specifically character-by-character on-yomi reading (字音で読まれる), and jukujikun and ateji assign readings outside that compositional mechanism.21920
大人 here reads otona as jukujikun and patterns as wago. Reading it with on-yomi (daijin, "an important person") would yield a different kango word.
Four scripts, four flavors as a working mnemonic
The script-to-stratum link is not a definition. Still, the high-probability mapping is durable enough to hold in working memory while reading.
| Surface form a reader sees | High-probability stratum |
|---|---|
| Hiragana or kanji + okurigana | wago |
| Two-kanji jukugo, no okurigana, on-yomi | kango |
| Katakana | gairaigo |
| Mixed-script compound | konshugo |
Sources: definitions and writing conventions across 大辞泉, ニッポニカ, ブリタニカ, and 日国.1234
The meaning shift of 和 in 和語 / 和食 / 和風
The character 和 in the labels 和語, 和食, and 和風 marks Japan's self-designation. The original Chinese exonym, or outside name, for the Japanese was 倭 ("dwarf" or "short people"). During the Nara period (8th century), the Japanese substituted the homophonous 和 ("harmony, peace") and adopted 大和 (Yamato, "great harmony") as the country's name.
The label 和語 thus encodes a millennium-long native-versus-foreign contrast: the "wa-" prefix labels things as Japanese-native against foreign (漢, 洋, 外).32
Why Chinese loans are not called gairaigo
ブリタニカ and 世界大百科事典 explicitly exclude 漢語 from 外来語. The reasoning is structural: kango borrowing predates the later loanword layers and underpins the language. The kanji writing system, the on/kun reading distinction, and much of the content vocabulary all flow from the kango stratum. So kango is treated as its own category rather than as a subset of gairaigo.21257
The dictionary threshold is system-internal assimilation. Kango has been so thoroughly assimilated since the Asuka–Nara periods that it no longer feels foreign to speakers. By the same logic, 世界大百科 frames ラジオ as 外来語 but a hypothetical English-faithful レイディオ as 外国語.7
See also
- Yojijukugo (四字熟語): The Japanese Four-Character Idioms Explained
- How to Read Long Kanji Strings: Chunking Three, Four, Five, and Six-Kanji Compounds
- Rendaku: When K Becomes G in Compound Words
- Mixed Script: How Japanese Combines Kanji, Hiragana, and Katakana
- Japanese Speech Levels: Plain, Polite, Formal, and Literary Register
- Japanese Onomatopoeia: The Four Classes (giongo, gitaigo)